Kuro5hin.org: technology and culture, from the trenches
create account | help/FAQ | contact | links | search | IRC | site news
[ Everything | Diaries | Technology | Science | Culture | Politics | Media | News | Internet | Op-Ed | Fiction | Meta | MLP ]
We need your support: buy an ad | premium membership

[P]
Casting doubt on hi-fi reviews

By Mwongozi in Technology
Tue Jul 09, 2002 at 02:37:13 PM EST
Tags: Technology (all tags)
Technology

"particularly solid and clear bass, extended but without ever being excessive"
"tonality of most instruments and voices seems generally natural, with just a hint of dryness occasionally"
"a slight loss in tangibility of sound, which in turn makes for a little less listener involvement"
"Imaging is excellent and there is no detectable change in character with level."

On their own, you wouldn't be surprised to learn that these quotes come out of a hi-fi magazine. However, you might be surprised to learn that these quotes are describing different digital interconnects.


For those who are unfamiliar with digital audio, the beauty of it is that you can transmit it with no error, even through dodgy cabling. In other words, a digital cable has no effect on the sound whatsoever. Unlike an analogue audio cable, the only two bits of information that these cables need to carry are "1" and "0", or "On" and "Off". These bits are carried as either pulses of light, or pulses of electricity, depending on the cable. The data is reassembled into an analogue signal at its destination. Even if the cable was really poor, and introduced a lot of noise, it would take an awful lot of noise to make a "1" look like a "0", or vice versa.

Even if somehow enough noise was introduced to turn a "1" into a "0", this would not "distort" the sound, it would actually remove the sound altogether. The data contains checksums to ensure that it has been transmitted correctly, if it is corrupted, no audio is received.

So what's funny is that an otherwise reputable hi-fi magazine seems to be inventing differences between these cables when in fact, there is none. It's also disturbing to learn that people believe what they say, because one of the cables costs 300! Presumably, people must be buying it.

An e-mail to the magazine informing them of their obvious error has gone un-replied, are the magazines in bed with the cable manufacturers in order to sell wildly expensive cables to the ignorant public?

Remember, when buying digital cabling, there will be no change in your audio whatever cable you use. Buy the cheapest you can possibly find.

Sponsors

Voxel dot net
o Managed Hosting
o VoxCAST Content Delivery
o Raw Infrastructure

Login

Related Links
o hi-fi magazine
o digital interconnects
o Also by Mwongozi


Display: Sort:
Casting doubt on hi-fi reviews | 159 comments (154 topical, 5 editorial, 0 hidden)
Other considerations... (3.66 / 3) (#1)
by delmoi on Tue Jul 09, 2002 at 09:12:36 AM EST

Well, what about things like durablity? And length? (ok, that one was probably implied).
--
"'argumentation' is not a word, idiot." -- thelizman
cat5 vs OC-48 (none / 0) (#99)
by RandomPeon on Tue Jul 09, 2002 at 11:27:43 PM EST

They're not talking about these things. I'm not going to get better sound if I stream mp3s over cat5 or if I use OC-48 fiber. If there's enough bandwith in the channel it's Good Enough. (Use Shanon's theorem to determine the answer).

It's the same with any other digital media, all that matters is that it doesn't drop bits and can handle the your bandwith requirements.

[ Parent ]
digital information in an analogue world (3.81 / 11) (#2)
by beak on Tue Jul 09, 2002 at 09:24:44 AM EST

10-20 years ago when CD's first came out, people were saying things like 'But it's digital, it can't change!', and could not understand how, for instance, CDs with the same binary data could 'sound different' on the same equipment.

However since then the phenomenon of jitter has been established. Jitter is the variation in the timing of the 1's and 0's.

This can be cause by various effects (such as noise, filtering, or even multi-path delays in an optical cable) but results in the (prefereably) nice clean threshold between the 1 and 0 values being blurred. This blurring results in minor timing differences showing up when the threshold is crossed. These timing differences then introduce errors into the sound.

Yep, it sounds like Star trek techno-babble, but the wide range of recognition for this phenomenon either means that the speaker cable manufacturers either have a massive conspiracy going on, or it is real..

Digital Domain: What is jitter
MF Electronics Jitter FAQ [PDF]

There are now 'Jitter meters' being build to measure this effect.

Bottom line with audiophile stuff -- listen to it yourself: If you can tell the difference, then decide whether the difference is worth it... If not, then just take the cheapest one.

Timing Errors (none / 0) (#6)
by Bad Harmony on Tue Jul 09, 2002 at 09:43:10 AM EST

Jitter can be easily removed from a digital signal by regenerating the bit stream. This can be done with a fixed frequency clock if you don't mind an occasional bit slip. This is the approach used by the telephone company. Another method is to use a phase-lock loop to synchronize the local clock with the incoming bit stream. Either way, you sample the waveform of the incoming data in the middle of the bit time and use the value to regenerate the bit stream using the local clock.

5440' or Fight!
[ Parent ]

Jitter? (4.75 / 4) (#7)
by dennis on Tue Jul 09, 2002 at 09:47:11 AM EST

Then why don't I need a 300 cable between my motherboard and hard drive? That transmits data at a much higher rate than a CD, so it should be even more vulnerable.

If you're talking about the final output to the speaker magnet, that's another thing. Otherwise, this sounds like it's in the same league as the audiophile cd players, specially balanced to spin at a nice even rate. It's all going into a RAM buffer anyway, it doesn't matter.

[ Parent ]

No analogue source (5.00 / 1) (#9)
by Mwongozi on Tue Jul 09, 2002 at 09:58:52 AM EST

Jitter, in this context, would only apply from an analogue source being run through an ADC. However, digital interconnects carry data from a digital source, there is no sampling going on.

The bits will arrive at the destination in the same order in which they are sent, and it's up to the receiving device to have a reliable clock in which to pick them all up.

At no point can jitter be introduced by a simple cable.

There's a good article on jitter here.

[ Parent ]

Not a problem (5.00 / 2) (#10)
by tftp on Tue Jul 09, 2002 at 09:59:38 AM EST

Yep, it sounds like Star trek techno-babble

On K5 it sounds like kindergarten talk :-)

Serial communication always has a clock. It is either explicitly provided over a separate channel (wire), or it is embedded into the bitstream so that it can be easily recovered by the receiver. Jitter is nothing but a parasitic phase modulation of that clock. It is very easy to "clean" it, by for example synchronizing a LO to the recovered clock. Even simplest hardware implementations will be able to correct up to +- 50% of symbol time. Digital systems with memory have practically no limitations (no-skip CD players, for example, work this way.)

[ Parent ]

sounds like Star trek techno-babble (4.00 / 2) (#12)
by FredBloggs on Tue Jul 09, 2002 at 10:13:40 AM EST

Anything which suggests that a £2000 transport is `better` at getting digital data from a CD than the cheapest CD-ROM drive is forever destined to sound like babble. You can either read data 100% accurately (i`m talking about after its been had its clock bits or whatever dealt with perfectly), or you cant. People confuse the issue with stuff about DACs and aliasation etc, but actually getting the data off the CD does no require expensive parts.

Yes, i`ve listened to them, and no, I cant tell the difference. Same goes for uni-directional speaker cable, green dye on the edge of CD's, oxygen-free interconnects etc.

Lets face it - if i worked in a HIFI store, i`d be too honest to make them much money!


[ Parent ]

unidirectional speaker cable... (none / 0) (#157)
by Curieus on Fri Jul 12, 2002 at 08:38:09 AM EST

Now i am no audio phile, but there can be a reason for unidirectionability (word?) of speaker cable.
This would go for a shielded cable (for ex triaxial cable, or shielded twisted pair) where the shield is intended to dampen external interference signals.
In this case it is relevant on which side you earth the shielding.

But whether it has a serious impact on the quality of sound in normal households, i doubt it

[ Parent ]

Horseshit (4.00 / 2) (#40)
by trhurler on Tue Jul 09, 2002 at 12:39:49 PM EST

If this was a serious problem, it could have been completely solved a decade ago. Think. The timing of signals on modern computer networks has to be accurate to within hundreds of millionths or billionths of a second. The technology to do this stuff is actually pretty cheap; what makes network gear expensive is the engineering expense in coming up with newer better products every six weeks or so - NOT the actual technology used.

What's happening here is that marketing people at audio companies are desperately trying to find excuses for you to pay lots of money for magic fairy dust, because technology has simply moved beyond audio being a quality problem. Not a "conspiracy" so much as a lot of people with identical goals, all pursuing them in concert.

--
'God dammit, your posts make me hard.' --LilDebbie

[ Parent ]
Babble (4.50 / 2) (#47)
by chbm on Tue Jul 09, 2002 at 01:13:39 PM EST

Techno-babble indeed and you don't understand it.
The stuff you mention (like "multi-path delays in an optical cable") is the kind of stuff learned in EE courses not rag-mags. It's also very handy to be picked up by marketroid types as buzzwords to sell 1 m of low grade fiber for 450 EUR, which by the way is a ridiculous price even for 20 m of good grade fiber with well welded plugs used to transmit 155Mb/s not 1.5Mb/s.

The explanation in your first link is mostly babble too. The AES/EBU paragraph is manour. Moreover it's from 1996. The PDF seems like a good technical piece but bears no importance in the cable issue.

We can transmit at over 2Tb/s over 20km of cable. Sending 1.5Mb/s over 1 m of cable is not exactly rocket science and doesn't involve 450 EUR cables either.

-- if you don't agree reply don't moderate --
[ Parent ]

Physics (5.00 / 1) (#67)
by dennis on Tue Jul 09, 2002 at 04:37:31 PM EST

It's also very handy to be picked up by marketroid types as buzzwords to sell 1 m of low grade fiber for 450 EUR

A friend of mine once was in an audiophile store when the salesguy started spouting those buzzwords. It didn't make any sense to my friend, so he said, "I should warn you, I'm a graduate student in physics." The guy didn't miss a beat - he said "Oh, then you know all about this!"

[ Parent ]

2Tb/s over 20km? (none / 0) (#155)
by sgp on Thu Jul 11, 2002 at 09:45:14 PM EST

Can you back this up with a respectable URL? I've never heard any such claim before.

There are 10 types of people in the world:
Those who understand binary, and those who don't.

[ Parent ]

Buffering (none / 0) (#62)
by p3d0 on Tue Jul 09, 2002 at 03:30:02 PM EST

I don't need to be the 10th person to tell you this jitter issue is garbage. However, consider this:

If jitter were a problem, then a simple buffering scheme would fix it. If the speakers would just buffer, say, 1/100 second of digital data and play it back at a perfectly constant rate, then any jitter problem would vanish.
--
Patrick Doyle
My comments do not reflect the opinions of my employer.
[ Parent ]

Jitter (none / 0) (#134)
by phliar on Wed Jul 10, 2002 at 03:44:22 PM EST

Yep, it sounds like Star trek techno-babble, but the wide range of recognition for this phenomenon either means that the speaker cable manufacturers either have a massive conspiracy going on, or it is real..
You misunderstand the very page you provide a link to.

Jitter is an artifact of the D/A or A/D conversion process. Speaker cables and interconnects are "all analog." Digital cables are "all digital." No D/A or A/D conversions anywhere so no jitter.

... nice clean threshold between the 1 and 0 values being blurred.
You could use a finger to mark the bits on a disk of peanut butter as your digital medium, and you'd still get a perfect copy (although the data rate may be low).


Faster, faster, until the thrill of...
[ Parent ]

Wine and audio reviews (3.00 / 2) (#5)
by sien on Tue Jul 09, 2002 at 09:37:27 AM EST

This is very similar to comments in the diary, but the issue is worth expanding because it is an interesting juncture of subjectivity and objectivity.

Audio and wine reviewers are reviewing for affluent but somewhat silly people who have some idea and some money but who above all else enjoy reading about audio equipment or wine.

With audio equipment - another comment sums it up perfectly - see if you can hear the difference, if so buy something better, otherwise buy the cheapest.

The other thing about audio reviews is that there is such a strong fundamental limit in the attainable quality which is the quality of the equipment used to record the original sound. People who spend more than say about $15K on audio equipment are highly amusing. And of course the massive diminishing returns on what you get after about $2K are well worth considering.

As for wine, I read an article somewhere where they did some really careful double blind tests on wine reviewers and found that they were quite accurate, but had nothing like the precision to which they profess. But hey, I enjoy reading the odd wine review, they are fine style of creative writing.

But at the end of the day it's simple. Take some time and enjoy what you spend your money on and don't take reviews too seriously. It's all fun.

Intersection of two simultaneous phenomena? (3.66 / 3) (#11)
by LairBob on Tue Jul 09, 2002 at 10:03:32 AM EST

I think this is an interesting idea to kick around, because it seems to raise two very different issues at once:

1) Digital data / Analog media
Going back to the earlier "jitter" comment, we often misplace and overestimate our confidence in the reliability of digital info. When it comes down to it, especially in the instance of speaker cable, it's all still fundamentally the transmission of analog information. Translating the soundwaves into digital form before sending them through the cable is almost more a way of mitigating analog interference than it is a whole new "error-free" mode. Although I intuitively get the point of this story completely, I do think that at a fundamental level, it does mistakenly judge an analog mechanism by digital criteria. Technically, if you look hard enough, there's almost guaranteed to be a measurable difference in the actual transmitted signal between any two cables, let alone any two separate brands.

2) Self-justifying criticism
Nevertheless, it raises an issue that constantly concerns me with "whatever-o-phile" reviews...
Let me start by saying that I am not personally a highly sensitive connoisseur of anything--while I feel I can judge "good" from "crappy" for myself in stereos, wines, food, music, etc., I can comfortably state that none of my five senses are particularly well-honed.
Let me then add that I firmly believe that there are people out there with specific senses that are significantly and reliably more acute than mine. I have good friends with perfect pitch, or remarkably refined senses of taste, smell, etc. I trust that they can consistently and precisely experience the details of wine, sound, etc. better than I can, and I also trust that most reviewers fall into these groups.
Nevertheless, I have to agree on a gut level with the point of this story. I have a hard time believing that any reviewer--or more relevantly, multiple reviewers--would find consistent and predictable differences in these cables' sound qualities. Defining oneself as a reviewer pretty much requires that you establish distinctions between things. If you work back from that initial position, then in a case like this, you're essentially forced to introduce subtle and subjective/unproveable observations to justify the review's existence.
I might well be proven wrong, but I really have a feeling that these distinctions would not hold up to a statistically rigorous, "double-blind" kind of study. I do think that most hi-fi reviews are carried out in good faith, and discuss legitimate shadings in audio quality and character, but I also think the entire review industry works on a slippery slope, where the need to be authoritative drives a need to feel and appear precise at any cost. Like the author of this story, I feel that these digital cable reviews have probably slipped over the cliff.

WTF? (4.00 / 1) (#14)
by jmzero on Tue Jul 09, 2002 at 10:23:13 AM EST

Translating the soundwaves into digital form before sending them through the cable

Is this what your CD player does?  I hope not.  Why not read the digital information off the disc, and just send it?  Why would it convert it to analog, and then back to digital for the trip down?
.
"Let's not stir that bag of worms." - my lovely wife
[ Parent ]

Well, that part happens _way_ upstream... (3.00 / 1) (#16)
by LairBob on Tue Jul 09, 2002 at 10:38:32 AM EST

I was just talking about how sound waves that begin as clearly analog signals in the studio are translated into bits along the way before they go through the cable--granted, if you're playing a CD the digitization has happened well before you unwrap your personal copy, but my point isn't really dependent on where the digitial translation happens upstream of the cable.

[ Parent ]
Fair enough (none / 0) (#20)
by jmzero on Tue Jul 09, 2002 at 11:02:59 AM EST

Going back then..

whole new "error-free" mode

Indeed, I wouldn't suggest that digital recording is some error-free magic.  I would suggest, though, that between digitization and the D/A converter, it's very possible that music has not gained error.  And if there is some timing issue in the transport between these two places, then it seems like an issue of very poor design.

44 kHz is not stretching the limits of today's electronics, we should be able to get that signal there more or less perfect.
.
"Let's not stir that bag of worms." - my lovely wife
[ Parent ]

"Perfect" digital delivery via an analog (none / 0) (#24)
by LairBob on Tue Jul 09, 2002 at 11:46:22 AM EST

Well, I guess one good place to look at things is at the last gatepost you mention--the D/A converter within the speaker. Assuming that an inflow of digital data arrives, correctly check-summed, with all the right bits synched to correct timecodes, etc. at the speaker's D/A converter, then does it matter how it got there?

The straightforward assumption is that no, any analog variations introduced in the transmission, no matter how slight, are completely erased once the data crosses that threshold correctly and gets lined up in the digital buffer waiting for conversion. That, I believe, sums up the author's initial hypothesis, and underlies the assertion that any 'working' cable is as good as any other.

Is that too simplistic an assumption, though? I don't have near the EE background to know this, but even the simplest intro to the mechanics of electronic transmission makes it clear that even when you're just talking about transmitting an on-off signal across a wire, there are really hundreds of "moving parts"--things like switch bounce, signal resonance, etc. all introduce variations and noise into the signal. When you take a "simple" system like that and butt it up against something like a D/A converter, with its own internal "mechanics", is the final data delivery always consistently identical and flawless?

Is that a practical assumption in the real world? This really requires someone with a lot more insight into the EE details than I've ever had, but I'm definitely curious what more informed people think.

[ Parent ]
How it works (5.00 / 1) (#36)
by jmzero on Tue Jul 09, 2002 at 12:22:05 PM EST

I don't know how these things work.  But I know that they should.  If these things aren't using buffers and error correction and such, they should be.  It should be pretty simple technology these days to get a perfect digital signal from its digitization to your D/A converter.

I know that a CD is capable of doing this, because I use CD's for data.  I know that computers are capable of reading and writing perfect data on CD's.  I know that cheap cables can transmit digital data in lossless fashion.  Cheap cables all around my computer are doing this all the time.  

If all this stuff isn't happening in your $1200 digital audio setup, then that's a monument of poor engineering (and consumer apathy).
.
"Let's not stir that bag of worms." - my lovely wife
[ Parent ]

miracles of modern technology (4.00 / 1) (#57)
by tps12 on Tue Jul 09, 2002 at 03:00:43 PM EST

Think about your computer. Billions of digital signals whizzing around. But how many times do you open a file that used to say "tube" and find that it now says "cube?" It's just exceedingly improbable. In the rare case of hardware failure, you get complete failure, not subtle random errors.

Is it crazy? Yes, there is a lot of shit flying around in there. But compared to other feats of engineering (suspension bridges and the internal combustion engine...whatever got people to the moon), that we can control it is not so surprising.

[ Parent ]

Question(s) (none / 0) (#96)
by eightball on Tue Jul 09, 2002 at 09:45:57 PM EST

Wouldn't the errors be more along the lines of timing errors rather than data errors? Would you expect to notice a millisecond lag in accessing this file? I have less doubt that you could discern the difference of a millisecond difference between two speakers.

[ Parent ]
Uhh.. (none / 0) (#127)
by jmzero on Wed Jul 10, 2002 at 10:31:21 AM EST

Of course such timing errors could be introduced during recording, or at the D/A converter - but with proper design it's impossible between those two places.

A D/A converter shouldn't just play a bit whenever it happens to get there, it should be keeping a small buffer and queueing them out according to its own timer (I would think).  Again, I don't know if this is practice, but it seems logical enough.  

In any case, I wouldn't expect inconsistent timing errors.  Digital timers typically have to be a lot more accurate than "a few milliseconds".  The timing errors I would expect would be a consistently too long or too short cycle - but even then not by much.
.
"Let's not stir that bag of worms." - my lovely wife
[ Parent ]

I don't think you understand (none / 0) (#65)
by tzanger on Tue Jul 09, 2002 at 03:53:23 PM EST

Technically, if you look hard enough, there's almost guaranteed to be a measurable difference in the actual transmitted signal between any two cables, let alone any two separate brands.

Not with digital signals! If the protocol corrects for jitter, drops, noise, etc. then there is no difference -- With digital audio cables going 50 feet, unshielded telephone cable and high-quality optical cable will have zero difference. Now if you're out in space and getting bombarded by enough radiation and EMI that the EDAC can't cope then sure higher quality cable will help. However we're not talking about that; we're talking about the three feet of cable that goes between your CD player and your receiver. There will be no difference.



[ Parent ]
He has a point (3.00 / 2) (#100)
by RandomPeon on Tue Jul 09, 2002 at 11:36:39 PM EST

just not a very strong one. What he's talking about is the error introduced in the A-D conversion. You can't perfectly represent an arbitrary wave function in a fintite number of bits. This could be a problem. As a more general problem, it happens with any type of analog-digital conversion. Colors for example- there are a far more than 2^24 colors, but that's how many colors your monitor can display. Fortunately, the solution is easy - use enough bits so people can't tell. You eyes can't distinguish between 2 colors in a 24-bit color scheme, and your ears probably can't tell the difference in the A-D conversion in sound. If this were a real problem, CDs would never take off.

[ Parent ]
No ADC (4.00 / 1) (#121)
by Mwongozi on Wed Jul 10, 2002 at 04:44:10 AM EST

While this is true, there is no ADC in a digital cable. Any problems with the ADC are obviously not going to be affected by which cable you use.

[ Parent ]
What? (3.00 / 2) (#13)
by jmzero on Tue Jul 09, 2002 at 10:20:05 AM EST

You mean my $600 "bit sweetener" won't do anything :)

I can't believe how many people here believe in jitter.  Wow.  In any case, what does it matter?  CD's sound horrible - with any transport and any amp and any speakers.  I guess that's not fair - sometimes they sound OK.  But, in any case, it's pointless to talk about "hi-fi" 14.4 kHz CD's.  

If there was such a thing as meaningful jitter, or any differences in the devices for reading off the CD, then I suggest audio-philes try using my computer CD-ROM and the cheesy ribbon cable on the back of it.  In all the days I've used it for reading data CD's I don't think it has ever misread or "mistimed" a bit.  I could guarantee that all the information gets to the D/A converter correctly and "on time".

Then you only have to buy $6000 analog cables to connect your D/A converter to your amp to your speakers.  All to play your crappy CD music.

(and if you have an SACD player, then I'm not talking to you - as I don't have one yet and have no idea what they sound like)
.
"Let's not stir that bag of worms." - my lovely wife

"CD quality" (4.66 / 3) (#15)
by Mwongozi on Tue Jul 09, 2002 at 10:27:48 AM EST

CDs have a sample rate of 44.1khz, not 14.4. And now, if I may, I'd like to quote R3mix:

"CD was invented to be perfect sound without waste. The 90db signal to noise and dynamic range provides a noise floor that is lower than you can get from any analog source in the recording studio today. The air current in the room of the recording studio is louder than the noise floor on CD. When you use ANY microphone, you will pick up the room air noise. This means that CD already does a better job than we need it to. I already run into problems where CDs can record sounds too loud for analog equipment to safely amplify. If DVD audio is to be believed, then you could record a dynamic range wide enough to capture a jet engine's loudness. This is not possible to reproduce on current analog equipment without distortion and serious damage to your hearing. Again, CD is perfect. Current recordings on CD barely use any dynamic range. Most modern music has a "compressed" dynamic range. Constantly loud and rarely uses a sound below -15db on the level meter. This is a mastering problem. The mastering engineers master modern music for radio play to get their song louder than their competitors so people will pay attention when their song comes on. Take any 1980's or early 1990's CD and put it in your CD player, then listen to the volume. Now take a modern rock or pop music CD and play it. The volume of the modern music is always near or at the MAXIMUM peak level possible. The dynamic range squeezed out. Now, simply put in the older 1980's or early 1990's CD and turn up the volume on your stereo. You'll notice how much BETTER the older recordings sound. There is IMPACT in the drums. Details in the sound. It's more realisitic sounding overall. The older (but still modern) recordings are easier on your ears at louder volume and seem more natural. This is how the CD medium sounds at its best. Do not listen this way on PC or boombox speakers. You need a decent stereo or good headphones to hear the difference."

This equally applied to SACD - other than multiple channels and longer playing time, the increased quality is utterly pointless.



[ Parent ]
Mmmm (2.50 / 2) (#18)
by jmzero on Tue Jul 09, 2002 at 10:56:17 AM EST

Regardless of what particular sounds can be reproduced, there's a question of how much information can be encoded.  

Are we sure that 44kHz (at 16-bit precision, I think - and I'm probably wrong here too...) is good enough to outmatch human hearing in terms of how much information can be detected and represented?  

I don't know - and, as I say, I don't know whether SACD helps anything.  Perhaps the problem is in the D/A conversion process.  Or in the recording.  Or in the mastering as your quote suggests.  

All I know is that CD's sound horrible compared to analog sources like a good record.
.
"Let's not stir that bag of worms." - my lovely wife
[ Parent ]

Do you know what a -90dB noise floor means? (none / 0) (#22)
by fluffy grue on Tue Jul 09, 2002 at 11:31:46 AM EST

It means that in order for the sample noise to be audible, the peak signal needs to be about as loud as a jet engine.
--
I am a calm and tranquil flower.

Try the new Aborted Fœtus McFlurry! Cool and refreshing!
[ Hug Your Trikuare[ Parent ]

Noise... (3.00 / 1) (#29)
by jmzero on Tue Jul 09, 2002 at 12:04:07 PM EST

It means that in order for the sample noise to be audible, the peak signal needs to be about as loud as a jet engine

I didn't know that little factoid.  What I'm wondering is something like this:

Regardless of what particular sounds can be reproduced, there's a question of how much information can be encoded.

I'm no expert on sound technicals, and this may show through vividly...

Imagine you were surrounded by 200 people talking.  I realize that all the sound they're producing would sum together and interfere and such, but nonetheless there's going to be a vast amount of sound information filtering about.

I think humans have an extraordinary ability to glean information from this mess - perhaps more information than can be represented by a 44kHz signal - no matter what its noise floor is.

Perhaps these two are intermingled in a way I don't understand.  Perhaps having this low noise floor somehow proves it's able to store sufficient information.  If so, perhaps you could explain it.
.
"Let's not stir that bag of worms." - my lovely wife
[ Parent ]

2 different things (5.00 / 1) (#39)
by chbm on Tue Jul 09, 2002 at 12:39:49 PM EST

In that cross talk mess you can still only hear up to 19 or 20 kHz. There's a weird human listening phenomena that's some people are able to tune in a given voice and hear it over a lot of other voices. I'm not sure if it's because those people can filter out the other tones or if it's because they have better ears that can diferentiate spatial sound sources better but either it's got nothing to do with sampling frequencies.

Some people having an higher dinamic range would mean they could hear someone wispering over a jet engine roaring. I never heard of such case, in fact the tonal characteristics of human hearing make it so you can't even hear the person let alone get the the message.

-- if you don't agree reply don't moderate --
[ Parent ]

Three comments, not all to the parent post... (4.50 / 2) (#83)
by pla on Tue Jul 09, 2002 at 07:29:45 PM EST

I'm not sure if it's because those people can filter out the other tones or if it's because they have better ears that can diferentiate spatial sound sources better but either it's got nothing to do with sampling frequencies

Humans do that by predicting, both in terms of freqency and semantics, what will come next in the "thread" they wish to follow. If you take a single person and have them read two dissimilar passages from books, mix those together and play them back simultaneously from a single sound source, a listener can separate the two (though only comprehend one at a time). If the two passages have a high degree of similarity (about the same topic, similar turns-of-phrase at the same time, etc), people will have a harder time telling them apart (though may still manage it, with some effort and a few repeated listenings).


Okay, point #2... A high Dynamic range does *NOT* mean that a given sound has to have the loudness of a jet taking off nearby. The loudest sound may well only match a low whisper for intensity. Thus, we have a knob on the stereo labelled "volume". The dynamic range only describes the maximum loudness *difference* that the signal can express. The idea of a jet comes from the fact that the DIFFERENCE in sound intensity a CD can convey equals a bit more than the difference between the human hearing threshold and a jet taking off. As someone else pointed out, most music you hear has a fairly low dynamic range, because most of the instruments/singers get mixed at a similar level (in terms of RMS, not peak). That *doesn't* mean all music sounds as loud as a jet taking off, or even that most music has no "quiet" sections. It just means that the music will have few magnitude-based subtelties in it (drums should peak quite a lot higher than everything else... The only way to make it "sound" louder overall consists of boosting the RMS, which means either clipping loud sounds (undesireable) or boosting everything else relative to the loudest sounds.)


Finally, humans have an unusual (or rather, inexplicable) ability to perceive sub-threshold sound. This has no mystical connotations, just that, based on the understanding we have of how humans hear things, we can hear things we really aught not (like -12db beyond our theoretical lower limit). This and this alone lends *some* credit to the idea that analog audio can outperform adequately sampled digital. This limit, though, doesn't involve the sampling *rate*, but the resolution. More bits means a larger dynamic range. A larger dynamic range means more subtle changes in sound pressure. If those changes go below what we can perceive, then no, we cannot tell the difference. (As an aside, many people *can* hear above 22khz. I personally can hear up to almost 30khz, just barely. Our sensitivity in this range leaves something to desire, though, thus most people (even who can hear that high) don't even notice the absence of such extreme sounds).


[ Parent ]
Volume control (none / 0) (#90)
by fluffy grue on Tue Jul 09, 2002 at 08:17:37 PM EST

If something is properly mastered, then there should be no reason to touch the volume control. If something is recorded in such a way that you have to turn up the volume enough that the noise floor is audible, then the recording engineer should be sacked, because it means the recording isn't using the format's dynamic range.
--
I am a calm and tranquil flower.

Try the new Aborted Fœtus McFlurry! Cool and refreshing!
[ Hug Your Trikuare[ Parent ]

Yes (4.00 / 1) (#26)
by chbm on Tue Jul 09, 2002 at 11:53:14 AM EST

Are we sure that 44kHz (at 16-bit precision, I think - and I'm probably wrong here too...) is good enough to outmatch human hearing in terms of how much information can be detected and represented?
In terms you can understand: unless you're a dog or a bat, yes.
All I know is that CD's sound horrible compared to analog sources like a good record.
You should stop listening to Celine Dion CDs then.

-- if you don't agree reply don't moderate --
[ Parent ]
a lot of the problem is production (5.00 / 1) (#32)
by DavisImp on Tue Jul 09, 2002 at 12:07:33 PM EST

CDs, especially recently, have been produced to maximize loudness. The idea is that if the CD sounds a little bit louder than everything else you're listening to at the same volume, you'll be more likely to buy it (kinda like orchestras all tuning up a little bit higher than 440 Hz). Consequently, most modern CDs are highly compressed, eliminating dynamic range and making them all sound kinda shitty.

It's not a problem of the media, it's a problem with the post-production.

That said, r3mix is wrong about CDs being an ideal format. DAT has a higher dynamic range, and DVD audio can encode at a much higher bitrate. The truth is that you can tell the difference between a truly analog and CD digital source -- it's subtle, but it's there.

[ Parent ]

Um (none / 0) (#66)
by tzanger on Tue Jul 09, 2002 at 04:04:13 PM EST

CDs, especially recently, have been produced to maximize loudness. The idea is that if the CD sounds a little bit louder than everything else you're listening to at the same volume, you'll be more likely to buy it (kinda like orchestras all tuning up a little bit higher than 440 Hz).

Um how does loudness equate to detuning?



[ Parent ]
Detuning (none / 0) (#77)
by Bad Harmony on Tue Jul 09, 2002 at 06:18:12 PM EST

It's an infamous trick used by some radio stations. They play their records, CDs or data files at a slightly higher speed than the normal speed. It is supposed to make the radio station sound better than the competition, and allowed the station to insert more commercials.

5440' or Fight!
[ Parent ]

Records (5.00 / 1) (#35)
by jmzero on Tue Jul 09, 2002 at 12:12:46 PM EST

All I know is that CD's sound horrible compared to analog sources like a good record.

You should stop listening to Celine Dion CDs then.

As far as I'm concerned, this point is beyond contention.  Sometime you'll have to come to my brother's listening room.  We've got a good few identical recordings on record and on CD.  I can switch quickly between the two (with the songs lined up and the levels about the same).  

The CD loses badly.  Not in a "golden ear" sort of way, but in a "Is the CD player broken?  Are you missing a channel?" sort of way.  The difference is startling.  That said, if I play the CD first, people say it sounds great.  You don't know anything is wrong unless you compare the two.

As I say, I don't know why this is.  Perhaps it's a problem in recording or mastering of the CD.  Perhaps we've got a poor CD setup, although it's rated highly and certainly sounded good until we set up the turntable.

I suggest you try the test yourself.
.
"Let's not stir that bag of worms." - my lovely wife
[ Parent ]

Technical reason that CDs can suck (5.00 / 1) (#45)
by frankwork on Tue Jul 09, 2002 at 01:08:15 PM EST

One of the "mistakes" made when designing the CD spec was to encode the data on a linear basis rather than the logarthmic basis with which humans hear. To double the percieved amplitude of a sound, the actual amplitude has to be something like 10 times higher.

What this means is that while a CD has 16 bits to encode the loudest parts of a CD, the quiet parts often only have a handful of bits to work with, resulting in really disgusting levels of distortion on many sections of the recording.



[ Parent ]

Interesting (none / 0) (#48)
by jmzero on Tue Jul 09, 2002 at 01:16:18 PM EST

Your explanation certainly sounds feasible to me, as do 10000 other explanations I've heard about curves and nuances.  

Also, the people who talk about how CD's actually have way more range and detail than a human can discern seem to make convincing cases.  

All I know is that CD's usually don't sound good.  The whole science of audio seems to be a mess. Different equipment in different configurations sounds drastically different.  I can understand how speaker placement and room acoustics can be an art, and why it's complicated - but you'd think someone would be able to get sources and amplification figured out so it just sounds good.  
.
"Let's not stir that bag of worms." - my lovely wife
[ Parent ]

Yummy analog goodness. (5.00 / 2) (#52)
by ktakki on Tue Jul 09, 2002 at 01:39:56 PM EST

The reason many (not all) people perceive vinyl as sounding better than CD has a lot to do with the artifacts of analog recording: equalization curves, "soft" distortion, tape compression, and "head bump".

When the lacquer master record is first cut, the master tape is compressed and equalized -- bass is cut and treble is boosted -- as the signal is sent to the cutting lathe. This keeps the phonograph needle from mistracking lows and insures that the high end of the spectrum survives the manufacturing process. On the consumer end of the signal chain (preamp or receiver) there's a corresponding EQ stage that cuts the treble and boosts the bass according to this standardized (RIAA) equalization curve.

None of this is necessary when mastering CDs, where a laser scans a series of pits instead of a needle tracking a physical groove. Mastering a CD does involve EQ and compression, but this is done for aesthetic rather than technical reasons, though it is important to limit the signal; digital distortion sounds harsh.

Recording on to analog tape allowed a certain amount of "headroom", a soft limit before distortion could be heard. In certain cases (e.g., snare drum tracks) pushing into the range of distortion was desirable, something not possible with digital tape. Analog tape decks process the signal in predictable and interesting ways: compressing, limiting, and emphasizing certain frequencies. Garbage in, ice cream out.

k.
--
"In spite of everything, I still believe that people
are really good at heart." - Anne Frank

[ Parent ]

Does audio suffer from aliasing effects? (none / 0) (#46)
by squigly on Tue Jul 09, 2002 at 01:10:25 PM EST

Admittedly, I know next to nothing about audio, but I would have thought it would be quite hard to represent a 40kHz signal using 44000 samples per second.  Doesn't this give an effective 4kHz interference or something?

[ Parent ]
No (none / 0) (#50)
by chbm on Tue Jul 09, 2002 at 01:18:25 PM EST

Humans only hear up to 20kHz so CDs are designed to sample up to 22050 Hz ("real equipments suck" slak).
Niquist says fc = fs/2 so you wouldn't get 4kHz alias but 18kHz.

-- if you don't agree reply don't moderate --
[ Parent ]
I think you mean Nyquist [n/t] (none / 0) (#56)
by cyberdruid on Tue Jul 09, 2002 at 02:50:46 PM EST



[ Parent ]
Yes (none / 0) (#69)
by X3nocide on Tue Jul 09, 2002 at 04:54:48 PM EST

A rather famous engineer's theorem (I believe it was Shannon, but I'm not EE) condensed: any signal at more than double the sampling rate is not perceivable. The human ear has been rated at around 22k Hertz. So a 44.1k signal is more than enough to capture what humans can perceive. You're nutters if you actually understand what I said and believe that vinyl is techincally better.

pwnguin.net
[ Parent ]
Technically maybe... (5.00 / 1) (#70)
by jmzero on Tue Jul 09, 2002 at 05:03:25 PM EST

You're nutters if you actually understand what I said and believe that vinyl is techincally better.

I don't really care if vinyl is techincally better, it sounds better.  And I would have to be nutters not be believe that, given my own experience with both formats.

When was the last time you listened to a mint record on a quality turntable, a good tube amp, and hi-fi speakers?  Have you experimented with the two formats?  

We don't need theory to decide which sounds better - there's ample opportunity for empirical (if subjective) proof.
.
"Let's not stir that bag of worms." - my lovely wife
[ Parent ]

Harry Nyquist (none / 0) (#72)
by nixman on Tue Jul 09, 2002 at 05:46:07 PM EST

It was Nyquist who discovered/proved fs/2.

[ Parent ]
Something else (none / 0) (#73)
by nixman on Tue Jul 09, 2002 at 05:51:29 PM EST

You didn't state the theorem quite right: sampling at twice the maximum frequency allows (in theory) to exactly reproduce the original signal. In practice, most people use 20% padding.

[ Parent ]
Addendum.. (none / 0) (#92)
by mindstrm on Tue Jul 09, 2002 at 09:05:30 PM EST

That was only to reproduce a sine wave.

I realize you may see that as implicit.... but others won't.

[ Parent ]

Sum of sine waves (none / 0) (#105)
by pin0cchio on Wed Jul 10, 2002 at 12:23:42 AM EST

[The Nyquist-Shannon sampling theorem] was only to reproduce a sine wave.

Or any sum of sine waves below f/2, where f is the sampling frequency. Before quantization, the sum of samples of sine waves is the sample of their sum. The Fourier transform guarantees that any discrete signal can be represented as the sum of sine waves.


lj65
[ Parent ]
Almost (none / 0) (#151)
by vectro on Thu Jul 11, 2002 at 01:14:55 PM EST

Actually, Nyquist showed that you can reproduce any frequency up to but not including the sampling frequency. It's strictly less-than.

Alternatively, you cannot reproduce any frequency at or higher than the sampling frequency.

“The problem with that definition is just that it's bullshit.” -- localroger
[ Parent ]

Analog vs. Digital (none / 0) (#132)
by phliar on Wed Jul 10, 2002 at 03:05:47 PM EST

(God, I can't believe I'm getting sucked in again....)
All I know is that CD's sound horrible compared to analog sources like a good record.
I'm tempted to say that you're fooling yourself but I won't. Let me rephrase what you wrote: you like the sort of sound you hear from LPs you own -- which might say something about the type of mastering that is in fashion for the kind of music you listen to, but says nothing objectively about the abilities of CD and vinyl to accurately record sound.

People are used to the distortions added to the sound by the limitations of the physical needle/groove system. They're also used to the transfer characteristics of voltage devices like vacuum tubes. That's fine; if you like something, do it. But when some people start spouting mumbo-jumbo technobabble about "that rich analog warmth" (because they remember that their grandfathers' amps glowed and were warm)... well, I swiftly come to a decision on their intelligence and gullibility, smile, nod, and walk away.


Faster, faster, until the thrill of...
[ Parent ]

I wish... (none / 0) (#141)
by jmzero on Wed Jul 10, 2002 at 06:54:28 PM EST

I'm tempted to say that you're fooling yourself but I won't.

I was on the other side of this fence, until I heard a good record player.  I find that about 90% of the people who mock me for preferring records have never actually done much testing themselves.

you like the sort of sound you hear from LPs you own

I don't just like it more, but I also find it to be more defined.  As has already been demonstrated, I have no intelligent ideas on why this is.  If it's imaginary, then it's a shared hallucination with 20 or 30 people who have done the CD/record test at my brother's listening room.

I don't know why this is, but you have to understand that it's utterly undeniable to me.  You're welcome to come over and experience the difference any time...

I have no interest in promoting records other than my own wish that consumer audio was better.  I personally don't own a turntable - I have a cheap CD player in my basement.  My work doesn't have anything to do with audio.

"that rich analog warmth"

The subjective quality of sound is very hard to describe - so people use vague words like "warm", "bright", and "snuffle-grouty".  I tend to stick with "Is it enjoyable to listen to?" and maybe "Can I identify and follow distinct players?"
.
"Let's not stir that bag of worms." - my lovely wife
[ Parent ]

Don't get me wrong (none / 0) (#142)
by phliar on Wed Jul 10, 2002 at 11:05:33 PM EST

I was on the other side of this fence, until I heard a good record player.
I have lots of records and a good record player. You won't hear me on the side of those who say LPs are outdated and everyone should only listen to CDs etc. And I'm not saying that you suck because you "don't just like it more, but also find it to be more defined" either. What I'm saying is that "more defined" and "like it more" are the same thing.

An analogy: I have some old Charlie Parker vinyl. By every objective measure it sucks. The levels shift, the miking is terrible, and the pops and crackles are like a thunderstorm.I also have some CDs by Joshua Redman that are decently mastered etc. The Parker LP transports me to the bar that he's playing in; the Redman CD is just a record. The distortions of the LP are more "compatible" with my expectations of what jazz sounds like. I'd also say that a good recording engineer could make a CD sound like that, just like a good hardware engineer could make a solid-state amp that sounded just like a tube amp.

I mentioned my idea for a "digital null box" in message 133 in this thread. That would be a good starting point for a thought experiment. Do you think you'd know if the box was in the circuit or not?


Faster, faster, until the thrill of...
[ Parent ]

Question (none / 0) (#148)
by jmzero on Thu Jul 11, 2002 at 11:06:05 AM EST

Do you think you'd know if the box was in the circuit or not?

I have no idea...  

I, probably as much as you, would like some 'closure' on this debate.  You can find bafflegaff 'research' on this sort of thing all over the web, none I've seen conducted seriously.

I suppose those who write vague audiophile reviews can only lose by introducing more objectivity.
.
"Let's not stir that bag of worms." - my lovely wife
[ Parent ]

44 kHz at 16 bits (none / 0) (#133)
by phliar on Wed Jul 10, 2002 at 03:20:43 PM EST

Are we sure that 44kHz (at 16-bit precision, I think - and I'm probably wrong here too...) is good enough to outmatch human hearing in terms of how much information can be detected and represented?
What does "human hearing" mean? Whose hearing?

As far as numbers go, think 90 dB and a 20 kHz bandwidth. Is that enough for you? I only know that it's enough for me. Someone has already pointed out what 90 dB means in the real world. Let me add that the highest note on a piano has a fundamental of 4 kHz. That still leaves room for 5x overtones.

I've often wanted to build a "digital null box" that samples and quantizes audio input to 44.1kHz/16bit and then converts the digital data back to audio. This would be the test for anyone claiming to be an audiophile: for a signal of your choice, on equipment of your choice, can you tell whether my "digital null box" is switched in or out. If you can't do that -- end of argument.


Faster, faster, until the thrill of...
[ Parent ]

Aliasing for dummies - and a demonstration for all (none / 0) (#135)
by Blarney on Wed Jul 10, 2002 at 04:04:29 PM EST

In my humble opinion, aliasing can be a problem with CD audio. A CD can perfectly represent audio between DC and 22050 Hz - I will not argue with that. However, many audio sources such as microphones are capable of outputting signals higher than 22050 Hz. These signals, of course, cannot be heard by the human ear. If they make it onto the CD, though, their frequency will be shifted down to the point where it will be audible as interference.

For example, suppose you are recording an electric guitar playing a 440 Hz A with a very bright tone with lots of harmonics. The 51st harmonic of the signal will be 22440 Hz - but may well make it onto a CD and will show up at 22440-22050 = 390 Hz, well within the audio spectrum. Worst of all, 390 Hz is not a harmonic of the original 440 Hz signal, and will sound terribly dissonant.

This can be avoided while recording by carefully filtering off the high frequencies of the signal source before digitization, but this can introduce distortion. I understand that the modern way to handle this problem is by converting analog to digital at a much higher sampling rate, such at 96 KHz, and performing filtering and sample rate conversion digitally. However, there are many CDs out there that do have detectable aliasing.

Here's a demo for anyone, even those without "golden ears". Go and buy a vinyl copy of Sergeant Pepper by the Beatles. Go and buy a CD copy of the same. Listen to the end of the vinyl record, after "Day in the Life". There is a moment of silence right at the end which actually contains an ultrasonic note - if you have a dog, it will howl. Now play the CD - you will hear a high-pitched tone at the end. It has been shifted down into the human hearing range by aliasing, to a new note with no harmonic relation to the original.

[ Parent ]

This shouldn't happen (none / 0) (#156)
by Mwongozi on Fri Jul 12, 2002 at 05:55:08 AM EST

The Phillips spec for CD audio says that all analogue sources should be low-pass filtered at 20khz to avoid aliasing anyway. Double-blind tests have shown that no-one can hear the difference between full-range audio and 19.5khz lowpass audio anyway, so if the CD manufacturer fails to follow the spec, it's their fault, not the fault of the format.

[ Parent ]
Suggestion (none / 0) (#41)
by trhurler on Tue Jul 09, 2002 at 12:46:25 PM EST

Tell us what record, what CD, what your equipment is, how it is connected, and whether you've done this test double blind or not. Without that info, what you have is a lot of crap. As others have noted. you're not thinking of 14.4, which was an old modem speed.

I've never heard of a double blind test where people could tell the difference, IF you used a CD that had come from an analog master so that it had the same background noise(ie, DISTORTION OF THE INTENDED SOUND) as the record. If you use a pure digital recording and then press a record from that, then yes, people can tell the difference - but notice that the information that has been "lost" according to them could not possibly be added back in during the pressing of the record FROM THE CD. What they hear is NOISE.

--
'God dammit, your posts make me hard.' --LilDebbie

[ Parent ]
CD vs Record (none / 0) (#44)
by jmzero on Tue Jul 09, 2002 at 01:07:26 PM EST

We've done this test a few times.  I don't remember the exact equipment, but we tried a few configurations involving Antique Sound Labs amps, an Onkyo transistor amp, Martin Logan ReQuest speakers, B&W cone speakers, my Paradigm cone speakers, some Magnepan panels.  We used a few different CD players from regulars like Sony and from more high end focused brands.  I'll admit now I don't remember the brands or models - I'm sure my brother would if you're particularly curious.  

In any case, the results were clear and consistent.  Our test wasn't exactly double blind, but it was good enough for me.

We'd start the CD player and the record player on the same song (the same performance, and at about the same level).  With the Mackinnon preamp, we'd switch back and forth between the two as the song progressed.  We didn't say which was which, but of course it's fairly easy to identify the record simply by listening for artifacts.

Nobody questioned there was a difference, even those who aren't "into" music.  This wasn't "golden ear" sort of difference.  People (including myself) thought the CD player was broken, or that one of the channels wasn't hooked up.  That said, many of those same people were very impressed by the CD player before we played the record.  

Again, I don't know whether the CD recordings are flawed in some way or whether we just didn't use a good CD player, or whether the CD format isn't capable of making a good recording.  Whatever the case, the difference is not imaginary.  
.
"Let's not stir that bag of worms." - my lovely wife
[ Parent ]

good topic for an article (none / 0) (#59)
by tps12 on Tue Jul 09, 2002 at 03:14:39 PM EST

I have seen this vinyl vs. CD argument so many times, that I'm surprised there's not a good article online discussing a real double blind test, with recordings from both digital and analog masters.

Any interest in recreating your experiments more rigorously and writing something for k5? I know, it sounds like a lot of work, but there'd definitely be an interest (and another flame war, which is always good).

[ Parent ]

But... (none / 0) (#94)
by mindstrm on Tue Jul 09, 2002 at 09:09:11 PM EST

I think the test he's referring to is taking a CD, using it to master some vinyl, then comparing the two in a blind test where the listener has no knowledge of the process used to create them, then havin ghtem decide which one is better.

IF they come up wtih teh same argumetns about vinyl being better, then obviously there is something weird.. because there is no way the vinyl can be more accurate.


[ Parent ]

That would indeed be wierd (none / 0) (#125)
by jmzero on Wed Jul 10, 2002 at 10:15:47 AM EST

If vinyl made from a CD master sounded better.  

Perhaps not so surprisingly, I don't know anyone who could give this one a test.
.
"Let's not stir that bag of worms." - my lovely wife
[ Parent ]

Interesting (none / 0) (#126)
by jmzero on Wed Jul 10, 2002 at 10:25:10 AM EST

I've never heard of a double blind test where people could tell the difference, IF you used a CD that had come from an analog master so that it had the same background noise(ie, DISTORTION OF THE INTENDED SOUND) as the record.

In practice, I find it's easier to identify individual instruments on vinyl.  For example, it's easy to pick out what the drummer is doing on Brubeck's Take 5 on vinyl.  On the CD (which was transferred from the same analog master - I'm pretty sure it wasn't DDD) that detail is lost in the munge.

I find this is often the case.  On a record, it seems as though my ears are able to distinguish individual sound sources - as I would be able to distinguish individual conversations in a crowded room.

Again, I don't know why this is.  Perhaps some bad mastering practice, faded source, or such is at fault in the CD's poor performance.  Perhaps the "DISTORTION OF THE INTENDED SOUND" is what makes it sound more like a live performance.  I don't know.
.
"Let's not stir that bag of worms." - my lovely wife
[ Parent ]

PC Ragazine (4.66 / 3) (#17)
by jabber on Tue Jul 09, 2002 at 10:42:23 AM EST

are the magazines in bed with the cable manufacturers in order to sell wildly expensive cables to the ignorant public?

Good morning. Glad to see you're awake.

There are precious few "industry" or "special interest" magazines that are anything more than advertisement delivery mechanisms.

For most magazines, the news-stand and subscription income is gravy on top of the money they get from advertisers, either for ads, or for reviews.

This really should not come as a shock, but if it does, I envy you your innocence. If ignorance is bliss then anyone who believes "reviews" must be orgasmic.

[TINK5C] |"Is K5 my kapusta intellectual teddy bear?"| "Yes"

The quest for impartiality (none / 0) (#19)
by Mwongozi on Tue Jul 09, 2002 at 10:57:55 AM EST

So of course the big question is, where can you find impartial reviews of audio equipment?

[ Parent ]
Reviews of audio equipment (none / 0) (#21)
by jabber on Tue Jul 09, 2002 at 11:09:08 AM EST

Why, anywhere you can play with it yourself! :)

[TINK5C] |"Is K5 my kapusta intellectual teddy bear?"| "Yes"
[ Parent ]

Also, not to mention: (4.60 / 5) (#23)
by fluffy grue on Tue Jul 09, 2002 at 11:38:37 AM EST

The audiophile crew always like to bring up "jitter" and how it'll "distort the audio" by delivering bits out of order and so on, failing to realize that in a digital signal, if you manage to deliver bits out of order, you don't get subtle distortion, you get huge-ass pops and clicks.
--
I am a calm and tranquil flower.

Try the new Aborted Fœtus McFlurry! Cool and refreshing!
[ Hug Your Trikuare

Hmm, no. (5.00 / 1) (#31)
by chbm on Tue Jul 09, 2002 at 12:05:58 PM EST

Jitter means either the A/D or D/A converter has a bad clock signal and you deliver the samples slightly off timing. People trying to pass as audiophiles claim they can hear that.
Some people also refer to "jitter" when they mean clock skids between TX and RX of diferent equipments. That basicly causes 1 bit in every n bits to be lost in transmition (cables don't help though :)). I don't have present the line code used in SPDIF but I think it's safe to assume it's diferential and RX gets line clock.

-- if you don't agree reply don't moderate --
[ Parent ]
Okay... (5.00 / 1) (#53)
by fluffy grue on Tue Jul 09, 2002 at 01:57:55 PM EST

See, audiophiles also don't seem to even agree on what these bullshit terms mean. And anyway, the things you describe as being "jitter" would still cause more than "a very subtle distortion" in the audio stream.
--
I am a calm and tranquil flower.

Try the new Aborted Fœtus McFlurry! Cool and refreshing!
[ Hug Your Trikuare[ Parent ]

No kidding (5.00 / 1) (#81)
by pla on Tue Jul 09, 2002 at 06:49:16 PM EST

And anyway, the things you describe as being "jitter" would still cause more than "a very subtle distortion" in the audio stream

No kidding... Although digital transmisison sends information bitwise, DACs use a fixed word size (16 bits, for example). Losing or gaining bits due to clocking errors would most certainly *NOT* cause subtle, easily ignored errors. It would have the same effect as performing a rotation on each sample in software, ie, you don't get values wrong in one bit position (and *certainly* not just the LSB), you get values off by a multiple of a power of two. And you don't get just one sample like that, you get *every* sample like that until the clock slew between the two sides causes a multiple of 16 errors (in the same direction).

Without having the spec in front of me, I can still safely say that either the clock tolerances involved beat those of a typical cesium-133 atomic clock, or the signal itself uses a data framing protocol (much more likely). Translation - Jitter, as defined in this context, simple cannot exist without causing a catastrophic degradation of audio quality.


[ Parent ]
SPDIF (3.50 / 2) (#71)
by miles b on Tue Jul 09, 2002 at 05:29:13 PM EST

Hmm...By the name, one would think it's differential, but I'm inclined to think no.  It's a two conductor serial cable, so it couldn't possibly be differential in an XLR sense.  I'm thinking that the only way it could be differential is if they use multiplexing on it. I believe that AES/EBU actually is differential, but most people find that it's inferior to the sound quality of SPDIF.  From what I can tell, AES/EBU is mostly used in the pro audio community because the connectors are much more rugged than SPDIF.

However, here's the thing:  these cables are used to connect a digital source to a processor, both operating independently, with their own clocks.  Older digital sources may not produce a signal that's within range, so it may show up as 44.1 kHz, but the processor won't be able to lock onto it's clock.  It'll still be usable, but you'll miss a lot of detail.  Since it's not really possible for consumers to own a global clock to run all of their digital equipment off of, and since different digital equipment is manufactured to different tolerances, it is possible to have clocks that are slightly off, yet still work together.  A lot of the smarter DAC's will just internally convert it down to a 22.05 kHz signal, and throw away the values that are grossly erroneous. Different products do this to different extents.  On some CD's, you can hear loud pops and clicks through my DAC, but if I just use the built-in analog outs on my CD player, they're totally gone.  For most stuff, the external DAC sounds much much better, so I just live with the pops and clicks.  They're only on a few CD's anyway.

[ Parent ]

What part of the name tells you that? (5.00 / 1) (#88)
by fluffy grue on Tue Jul 09, 2002 at 08:10:46 PM EST

S/PDIF is short for Sony/Philips Digital Interface Format. Nothing about differential in there. :)

Also, I think you're confusing "differential" with "balanced." XLR is balanced. Regular two-wire interconnects are differential. Or at least in the form of terminology I'm familiar with. (Of course, to make things more difficult, the specific line-level digital encoding could be differential, such as Mersienne differential encoding, as used by Ethernet.)

Additionally, your "analysis" of what a DAC does when clock skew is grossly inaccurate. It doesn't "halve the data rate" or anything (since it'd require the sender to do that as well); instead, the receiving end synchronizes to the sending end. Not that hard, really, since there's only one sender and one receiver. There's no need for a global clock because the protocol is self-clocking. The DAC just syncs to the source.
--
I am a calm and tranquil flower.

Try the new Aborted Fœtus McFlurry! Cool and refreshing!
[ Hug Your Trikuare[ Parent ]

Sync to the source is what causes jitter (none / 0) (#103)
by pin0cchio on Wed Jul 10, 2002 at 12:18:11 AM EST

There's no need for a global clock because the protocol is self-clocking. The DAC just syncs to the source.

But if the signal becomes distorted in some way while crossing the interconnect, then so will the embedded clock signal. This is the primary cause of jitter in a DAC system. Jitter manifests itself audibly as harmonic distortion in the upper frequencies.

The way to eliminate jitter is to stick a ring buffer before the DAC that evens out the clock on the incoming signal. All CD Digital Audio players do this internally; in fact, most control the angular velocity based on the ring buffer.


lj65
[ Parent ]
WTF? (none / 0) (#104)
by fluffy grue on Wed Jul 10, 2002 at 12:22:38 AM EST

But if the signal becomes distorted in some way while crossing the interconnect, then so will the embedded clock signal. This is the primary cause of jitter in a DAC system. Jitter manifests itself audibly as harmonic distortion in the upper frequencies.
Care to explain how that works? Better yet, lend me some of the drugs you're on. I could use them after seeing some of the asinine idiocy that the audiophile apologists have posted to this story.
--
I am a calm and tranquil flower.

Try the new Aborted Fœtus McFlurry! Cool and refreshing!
[ Hug Your Trikuare[ Parent ]

references (none / 0) (#107)
by tuj on Wed Jul 10, 2002 at 01:07:32 AM EST

http://www.digido.com/jitteressay.html

http://www.amek.com/oldsite/datashee/aesebu.htm



[ Parent ]

Yes, I've read those (none / 0) (#108)
by fluffy grue on Wed Jul 10, 2002 at 01:14:39 AM EST

All those sites do is show what would cause jitter. They don't do anything to show that it actually makes a difference!
--
[ Hug Your Trikuare ]
[ Parent ]
jitter does cause distortion (none / 0) (#109)
by tuj on Wed Jul 10, 2002 at 01:33:49 AM EST

You said that those sites "don't do anything to show that it [jitter] makes a difference."

from the website:
"A properly dithered 16-bit recording can have over 120 dB of dynamic range; a D to A converter with a jittery clock can deteriorate the audible dynamic range to 100 dB or less, depending on the severity of the jitter. I have performed listening experiments on purist, audiophile-quality musical source material recorded with a 20-bit accurate A/D converter (dithered to 16 bits within the A/D). The sonic results of passing this signal through processors that truncate the signal at -110, -105, or -96 dB are: increased "grain" in the image, instruments losing their sharp edges and focus; reduced soundstage width; apparent loss of level causing the listener to want to turn up the monitor level, even though high level signals are reproduced at unity gain. Contrary to intuition, you can hear these effects without having to turn up the listening volume beyond normal (illustrating that low-level ambience cues are very important to the quality of reproduction). Similar degradation has been observed when jitter is present. Nevertheless, the loss due to jitter is subtle, and primarily audible with the highest-grade audiophile D/A converters."

Most jitter, contrary to a previous poster, has subtle effects on audio, due to the fact that most jitter is not extreme.  

again, from the website:
"But the closer you look at the phenomenon of jitter, the more you realize that even minute amounts of jitter are audible, even through the FIFO (First in, First Out) buffer built into every CD player. "

Now you might say "that is a subjective opinion" which is true, but, he also says on the site:
"I have participated in a number of blind (and double-blind) listening tests that clearly indicate that a CD which is pressed from a "jittery" source sounds worse than one made from a less jittery source."

My apologies if you noted all of these things in the article on that site, however, your response indicated that you did not.

(quotes from http://www.digido.com/jitteressay.html)



[ Parent ]

Yes, I noticed all those things (5.00 / 2) (#112)
by fluffy grue on Wed Jul 10, 2002 at 02:08:47 AM EST

There is one piece of simple evidence which will indicate to me that these people are not totally full of shit, and that would be a high-precision oscilloscope display taken directly off of the analog output from the same amplifier being fed with the digital input from two CD players, one of which is jittery and the other isn't (all other things being equal), showing the difference between the waveforms.
--
[ Hug Your Trikuare ]
[ Parent ]
sceptism is healthy (none / 0) (#138)
by Sawzall on Wed Jul 10, 2002 at 04:30:07 PM EST

but jitter does create an issue in chains - production or broadcast. Thus in this world, you use master clocks as opposed to having each device attempt to recover the timing on its own, and thus creating its own little errors. To take it to the extreme, but common place, example.

I play out a digital stream from a storage device, it is taken from AES to IP, then encapsuated, muxed, and then spit up a DVB encoder to a HPA transmitter to a satellite. The satellite transponder in turns sends it back down through the antenna, LNB, IRD, then to "demuxer" tuner, and now you have an IP stream that is decoded back to AES. You better have a way to generate clock!

Does this mean shit to the consumer? No. Jitter has not meant anything to the consumer since about '84 when the second generation CD players came out.

[ Parent ]

Oscilloscope pics (none / 0) (#143)
by tuj on Thu Jul 11, 2002 at 01:06:12 AM EST

http://www.jitter.de/english/sound.html



[ Parent ]
Useless (none / 0) (#147)
by Mwongozi on Thu Jul 11, 2002 at 10:38:50 AM EST

That's now useful at all. That's an oscilloscope shot of the digital signal itself. As I explained in the article, any error in the digital signal is inconsequential.

What would be good is an oscilloscope shot of the analogue signal distorted as a consequence of jitter. This is, of course, impossible to cause with a simple digital interconnect.

[ Parent ]

jitter can be caused by interconnects (1.00 / 1) (#150)
by tuj on Thu Jul 11, 2002 at 12:31:23 PM EST

Your statement that jitter cannot be caused by interconnects (cables) is simply wrong.  

Explaination:  a longer (or worse quality cable) will exhibit more capacitance than a shorter (or higher quality cable).  The effect of this is that the rise and fall times of the waveform will be increased on the longer cable, potentially causing errors in the digital interpetation of the signal.

This guy's ( http://www.codebunny.org/audio/cdtransports.html ) test using a coat-hanger as a cable shows only a few errors in the signal.  However, this site ( http://www.audioprecision.com/publications/audiotst/jan96/jan961.shtml ) shows how the cable can have an effect on the signal to the DAC, and also, in Figure 11, how bandwidth limitations caused by cables can have an effect on the analogue signal.  

Does the cable matter?  Its difficult to say.  In 100+ meter runs (150 nsec rise/fall) of AES/EBU, it probably does have some effect on the signal out of the DAC.  In a 2 meter run from your cd player to your amp, its probably not going to have much effect.  But don't say that jitter is impossible to cause with an interconnect, as this simply isn't true.




[ Parent ]

Not Entirely True (2.50 / 2) (#25)
by marktaw on Tue Jul 09, 2002 at 11:51:53 AM EST

CD players are meant to make up for bad data. I have a good CD player that could skip over a 1-2mm scratch on the disc and you'd never know the difference.

If any data is lost due to a digital cable, the receiver should be able to make up for the lost data as well, or just not play that slice of the waveform - there are 44 thousand per second on a CD. Small enough problems shouldn't be noticiable, but an A/B comparison may reveal something. Of course, a double blind placebo controlled test would be the only way to ensure it wasn't in the mind of the testors.

Incorrect (5.00 / 1) (#27)
by Mwongozi on Tue Jul 09, 2002 at 11:58:47 AM EST

You make two mistakes. Digital cables cannot "lose" data, and receivers do not interpolate missing bits. Any signal drop out would be manifested as moments of silence, unlike a CD player.


[ Parent ]
You make mistakes (5.00 / 1) (#28)
by MMcP on Tue Jul 09, 2002 at 12:03:51 PM EST

1) Digital cables can make mistakes.  Cut the cable and notice the subtle but noticeable change from music to silence

2) Receivers certainly do error correction when they have onboard DA converters (which many do).

[ Parent ]

exactly (none / 0) (#60)
by tps12 on Tue Jul 09, 2002 at 03:20:26 PM EST

Cut the cable and notice the subtle but noticeable change from music to silence

That's the point. With digital data, you either get drastic effects, like silence or static, or you get perfect data. It just makes no sense for there to be subtle differences in "tonality" due to corruption of digital data. It's like if you drop your hard drive, and your digital copy of Moby Dick starts with "Call me Henry" as a result. It's possible, but far from likely.

[ Parent ]

"Digital" cables (none / 0) (#64)
by ucblockhead on Tue Jul 09, 2002 at 03:42:44 PM EST

It is important to remember that at some level, no cable is truly digital. When digital data is sent over a cable, it is sent as some sort of square wave-form. Something on the other end watches the waveform and calls things above the baseline '1' and things below the baseline '0'. The longer the cable, the more the signal degrades. This is why digital cables have a maximum length. Beyond that, you need some sort of booster that reads the data and then spits it out again in a nice, reformed square waveform.

If the signal degrades enough, it becomes hard to tell the '1's from the '0's, and you get errors. How often this happens depends on the length of the cable, the quality of the cable and the amount of noise in the environment.

Of course, the nice thing about digital communications is that you can detect these errors and do things like resend broken packets and the like. How well this works depends on how good the communication protocol is. What you here as "silence" because of a "line drop" is probably more a factor of the packets being so borked that they never passes checksum and are hence dropped. At a lower level, there is almost certainly data, erroneous data, coming through.
-----------------------
This is k5. We're all tools - duxup
[ Parent ]

No one caught this? (none / 0) (#79)
by pla on Tue Jul 09, 2002 at 06:31:15 PM EST

Four responses, and not one pointed out:

CDs do not store bytes. They store bits.

44100hz at 16 bits per sample, per channel comes out to a bitrate of 1,411,200, without considering the not-inconsiderable redundancy CDs use for error detection and correction.

That doesn't really affect your point, but it changes the numbers somewhat.

More importantly, the redundant information on the disk allows the drive to correct almost all errors. When a CD starts to skip or have pops in the audio, you need to stop using it as a coaster for drinks ;-).


[ Parent ]
Two Things (3.00 / 1) (#33)
by n8f8 on Tue Jul 09, 2002 at 12:08:02 PM EST

First, any review written for a magazine making money from advertising dollars from the same products they are reviewing should be suspect. Second, in my personal opinion, I'd sometimes rather read a review that related performance of an item to somthing tangible rather than strict scientific measures (.i.e. arcane values for THD and decibel measures outside normal hearing ranges.)

Sig: (This will get posted after your comments)
You are a bit boring really (3.00 / 2) (#34)
by drquick on Tue Jul 09, 2002 at 12:10:38 PM EST

Boring, mostly because you're right. Business is make believe.

Being an audiophile is stupid (2.00 / 1) (#37)
by Fon2d2 on Tue Jul 09, 2002 at 12:33:53 PM EST

There's literally no limit to how far it can go. Hundreds, thousands, or hundreds of thousands of dollars. And it all borders on obsessive-compulsiveness. That kind of review over digital interconnects does not surprise me at all. It might be legitimate. It might not. That's the thing. The experience is always so subjective and the details are often dodgy. All you've really managed to do is expose some of the absurdity in the high end audio market. It doesn't mean much besides the fact that some people are crazy about their stereos. Big deal. I already knew that. By the way, 300 pounds for a cable is nothing. Cables can cost in the thousands.

I could say the same about people who buy cars. (none / 0) (#51)
by Farq Q. Fenderson on Tue Jul 09, 2002 at 01:27:07 PM EST

Really, being an audiophile only requires a certain understanding of physics - and far less of it has to do with cables than you'd think.

Consider, for those who have golden ears, they can easily tell a difference that you have no hope of noticing. For those people, having a nice sound system is important. It's also important in studios.

farq will not be coming back
[ Parent ]

Shopping smart (none / 0) (#75)
by dennis on Tue Jul 09, 2002 at 06:00:08 PM EST

for those who have golden ears, they can easily tell a difference that you have no hope of noticing.

If I had golden ears, I'd take the money some people waste on silly cables, and spend it on speakers. The analog parts make a difference, the digital parts don't.

[ Parent ]

Digital speakers? (none / 0) (#78)
by fishpi on Tue Jul 09, 2002 at 06:29:35 PM EST

You have digital cable between your amp and speakers? Wow! Spare a thought for those of us stuck with antiquated analogue speakers.

[ Parent ]
Nope (none / 0) (#80)
by dennis on Tue Jul 09, 2002 at 06:37:29 PM EST

If it's an analog cable, it won't necessarily meet my definition of a "silly" cable. I don't know enough about analog cables to say, I just know that spending money to get higher-fidelity 1s and 0s is silly.

[ Parent ]
OK, misunderstood you (none / 0) (#82)
by fishpi on Tue Jul 09, 2002 at 06:50:03 PM EST

I was just pointing out that without digital speakers there will always be some analogue cabling in a system and this is worth spending a little money on.

[ Parent ]
Yes, of course - for the most part. (none / 0) (#110)
by Farq Q. Fenderson on Wed Jul 10, 2002 at 01:36:30 AM EST

There are some digital components that matter, but cables have nothing to do with it, unles bandwidth somehow becomes an issue.

That, and there's more to a soundsystem than cables and speakers.

farq will not be coming back
[ Parent ]

you give them too much credit (4.00 / 2) (#61)
by tps12 on Tue Jul 09, 2002 at 03:25:18 PM EST

The author's point is that anyone with a basic understanding of electrical engineering can perceive that this review is completely absurd. It makes as much sense as a car magazine publishing a review on what to eat for breakfast to get the best acceleration.

[ Parent ]
VooDoo Audio (5.00 / 3) (#38)
by thelizman on Tue Jul 09, 2002 at 12:35:29 PM EST

I have part of an online car audio terrorist organization for more years than I care to admit. We used to tear it up on rec.audio.car whenever some corporate slug or mind numbed consumer would dare tell us that you had to have type X or brand Y audio interconnects. The Voodoo science they used to justify the cost of things like Kimber cables was laughable to anyone with an iota of knowlede.

I think my favorite was (and still is) "Time-Correct" or "Time Aligned" winding in an RCA cable. The jist was that three wires were used per channel, with the bass wires being shortest and the treble wires being longest. This was supposed to realign the phase of the audio signal which is distorted in travelling down the wire. The problem with this is that the signals on the RCA wires are electrical: The voltage changes at the speed of light, far faster than a mere human can perceive. On the other hand, for a current of electrons to flow from one end of the cable to the other could take days, because electrons actually move quite slowly (depending on the size of the conductor and applied voltage - I know you've all seen those animations with electrons zipping around, it's an exaggeration). And of course, none of these wires ever had a built in crossover newtork to even bother separating the different bands to their respective wires. I believe Monster Cable charged about $15 USD/foot for their time-aligned cables.
--

"Our language is sufficiently clumsy enough to allow us to believe foolish things." - George Orwell
VooDoo Audio (4.00 / 1) (#55)
by bgalehouse on Tue Jul 09, 2002 at 02:12:13 PM EST

Once was buying a fiber interconnect at Circuit City. I figured that the cheapest name brand would be more than good enough to not lose bits. However, I had to ask where they were, as the place wasn't organized super well.

When the sales droned tried to sell me the next version up because it had 'better shielding' I almost threw a fit about high school physics education, but managed to reduce myself to a laugh.

I especially like the gold plated connector shells on fiber cables.

[ Parent ]

"Shielded" Cables (5.00 / 1) (#68)
by ephelon on Tue Jul 09, 2002 at 04:40:16 PM EST

I recently purchased an optical SPDIF cable, and let the salesman talk me up the next (not super bargain basement) model.

He explained that the cheap cable had no real protection on it, and was quite succeptable to being kinked.  (Light only bends so much.)  The one I purchased was still reasonably priced, but was stranded to protect the fibre core of the cable.
-- This is not my home; the cats just let me stay here.
[ Parent ]

Toslink (fibre) cable... (none / 0) (#136)
by Sawzall on Wed Jul 10, 2002 at 04:13:10 PM EST

the cable probably cost three times the transmitting and receiving chips. See a problem there? There is. Until you get into serious money equipment, you are better off staying with the S/PDIF coaxial connection (RCA).

In theory, the optical connection gives you better isolation between components. But until you spend some real money on optical/electical converters, they introduce more noise than they remove.

But its a nice scam to sell some espensive cables and it does not cost much to add them into the equipment.

[ Parent ]

Toslink v. Coax (none / 0) (#158)
by adolf on Mon Jul 22, 2002 at 01:29:39 PM EST

Sure, sure.

But I'd rather have my computer plugged into the stereo with fiber than copper, any day, even if it is a comparatively worse, and more expensive way of doing things.  (I gave up on using infernal sound card DACs a long time ago.)

There's so much noise there, killing my radio reception, and (under some circumstances) making odd noises.  Toslink helps.  It also absolutely removes one avenue of introducing a ground loop.

I did spend more on the cable than the sound card and toslink daughtercard put together, though...  ;)

For coaxial SP/DIF, I go to Radio Shack, and buy their "Video Grade" wire with RCA connectors.  It's incredibly cheap, 75-ohm, and is easily unzipped into two seperate cables.  Or, if I need a special length, plain old RG-59 with F-to-RCA adapters does well at distances of hundreds of feet.

[ Parent ]

A bit of physics. (5.00 / 1) (#102)
by artemb on Wed Jul 10, 2002 at 12:13:13 AM EST

While I agree that for all practical purposes of Hi-Fi audio cable length makes no difference, I cannot agree with your statement that "The voltage changes at the speed of light..".

According to some laws of physics (the part that deals with transmission lines), signal propagation speed depends on line's impedance and generally proportional to (LC)-1/2. For coaxial cables it's roughly 2/3 of the speed of light.

See EE535 or Physics 231 Lab manual [PDF]

These effects, though, have very little effect to the transmission of sound. What can make a difference is a *gross* impedance mismatch, but for sound frequencies you'll have to to make substantial effort to make it happen.

Bottom line - cables/interconnectors do matter, but the effect is generally negligible comparing to other factors (i.e. speakers and/or amp non-linearity).

[ Parent ]

CDs and quality (3.00 / 4) (#43)
by steveftoth on Tue Jul 09, 2002 at 01:04:29 PM EST

Personally I think that the whole cd sounds worse then record (or your favorite source ) argument is moot because CDs are simply the best way we have of distributing music. They are cheap and easy to produce, easy to play, small, have more features then tape, and sound good. The only format that I think is better is solid state, because it doesn't skip, ever. However solid state is expensive, so I don't expect to see it being used any time soon.

Records do sound better if you have a good setup, especially if you are slowing them down (like DJs tend to). However they are a pain to play if all you want to do is listen to the music. For the price of a good record player, you can buy a hundred disc changer that will play your entire collection of CDs.

Once... (none / 0) (#76)
by pla on Tue Jul 09, 2002 at 06:18:07 PM EST

Records do sound better if you have a good setup

I don't know that I agree with that (*I've* never heard a record that sounds better than a CD, but I will accept that they *could*, theoretically). However, the biggest problem arises from how long they will sound better. The very act of playing a record damages it, to some degree. Even on the best equipment possible, after some number of plays (most likely a small number), the quality *must* degrade to below that of a CD. Worse still, it wouldn't tend to degrade gracefully... The way records store information, the high-frequency sound will erode faster than the low frequency.


argument is moot because CDs are simply the best way we have of distributing music

I agree for the most part, with the exception that "best in use" does not equal "best possible". As the most trivial proof, imagine having uncompressed 5:1 sound at 24 bits per sample, 96khz per channel. That would fit on a two-layer DVD (virtually all manufactured DVDs have two layers, while most home burners still only support one layer), and have almost exactly the same amount of music (time-wise) per disk. This would not cost significantly more than producing an ordinary CD (for manufacturers... I will not argue that consumer-oriented blank DVD media cost as much as ten times more than blank CD media, though that *still* comes out to about a buck per disk. *I'd* pay an extra buck for that much of a difference in quality).

With dual-sided media and a modest level of nearly lossless compression (something like 8-bit ADPCM with step tables optimized for each track), you could see complete libraries of music (ala the complete works of Bach, or Pink Floyd, or others of similarly large bodies of material) on a single disk, with much higher quality than similar collection on half a dozen CDs.


[ Parent ]
Complete works of Bach? (4.00 / 1) (#84)
by ghjm on Tue Jul 09, 2002 at 07:33:09 PM EST

I think you must not know just how large the works of Bach actually are. The Bach catalog contains over 1200 works. I estimate that a complete recital of Bach's works would require more than 150 hours.

[ Parent ]
Fair 'nuff (none / 0) (#85)
by pla on Tue Jul 09, 2002 at 07:43:13 PM EST

I estimate that a complete recital of Bach's works would require more than 150 hours

Ah, my bad, indeed. I did not know he had produced *that* much. Impressive! I figured *maybe* two DVDs, but not even close, apparently.

So it goes.


[ Parent ]
Dual-layer (none / 0) (#87)
by fluffy grue on Tue Jul 09, 2002 at 08:00:36 PM EST

Almost every DVD I own is single-layer; the only exception is the Tenchi Muyo OAV boxed set (each of the two OAV series is put on a dual-layer DVD, though the third disc, a multimedia encyclopedia, is only single-layer). Good thing, too, since lots of cheaper players (such as the Sony PS2) can't handle dual-layer discs.
--
I am a calm and tranquil flower.

Try the new Aborted Fœtus McFlurry! Cool and refreshing!
[ Hug Your Trikuare[ Parent ]

Dual layer very common elsewhere (none / 0) (#89)
by cafeman on Tue Jul 09, 2002 at 08:14:06 PM EST

Just about every disc I've got is dual layer (I'm in Australia). We prefer RDSLs, not flippers, and so single-layers are few and far between here. If I wanted to change the disc half-way through a movie, I'd go back to SVCDs.



--------------------
"No Silicon heaven? But where would all the calculators go?"


[ Parent ]
Eh? (none / 0) (#91)
by fluffy grue on Tue Jul 09, 2002 at 08:48:19 PM EST

I've never, ever had to "flip" a DVD, except in the case of episodic series collections, which is a circumstance where you typically don't watch through the whole thing at once anyway. What movies do you watch which are so long that they don't fit on a single DVD layer? (Last I was aware, a DVD layer could hold something like 4 hours of video...)
--
I am a calm and tranquil flower.

Try the new Aborted Fœtus McFlurry! Cool and refreshing!
[ Hug Your Trikuare[ Parent ]

Latest version of Pearl Harbour (none / 0) (#93)
by cafeman on Tue Jul 09, 2002 at 09:05:33 PM EST

Is spread over two discs (due to the DTS sound). I must be out of date - do they break up the extras onto a second disc in the US now? I though they used to break the movie in half ... but that was ages ago. Everything is normally on one disc here, 2 discs if there are heaps of extras or DTS.



--------------------
"No Silicon heaven? But where would all the calculators go?"


[ Parent ]
dts is a different matter (none / 0) (#97)
by fluffy grue on Tue Jul 09, 2002 at 11:10:22 PM EST

Most DVDs are ac-3 (dolby digital). dts wastes a lot of bandwidth, and really doesn't sound any better. Well, the dts tracks are usually mastered better, but that's a similar debate to CD vs. vinyl. :)

Most DVDs I have fit the extras on the same (single-layer) disc as the movie, though a few (like Fight Club) have so much extras that they just put it on a second disc. Apparently it's cheaper to manufacture two discs than a single two-layer disc, and again, many DVD players can't handle two-layer discs. And of course, they can justify charging a lot more for a two-disc set even though the manufacturing costs probably aren't much more.

The only other time I see a movie need more than a single disc is with Warner Bros., how they put both the 4:3 and 16:9 version on the same disc. They use two-sided discs for that, though; 4:3 on one side, 16:9 on the other. Personally, I prefer that they'd just only put on the 16:9 anamorphic (as opposed to letterbox) version with pan-and-scan cues and let the DVD player sort it out; that's how most movies come (in my experience, anyway), and every DVD player I've seen knows how to do The Right Thing (and because the average "What are these black bars?!" idiot doesn't know how to configure their player, every player I've seen defaults to 4:3 pan-and-scan).
--
I am a calm and tranquil flower.

Try the new Aborted Fœtus McFlurry! Cool and refreshing!
[ Hug Your Trikuare[ Parent ]

2 layer discs (none / 0) (#98)
by cafeman on Tue Jul 09, 2002 at 11:19:07 PM EST

Out of curiousity, what players have you seen that don't like RSDL discs? The PS2 is fine with them, I've got quite a few friends who have PS2s and have no problems with any local discs (90% of which are RSDL). The layer change is noticeable, but they don't not play the disc. I know the older Chinese ones didn't like them sometimes, but I haven't seen a player since the beginning of last year that had any problems with dual layer discs. RCE is a different matter ...

I agree with the anamorphic issue - I think they should stop bothering with reformatting discs, leave the damn movie in the state it was filmed (Harry Potter being one interesting exception - what the hell is "original format" in that case?). I've got a 4:3 tv, but I'd rather watch 2.35:1 if that's the way the movie was intended to be seen. If it was Super35 and they don't crop, give it to me in 4:3 (and let the widescreen TV people be damned).



--------------------
"No Silicon heaven? But where would all the calculators go?"


[ Parent ]
Oh, okay (none / 0) (#111)
by fluffy grue on Wed Jul 10, 2002 at 02:04:12 AM EST

From the reviews of the PS2 I'd read, early units had trouble with dual-layer discs. I guess they must have fixed that problem.

A lot of earlier cheap players (especially Toshibas) had trouble with dual-layer discs. I don't know if my current cheap player has a problem with them (I haven't gotten around to trying Tenchi on it yet, and again, that's the only dual-layer disc I have). My previous player (Sony DVP-S550D) had absolutely no trouble with layer changes, at least. It also had some really cool OSD stuff... like, you could turn on a little display showing exactly where the read head was on the disc (both radius and layer), and it also had a cool bitrate graph. Technically I could still use the Sony player, but I got tired of it not being able to play CD-Rs (not that I need to do that very oten), I'm sick of (and trying to move away from) Sony (for mostly pedantic reasons), and my new player is this cute little portable magenta thing, which I just can't pass up! :)
--
[ Hug Your Trikuare ]
[ Parent ]

Sony sucks (none / 0) (#113)
by cafeman on Wed Jul 10, 2002 at 02:43:55 AM EST

Bad company, average products, mucho marketing. They make pretty cool minidisc players though :)

I've got the Pioneer 533k. Asian equivalent of the American slimline, but with Kareoke! Big bonus, really. It's a good machine - does MP3s, CDRs, DVD+R, SVCD, VCD, and a whole bunch of other formats. Was thinking of one of the Chinese ones, but didn't want to take the chance - some are good, some aren't. Pioneer has a good rep, and it was on sale.

Gee ... a magenta DVD player ... wonder why you got that particular colour ;)



--------------------
"No Silicon heaven? But where would all the calculators go?"


[ Parent ]
Lots of really good cheap ones out there (none / 0) (#114)
by fluffy grue on Wed Jul 10, 2002 at 02:53:23 AM EST

Most of the cheap players (including my cute magenta one) are based on various ESS chipsets. I like ESS. They make good stuff, which is both high-quality and mass-produced to the point of utter cheapness. I often say how I've spent hundreds of dollars trying to find the perfect soundcard, and the best one turned out to be a $6 ES1868-based OEM card. This DVD player is also quite good; a few of the cheaper DVDs which are coming out now seem to make it skip now and then, but it seems to be a problem with the crappy optical pickup, not with the decoder chip itself. And it plays MP3s and JPEG photo albums, too! It has one of the better mp3 browsing interfaces I've seen in a hardware player, as well.

It's also pretty low-power (less than 1W, according to its spec sheet). I haven't tried this yet, but it should be fairly simple to attach an external battery pack and use it as a portable player. (The actual platform seems to be designed for this purpose, in fact... it's got great shock absorption.)

This particular player is the Koss "Hip Hues" line, available for $90 at Target. :) There's cheaper players out (based on the same chipset), but they're full stereo component-sized.

The only thing I don't like about these mini-players is they only have TOSlink (no S/PDIF), but I found a place online which sells 6' TOSlink cables for only $4 each. So of course I bought a whole bunch and am going to eventually sell them on eBay at a signifigant markup (since they have the polished spherical lens and everything else that you find in $300 audiophile-grade TOSlink cables... heh)
--
[ Hug Your Trikuare ]
[ Parent ]

TOSLink cables (none / 0) (#116)
by cafeman on Wed Jul 10, 2002 at 02:57:16 AM EST

They may be polished, but do they correct for jitter? ;)



--------------------
"No Silicon heaven? But where would all the calculators go?"


[ Parent ]
Sure! (none / 0) (#120)
by fluffy grue on Wed Jul 10, 2002 at 04:16:13 AM EST

They have a perfect optical matrix with an inverse-transform pseudocrystalline structure so as to minimize backreflections and photon lag. Plus, they have no light leakage, for perfect signal clarity!

(I get a lot of audiophile catalogs ever since I bought a pair of Grados a year ago.)
--
[ Hug Your Trikuare ]
[ Parent ]

Er, do you have something else in mind? (none / 0) (#129)
by pla on Wed Jul 10, 2002 at 01:26:31 PM EST

Most DVDs I have fit the extras on the same (single-layer) disc as the movie, though a few (like Fight Club) have so much extras that they just put it on a second disc. Apparently it's cheaper to manufacture two discs than a single two-layer disc, and again, many DVD players can't handle two-layer discs.

I think you have confused something for dual-layer other than what dual-layer really means. Do you mean double-sided? that differs entirely from dual-layer.

If you have a DVD drive on your PC, put a disk in. Look at the directory of the disk under "/video_ts/". Almost every DVD of a full-length movie (90-120 minutes) I have ever seen takes up around 7.5Gb. For reference, a single-layer DVD holds around 4.5Gb. 4.5Gb, assuming standard MPEG2 encoding, with no audio, will hold exactly two hours of video. The audio streams usually take up almost exactly a gigabyte by themselves, each (and many DVDs have two or three audio tracks).

Basically, you could only realistically fit slightly over an hour of movie+sound on a single-layer DVD.



[ Parent ]
No, I don't (5.00 / 1) (#130)
by fluffy grue on Wed Jul 10, 2002 at 02:26:55 PM EST

I know the difference between dual-layer and dual-sided. What I was saying is that almost all of the DVDs I have which try to put on more than 4.6GB of data use the dual-sided approach, and they only do that to have separate letterbox and 4:3 versions.

I think you're overestimating the amount of bandwidth which MPEG2 takes, not to mention AC3 - AC3 is (according to my Sony player's bitrate display) 384Kbps, which certainly doesn't take a gigabyte for 2 hours! That's more like 360MB.

For example, here is du on the 16:9 version of Dogma (which is a two-sided disc, and which provides two audio tracks and plenty of extras on the same single disc):

76 /cdrom/jacket_p
4256278 /cdrom/video_ts
4256356 /cdrom

Sure looks like less than 4.5GB to me!
--
[ Hug Your Trikuare ]
[ Parent ]

DVD's are a bane to the record industry... (none / 0) (#139)
by steveftoth on Wed Jul 10, 2002 at 04:34:17 PM EST

If dvd's were used to their full potential then it would really screw up the record industry. Either you can put a ton of music on a DVD or you can put slightly more audio at super high quality. It just depends on how high quality you want to go. There is no way that the record industry is going to release lots of DVDs with 5-10 hours of audio on them. It would cost the consumer like a hundred bucks a disc and not many people will pay a hundred bucks a disc.

Let's face it though, what percentage of people have a setup that can actually tell the difference between what you are describing, 24bit 96khz signal and a normal CD? Depending on the MP3 and the playback device it can be hard to tell which is the CD and which is the MP3. Most people do not care. Maybe you do, and you are willing to pay the difference for a good setup.

Personally, the only thing that I've heard that's better then a regular CD is a live performance.

[ Parent ]

What pays for magazines... (5.00 / 5) (#49)
by bobaloo on Tue Jul 09, 2002 at 01:17:22 PM EST

It's not the readers, it's the advertisers. I can recount my first-hand experience. Several years ago when I was advertising in a national magazine, it was made clear that if I spent "X" dollars per month with them they would feature one of my products in their Product Review section. I agreed, and when the time came for the review, I contacted the very well-known and respected reviewer to see about sending him a sample. His response was, "Kid, I'm an old man, I don't have time to screw with all that stuff, just send me some nice pictures and I'll take care of writing the review..."

product (none / 0) (#106)
by tuj on Wed Jul 10, 2002 at 01:00:26 AM EST

could you tell us the product, or type of product, and the mag or genre?



[ Parent ]
Hmmm (4.57 / 7) (#54)
by Phillip Asheo on Tue Jul 09, 2002 at 02:01:24 PM EST

a digital cable has no effect on the sound whatsoever.

It _should_ have no effect on the sound. It may be that it has no objective effect on the sound.

The act of paying $200 for a digital interconnect means the buyer risks severe cognitive dissonance should the sound not improve afterwards. In other words, he has an investment in it sounding better, so it will sound better.

Also, there have been experiments on workers productivity (I forget who did them, skinner ? that time and motion guy ?) which demonstrated that giving people some control over their environment (e.g. a temperature control), or intervening in any way whatsoever, (lowering the light levels, raising the light levels) resulted in improved productivity. Perhaps a similar effect is at work here. The act of tinkering with the hifi improves the subjective experience of listening to it.

Finally, jitter may or may not be an issue here :-)

--
"Never say what you can grunt. Never grunt what you can wink. Never wink what you can nod, never nod what you can shrug, and don't shrug when it ain't necessary"
-Earl Long

i don't know (3.75 / 4) (#63)
by tps12 on Tue Jul 09, 2002 at 03:36:03 PM EST

My own experimental evidence demonstrates the opposite.

I have spent the better part of the last 2 years of my life trying to get my stupid Lunix computer to work. If it's not one thing, it's another. Video card drivers, CD burner, CD reader, sound card, USB...the list goes on. Now I'm considering getting a DVD drive. I must be a masochist.

As a programmer with a degree in Computer Science, having a functional computer would be a badge of professional and educational competence. Not to mention the fact that I'd actually like to be able to use my computer for the kinds of things normal, healthy computer users do.

Despite all of this tinkering, I am probably less productive with my current setup than I was with my old Mac 7100 two years ago.

[ Parent ]

Justification (none / 0) (#154)
by sgp on Thu Jul 11, 2002 at 09:41:11 PM EST

And end-user who only does word-processing on a default install of Win9x with no interesting hardware should have a functional computer.

Maybe this is my way of justifying my own PC, but to be honest, it's limited to 1024x768 by its LCD display, and (at the moment, though I've only just started looking into it) it seems as if the Maestro2E can't actually grab analogue audio input with the Linux driver.

If that turns out to be the case, as a BSc (Hons) Comp Sci, Solaris consultant/admin/programmer, I should be able to fix the driver (yeah, right:) and wear my badge of honour (even link to it!) in the Linux kernel sources.

In reality, I'll just accept that I can't record my old vinyl onto CD, and live with it.

As for productivity - I've spend the whole day installing OpenBSD and FreeBSD onto my home PC. Nothing in my job spec. even mentions BSD (apart from mentioning Solaris, which has strong links). But it's experience I can take somewhere else.

As an IT pro, I don't really care about the state of my own machines, but about the state of my customers' machines. That's what I take pride in, they do exactly what I promised they would. (That's a 2-part thing: (a) only promising what can be delivered, and (b) delivering what was promised). Seeing myself as a customer for a moment, I promised myself a PC I'd have fun with, learn with and tinker with, might not be too stable, but wouldn't lose my data.

Okay, I owe myself a refund on the last one... re-installed my main laptop about a month ago to put Win98 onto it, found the backups I'd taken of the bootblock also included some of the partition table (which had been changed completely).

Still, I get that for free; if I did that for a customer, there would be huge *YOU MUST HAVE A FULL BACKUP OF ALL YOUR DATA* disclaimers in the contract. Did I back everything up when I mucked about with my own PC? Damn.

There are 10 types of people in the world:
Those who understand binary, and those who don't.

[ Parent ]

Research on LPs (none / 0) (#118)
by thebrix on Wed Jul 10, 2002 at 03:44:20 AM EST

A friend is in the same university department as someone who's doing research (a combination of engineering and psychology) on the curious phenomenon of people who believe LP sound is 'better' than CD.

This work has involved many tricks, such as playing increasingly (deliberately) distorted sound in a variety of situations; it turns out that LP lovers prefer sound with anything up to outrageously distorted levels provided that they believe a record player is playing it. I await the paper.

[ Parent ]

What's new about this? (5.00 / 5) (#58)
by jtown@punk.net on Tue Jul 09, 2002 at 03:13:33 PM EST

Hi-Fi magazines have been a joke longer than I've been alive. I just love things like water-jacket interconnects and optical digital cables that "add warmth" to sound. When I was a kid, hi-fi mags were a great source of amusement. Run a line around the edge of your CD with a green marker or glue a stabalizing ring to it to get more bass. Yeah, right.

The dolby-digital/DTS "revolution" has just opened up a whole new market of suckers for these scam artists. My digital audio gets from the DVD player to the decoder using a $0.99 RCA cable. I could go out and get a $999.99 cable and it won't make a bit of difference. My audio works fine. There are no skips or burps. That's all there is to digital data transfer. Either it works or it doesn't. Same with reading a CD back in the 80s. Either your player could read the disc or it couldn't. Drawing a green line or sticking a plastic ring on there did not (COULD NOT) make a bit of difference.

That green CD marker. (none / 0) (#101)
by static on Wed Jul 10, 2002 at 12:09:05 AM EST

Electronics Australia did a cool test with those green CD markers: they theorised that if it makes a difference when you put it all the way around, then it should make a difference half of the time if you put it only half-way around. So they put it halfway around the CD and looked for audio artefacts at some multiple of the CD spin speed.

They found none. None At All! :-)

Wade.


[ Parent ]

Better cable probably gives less errors... (2.33 / 3) (#74)
by cowbutt on Tue Jul 09, 2002 at 05:57:54 PM EST

...which in turn means the error correction facilities in the DAC go unused and the analogue signal decoded is what it's supposed to be.

BUT, I'd be extremely surprised if this was detectable with anything other than a good oscilloscope. I don't know about audiophiles, but I listen to music with my ears, not a 'scope. ;-)

I can absolutely understand the need for good speakers, particularly *big* speakers. My 12" guitar speaker sounds and feels so much better than the old 8" I used to use because it's moving more air around.

Good quality (i.e. undamaged and well-shielded) cable also makes sense. I'd be surprised if anyone can really tell much of a difference between the super-duper expensive stuff and cheaper stuff though. And the double blind tests bear this out.

Good amplifiers have well-specified power supplies and can cope with wide variations in signal loudness and not introduce too much noise to the signal. Likewise for source equipment. But there'll again be limits to what you can detect without getting that 'scope out.

In summary, my priorities in choosing a new hi-fi would be: as large/powerful speakers as I can bear in my living area, an adequately powerful quality amp followed by the cheapest source components that I can't tell from the next model up. Cable similarly; start with the freebies and compare it with the most expensive you can get your hands on (in the UK, Richer Sounds can loan you a case full of cables for a deposit) and use a binary search to find the cheapest that sounds as good as the next one up.

But what do I know - I'm still listening to my music on an Aiwa Midi that's eight years old and cost me 320 in 1994.

Please explain: (none / 0) (#86)
by fluffy grue on Tue Jul 09, 2002 at 07:56:54 PM EST

Why would lack of using error correction lead to a cleaner signal? Isn't the whole point to error correction that it, like, corrects the errors? It'd have to be pretty severely fucked-up to have to drop an entire sample, and at that point you'd probably have catastrophic data loss anyway.
--
I am a calm and tranquil flower.

Try the new Aborted Fœtus McFlurry! Cool and refreshing!
[ Hug Your Trikuare[ Parent ]

Limited number of detects/corrects (none / 0) (#115)
by cowbutt on Wed Jul 10, 2002 at 02:54:02 AM EST

All the error-correcting codes I've studied have limits on how many damaged bits they can detect and how many they can correct (the latter being lower than the former) for a given number of ECC bits. More details here. If there are more damaged bits to correct than the ECC data allows for, some will go uncorrected and the rest will possibly be "corrected" wrongly.

This is also why various CD Audio copy protection schemes such as Cactus result in the sale of sub-standard recordings; the ECC is there to deal with scratches and other mishaps a CD encounters during its life. If the data is purposefully corrupted (as such copy protection schemes do) then scratches will cause the audio data to be uncorrectable sooner than an unprotected CD.

Returning to the original point, having a perfect digital transfer is probably unnecessary as the ECC will usually do a sterling job of cleaning the audio data "enough" for human ears.

[ Parent ]

Except, no (5.00 / 2) (#119)
by fluffy grue on Wed Jul 10, 2002 at 04:13:13 AM EST

When the ECC fails, you get audible pops, clicks, and dropouts. (And yes, I know how ECC works, and I've actually studied the redbook ECC standard a bit. It's quite robust.)

If any bit errors make it into the final decoded stream, you will get much larger issues than "slight distortion." It will not just be a slight change in harmonic distortion; there's a 1/16 chance that it will be a very loud pop at -3dB (0dB being peak; I'm talking signal level, not SPL); so if your amplifier is at 70dB, the pop will be 67dB - hardly "subtle."
--
[ Hug Your Trikuare ]
[ Parent ]

Atrophied error correction! (none / 0) (#137)
by phliar on Wed Jul 10, 2002 at 04:15:11 PM EST

Better cable probably gives less errors which in turn means the error correction facilities in the DAC go unused
I've got it!!

You should use bad cables. You see, if you use a good cable, the error correction machinery in the component is unused, and just like your body (get off that couch and get some exercise!) it will become weak and unhealthy. However the cable can't be too bad since that would just tire out the error correction. What you need is the correct level of badness carefully tailored to gently exercise your error correction without tiring it out, and without the harsh jolts to the joints that other exercise programs... wait, where was I?


Faster, faster, until the thrill of...
[ Parent ]

On digital. (3.33 / 3) (#95)
by mindstrm on Tue Jul 09, 2002 at 09:27:44 PM EST

First... it is possible for digital cables to be better or worse; the codecs used involve error correcting, and some redundancy. A few dropped bits go largely unnoticed, but can and do affect sound.

So a crappy digital cable CAN degrade the sound. But, I mean, I'm talking about a CRAPPY cable. Really crappy.

SO yes the advertising here is mostly bunk.
Other pet peeves of mine:

Some page where some guy was talking about how jitter was such a big deal, and how burning audio cds at different speeds made a huge difference in sound quality, and how even digital players playing of hard drives could make a difference (like, different hard drives changing the sound quality)

Now.. I could *basically* believe most of it could be true, up until he got to the hard drive part.. because I *KNOW* the stuff is buffered in memory first. The platter it's coming from has no effect, sorry.

On Vinyl & Tubes: There is no point in breaking this down. Some people LIKE the sound of vinyl better. That's fine and great, I have no argument with them, whatsoever.
When they try to talk about how it's 'more accurate' and such though.. that's where I draw the line. It's simply NOT.

Some albums sound better when you add some post-processing effects to them. Does that mean the effects processor is more 'accurate'? no.

How good something sounds is subjective.

Now.. if you are that guy who's about to go look up the most expensive tube amps on the market and show me how flat their response curves are in order to show me how 'accurate' a tube amp is... I'll say this. That amp costs twice as much as a solid state one with the SAME CURVE (if not better). I'm not saying you can't make a great tube amp, just that tubes are not inherently more accurate.

The famous Which? review (3.00 / 2) (#117)
by thebrix on Wed Jul 10, 2002 at 03:27:45 AM EST

This reminds me of an incident several years ago (pre-mass Internet so no references I can find) when Which?, a consumer magazine, blind tested CD players ranging up to 1000+. They found worthwhile improvements in sound quality up to 200 but no substantial improvements thereafter; the howl from the 'specialist' industry was deafening :)

I was amused by one magazine this month which spent about six pages on a 'room conversion for home cinema', costing about 10,000, and raved about the 'exquisite quality' of the fittings and fixtures (never mind the electronics). Well, on the evidence of the single reasonably close-up photograph a door was obviously MDF and had the cheapest of hinges ...

How did this turn into such a wankfest? (2.25 / 4) (#122)
by ennui on Wed Jul 10, 2002 at 08:02:57 AM EST

Really, beyond the initial "all (I assume optic) optic cable should result in about the same quality of transmitted sound" there's not much to talk about here. All this rambling about EE credentials, audiophilism, how CDs work and assorted crap, generally accompanied by insults is shameful. You all have the biggest penis, now try to be civil when you condescendingly explain the life cycle of a photon.

"You can get a lot more done with a kind word and a gun, than with a kind word alone." -- Al Capone
Audiophiles (none / 0) (#123)
by Mwongozi on Wed Jul 10, 2002 at 08:53:41 AM EST

Audiophiles who spent massive amounts of money on their setup will leap to any barely-credible (And sometimes not even barely) argument to support their purchase.

And who can blame them? If I'd spent 300 on a one-meter cable, I'd be pissed off to find out that it was a total waste of money too.

Unfortunately, this means that it's very difficult to have a rational discussion with audiophiles, which is a shame, IMHO.

[ Parent ]

Go and sit in a darkened room then. (none / 0) (#124)
by synaesthesia on Wed Jul 10, 2002 at 08:55:39 AM EST

"all (I assume optic) optic cable should result in about the same quality of transmitted sound"

The discussion is over the word "about".



Sausages or cheese?
[ Parent ]
Not Unusual (none / 0) (#128)
by Lagged2Death on Wed Jul 10, 2002 at 11:18:21 AM EST

I've seen other discussions on this and extremely similar topics, and this is the way they've all turned out.

Snide remarks, condescending attitudes, blustering ignorance, and a dearth of hard data are the standard procedure - on both sides of the debate.

Starfish automatically creates colorful abstract art for your PC desktop!
[ Parent ]

My thoughts (3.50 / 2) (#131)
by riceowlguy on Wed Jul 10, 2002 at 02:28:58 PM EST

For a few years I was a pretty hard-core audiophile.  I stopped when I started living off-campus.  I have more important things to spend money on.

More importantly, at this point there is no way I could appreciate any improvements in my system because of the ambient noise levels in my apartment.

I have this homunculus jumping up and down in my head right now saying "of course your average K5er is anti-audiophile.  They do all their music listening in a room where there are three computers, all with 7 cooling fans in them.  They couldn't possibly appreciate decent quality equipment."  Same goes for the people who say there's no difference between CD audio and 192Mbps MP3s.  But then I realize such a comment would be more appropriate for /. and /.ers.

As someone who spent their formative years doing very little with their free time other than classical music endeavors, let me say that I can tell major differences between cheap crappy systems and good expensive ones, as well as between crappy expensive ones and good mid-prices ones, thank you very much.  I know what live classical music is supposed to sound like.  I know that most recordings don't come close to recreating that.  I know that some come much closer than most.

I've never played with digital separates, so I have no personal experience with whether digital cables have any effect on sound quality.  I personally think that one of two scenarios would come into play: either they would have no effect, or the effect would be so subtle that it wouldn't be worth plonking that kind of cash into cables, vs. better speakers etc.

Let's be clear on one thing - most people who say that they can't tell the difference between a $100 CD player and a $5000 one aren't lying.  They just aren't listening in a quiet enough room.  Since the majority of people, audiophiles included, can't dedicate the space, time or money to have a dedicated, custom designed, soundproofed and acoustically-treated listening room, they ARE stupid for spending more than the bare minimum on a stereo.

"That meant spending the night in the living room with Frank watching over me like some kind of Lovecraftian soul-stealing nightmare creature-Azag-Frank the Blind God of Feet, laughing and drooling from his black throne of madness." -TRASG0

We need hard data (none / 0) (#140)
by squigly on Wed Jul 10, 2002 at 04:56:30 PM EST

Surely if one were to sample a CD with digital out, using a sound card with digital input using a poor quality cable, this could be compared with a direct rip from a CD, or even a second sample from the same CD at the same point.  

These two samples can be compared by converting them to raw binary data, and then using diff on them.  If they are exactly the same, then jitter is a myth.  If they differ (and the samples were correctly matched), then the quality of cable does matter.  

Alternatively, simply get two computers, and transmit data between them using the audio cable several hundred times.  If there is barely noticable loss after a single copy, then this should be magnified quite substantially by multiple iterations.

Oscilloscope pics (none / 0) (#144)
by tuj on Thu Jul 11, 2002 at 01:08:15 AM EST

I posted this as a response to another poster, but I thought it was worthy of a root-level comment.  You be the judge.  Its measurements of jitter from cd players and the respective cables.

http://www.jitter.de/english/sound.html




Um, that still doesn't show anything useful (none / 0) (#145)
by fluffy grue on Thu Jul 11, 2002 at 02:19:05 AM EST

All that shows is the supposed jitter on the digital signal. HOWEVER it STILL doesn't show that jitter affects the resulting ANALOG SOUND at all.

Nobody is saying that the jitter in the digital signal doesn't exist. What everyone is saying is that the jitter doesn't make a single iota of difference in the resulting sound. THAT is what I want to see pictures of!
--
[ Hug Your Trikuare ]
[ Parent ]

analogue signal spectrum graphs good enough? (none / 0) (#149)
by tuj on Thu Jul 11, 2002 at 11:53:02 AM EST

I'll refer you to a paper ( http://www.nanophon.com/audio/jitter92.pdf ) for the AES assessing jitter in digital audio equipment.  At the end of the paper are several graphs of Fourier transforms showing the effects of jitter on the signal coming out of a DAC.  Notice the sidebands in the spectrum.

From the paper: "For a delta-sigma DAC any hitter at 150 kHz, for example, may modulate with the shaped modulator noise around that frequency creating modulation products falling in the most critical parts of the audio spectrum."

Most importantly, his conclusion was that of several consumer DAC units, only one behaved propertly when fed with jitter.

Have we now established that jitter does indeed cause distortion to the resulting analogue signal?



[ Parent ]

Thanks for the link (none / 0) (#152)
by fluffy grue on Thu Jul 11, 2002 at 02:14:37 PM EST

Hm. Unfortunately, the pdf itself is formatted oddly... it looks like the text is laid out in landscape mode but the paper is laid out in portrait, so large amounts of the text is cropped off. At least in ghostview. (I don't have acroread installed.)

Ah, it works fine in Preview on my Mac. Yay.

Okay, so I read through it. Interesting stuff. I do concede that it shows that jitter can affect the analogue output of a DAC. However:

  • The testing methodology doesn't appear to reflect a real-world situation. Real-world jitter is supposedly either random or dependent on the signal itself, not an even 2KHz oscillation.
  • It's easy to find resonance in DACs. There were no graphs showing the frequency output when there's no jitter added in - it's lacking a control test (it only shows graphs of signal+jitter and signal+jitter+PLL; there's no graph of just signal). Additionally, each DAC appears to be getting a different amount of induced jitter (some get 2kHz, some get 5kHz).
  • The maximum level of jitter-induced distortion appears to be around -85dB, which is only slightly louder than the sampling error of a CD to begin with (and is still less than the noise caused by the pumping of blood behind the ear), so even if these results can be believed, it still doesn't make a difference!

--
[ Hug Your Trikuare ]
[ Parent ]
more jitter info (none / 0) (#153)
by tuj on Thu Jul 11, 2002 at 04:25:59 PM EST

I hate to seem like I'm just posting links, but these pages explain jitter and its effects better than I.

http://www.audioprecision.com/publications/audiotst/jan96/jan961.shtml

That page shows, among other things, how the sidebands from jitter can be increased in amplitude by various conditions, such as rise/fall time.

http://www.jitter.de/english/sound.html

I posted that before, but its especially relevant because the experiment uses a commercial cd player and then uses an ultra-precision clock injector to create a signal free of sample-jitter.  They compare the audio subjectively and come to a very strong conclusion in favor of the jitter-free signal.  

There are various other reports on the net of similar subjective conclusions regarding jitter.  Whether they are to be believed or not is up to you, however it seems that jitter can indeed distort the analogue signal out of the DAC in ways that should be perceptible.  

Remember, no one said that everyone could hear the effects of jitter, or that the effects were dramatic.  I conjecture that some people, usually with very good equipment, can indeed hear the distortions caused by jitter.  But then again, some people can't tell the difference (one good speakers) between a 128kbps mp3 and a wav file.  

I couldn't find any relatively large double-blind studies of the subjective comparision regarding jitter, so I don't think anyone can say if jitter has an effect on the masses.  The audiophiles say that it does indeed have an effect, which I won't immediately disagree with.  I used to believe that anything over 44.1kHz 16-bit was overkill until I heard 48kHz 24 bit, and then 96kHz on good speakers (but that's another story).  There are quite a few reports of different jitter phenomena ( http://www.stereophile.com/showarchives.cgi?523 ).  BTW: thanks for good discussion.



[ Parent ]

300? (none / 0) (#146)
by dalinian on Thu Jul 11, 2002 at 04:39:27 AM EST

Is 300 a lot for a cable? No, except for maybe a digital cable. But then, some audiophiles are prepared to pay at least five if not ten times more for speaker cables. Yes, that's correct. This one hifi shop is selling, for example, a 3m cable that costs well over five thousand euro, which means over five thousand dollars or probably about three thousand pounds.

Of course you'd have to be insane to pay that kind of money for cables. But many audiophiles are just that - insane. They don't care if all digital cables are the same, because (quoting the page I linked)

"Audiophiles aren't into listening to music, playing it, dancing to it, or any of the things you are supposed to do with it - although oddly enough they also aren't into objective reality, hard facts, critical reasoning, or any of the left brained activities that one would suspect people who can't dance would be interested in."


A digital cable could make a difference... (none / 0) (#159)
by thogard on Sun Aug 04, 2002 at 10:43:56 PM EST

There is the chance that a noisy digital cable will cause the circutry to generate spikes that get feed into the analog section.  I've never seen this happen (outside of a GSM phone) but it could happen.

I've got an old Magnvox (aka Phillips) CD player.  One of the chips has a signal that goes high if an error is detected.  I hooked a LED up to it and put in a scratched disk and as soon as it tries to read teh scratched area, the light goes on.  I have never seen the light go on when the disk wasn't scratched so I'm assuming that the CD player is detecting about zero errors out of the 10 billion or so bits a day times 14 years for an error rate of  0:35,000,000,000,000.  Thats not to bad.  I'm not going to say someone wouldn't notice one bit being swapped around the wrong way but there are better odds of my friends winning the lottery and deciding I need to replace my old CD player.

Casting doubt on hi-fi reviews | 159 comments (154 topical, 5 editorial, 0 hidden)
Display: Sort:

kuro5hin.org

[XML]
All trademarks and copyrights on this page are owned by their respective companies. The Rest 2000 - Present Kuro5hin.org Inc.
See our legalese page for copyright policies. Please also read our Privacy Policy.
Kuro5hin.org is powered by Free Software, including Apache, Perl, and Linux, The Scoop Engine that runs this site is freely available, under the terms of the GPL.
Need some help? Email help@kuro5hin.org.
My heart's the long stairs.

Powered by Scoop create account | help/FAQ | mission | links | search | IRC | YOU choose the stories!