Talk:High-end audio cable

Clock jitter

The article claims:

However, since most modern DACs buffer the signal to some extent and the data recovery clock does not directly drive the DAC clock, data jitter rarely has any effect on the final output as long as it does not cause data corruption

This second part of this sentence is incorrect. The DAC clock has to go at the same rate as the data recovery clock, and so the DAC clock is phase locked to the data clock. The PLL acts as a low-pass filter, potentially changing the spectral profile of the jitter, but it does NOT bring it to a negligible level.

When you're looking at source of distortion for a high quality audio DAC, clock jitter is one of the dominant ones. The right architecture would be that there is a high-quality crystal oscillator at the DAC, and that clock is then transmitted to the source. This would have both measurably better performance (and, most likely, audibly better performance). With the clock at the source, clock quality, clock transmission quality, and clock recovery loop design all have impact on overall performance.

I do not mean to imply that magical fancy cables will have any impact here (beyond shielding and having proper line impedance, not much matters for digital transmission lines). But the argument as stated is incorrect. —Preceding unsigned comment added by 24.128.191.214 (talk) 11:29, 27 April 2011 (UTC)

Empirical Evidence that We Can Hear

I see citations for blind listening tests concluding people are unable to hear the difference between standard speaker wire and high-end speaker wire of the same size. This seems to merit elaboration and more prominent placement. I do not see any experimental evidence that people can hear the difference.

--RedHouse18 21:23, 26 September 2007 (UTC)

There's some info here: ABX test and [trials]. One of the problems with ABX testing of high-end cables is that the testing device itself is interposed between the source and the destination, inviting argument regarding whether or not it masks the high-end cable's best properties. Analog audio ABX test gear includes relays, enclosed metal housings and wire that aren't present in the normal listening setup. The late-lamented QSC ABX Comparator included microprocessor controls and tabulation of results. The complaint heard of audiophiles was that this accretion of EMF fields and multiple additional connections and routings had a detrimental influence on the resultant sound.

It's possible to test a high-end audio system versus the added effect of cheap cables and ADC/DAC converters in the audio path. Just compare a piece of audio vs. that same piece of audio played and recorded carefully with a decent soundcard and an inexpensive cable, and compare them in your audio system. Such a test is setup here[1]. Please don't tell me that there are many other variables involved in the test than the cables. It's true, but despite this AFAIK nobody has been able to tell apart in a blind test in a reliable manner the original, untouched file from the mangled one, presented at mentioned link. --80.24.18.223 (talk) 14:12, 27 May 2008 (UTC)

You wouldn't be able to conduct an ABX test with direct connection of high-end audio cables because there would have to be different source circuitry and different destination characteristics, however small. Two identical amps feeding two identical speakers? There's no such thing. Every component has tiny differences.

So it would be even easier to detect a difference, no?. --80.24.18.223 (talk) 14:12, 27 May 2008 (UTC)

Of course there will be differences. The point is whether you can _hear_ a difference (not measure it). This is not a test which is impossible to set up. — Preceding unsigned comment added by 84.215.72.60 (talk) 15:29, 27 October 2011 (UTC)

I was present at a listening test sponsored by the San Francisco chapter of AES back in 1989 (or so) when it was determined that a room full of audio engineers was unable to hear the difference between simple stranded-copper 12 gauge power cable and megabucks Litz and other boutique cables, including a very fine stranded Monster cable that had recently been released. The speakers were NHT studio monitors; pretty clean transducers sufficient for the test. I believe this test used a set of custom project boxes with high-power-rated toggles to switch back and forth between two kinds of cable at a time. I know of no written record of this AES chapter meeting but it helped form my opinion from then on. Binksternet 22:05, 26 September 2007 (UTC)

I think the difficulties of doing a blind test in realistic listening environments can be overcome. Start with two audio systems that have the same brand/model components but different cables. Play some music, quickly switch the systems (and speaker placement) and then play it again.

Ask listeners to describe differences and preferences. Repeat multiple times to get an average and also to determine how much randomness there is from one evaluation to the next. You might even play the same system twice or three times in a row to see if the listeners think they hear differences where there obviously aren't any.

--RedHouse18 19:13, 27 September 2007 (UTC)

"Common" engineering calculations done with the use of computers since the 1980s can handle foumulae with the terms which become important at frequencies above 100Hz. The formulae for alternating circuits [by Maxwell] have been around since 1890s.

A big problem is describing the circuit, impossible for a nonshielded cable. Another is knowing the alternating current properties of the insulations and of the commonly used braid shields.

It seems to this retired electrical engineer, with 10 years of cable installation and application experience, that the rest of the playback equipment is more important than the cables; especially the position of the listener in a given room, and the physical qualities of that room.

All recordings have had many physical limitations imposed. One can not recover data which has been discarded! The number of inputs [usually microphones], the placement of those inputs with respect to the sound source, the quality of the recording equipment, the skill of the "mixing" of inputs are critical. The person running the mixing board many times puts too much emphasis on one input, eg: the piano, which causes unremoveable distortion from the given recording.

I urge having someone with excellent hearing to judge a room and its associated playback equipment to listen to the playback of some recorded music which that person has heard during a "live" preformance, preferably during the recording of that music. After all, what is desired is having music played back without distortion of the original quality.

Missing from most of these is the fact that the shield should be connected ONLY at the source end to prevent circulating currents in the shields! This fact is well established in circuits involving critical measurements.

I believe this statement should be detailed to avoid confusion. In some cases connecting the shield only the source end can be much worse then connecting both ends, or none.

For unbalanced connections, since the shield also carries signal, both ends *must* be connected. Doing otherwise will be very noisy at best.

http://www.epanorama.net/documents/groundloop/ for more details

Needs clarification + digital

I have a problem with this part taken from the Cable theory section:

In his 2001 master’s thesis, “A New Methodology for Audio Frequency Power Amplifier Testing Based on Psychoacoustic Data that Better Correlates with Sound Quality”, Daniel Cheever points out a number of flaws in conventional testing that indicate the ear is extremely sensitive to tiny signals well above the normally-expected range of human hearing. The high-order harmonics to which he refers exert an influence on our perception of the sound that is vastly disproportionate to their strength, to the point that research from as far back as 1937 indicates that high-order harmonics may actually exceed the subjective effect of their lower-order cousins.

It sounds as if the citation to the thesis is a citation from the thesis itself to earlier (the 1937) research. Surely it would be more helpful to reference the 1937 research?

Since higher-order harmonics can be of much higher frequency and much lower amplitude than normal audio measurements measure, they provide one explanation for the apparent insufficiency of conventional electrical engineering theory in explaining the influence of high-end audio cables. Since this is but one theory, though, it is important that the reader familiarize himself or herself with the basics of audio measurement.

Tests I've seen for harmonic distortion resolve as low as -100 dB, where 0 dB is the test tone intensity. This should be possible with good quality testing equipment. It would be helpful if a figure is given for the amplitudes of these higher order harmonics, and whether or not these amplitudes are audible to the human ear. The 'higher frequency' part also needs clarification. Usually tests are conducted across the full audio bandwidth, 20 Hz to 20 kHz, and with high quality equipment it should be possible to cover an even greater range. Audio equipment is usually limited, either by design or through performance limitations, to roll-off frequencies above 20 kHz. Therefore, the average (analogue) audio cable won't be carrying anything important in the above 20 kHz range.

The article could mention digital audio cables separately, as these carry signals quite different to the normal analogue ones, requiring different designs.

It would be informative to talk about the electronic differences in a signal transmitted by a high end audio cable and a low end one. Any differences (if they exist) would certainly be detectable by machinery even if not by the human ear. I have a basic understanding of analouge signals and how they relate to sound, but this article's treatment of the electronic aspects of signals is superficial. It only talks about the audio. What I would like to see answered is, "How does an analogue signal change as it is transmitted over an audio cable?" I also agree that digital audio cables should be treated separately.

In response to the comments on the masters thesis, I believe that this research is a little cutting edge for wikipedia. I dont believe that wikipedia should present research that the scientific community has not reached at least some consensus on. Maybe the article should simply say, "It is unclear how much slight changes in sounds affect our perceptions of sounds." Marcusyoder 02:23, 23 September 2007 (UTC)

Article Must Move Towards Fact Based Content

Removed some sections that exhibited unsubstantiated bias against high-end audio cables and replaced them with fact based information. Naysayers must provide evidence of their claims if they are to remain in this article. The frequency response plot link that was added is a powerful example of how electrical engineering predicts the changes that cables will make. Capacitance Inductance and Resistance are not mysteries. They provide solid evidence of the SOME of the differences that cables will make. —Preceding unsigned comment added by Apblake (talk • contribs)

Of course those parameters have an effect on the signal. But, a) are those effects relevant at audio frecuencies, levels, and usual runs of cable? I'd say no for most systems, as long as you use an adequate gauge size of even cheap cord. And b) you can buy inexpensive cable with low capacitance, inductance or resistance from known providers that has nothing to envy (but the prize) to high-end cables in this respect.--80.24.18.223 (talk) 13:33, 27 May 2008 (UTC)

"Naysayers must provide evidence of their claims" ? Of course not! Anyone who has the slightest notion of logic knows you can't ask to prove a negative AND it's those who make a claim that should prove it (burden of proof). In the case of audio cables, it's those who state there's a difference to be heard that should come up with proof. Of course nobody has ever succeeded in blind testing. Please don't shift the burden of proof, this is not a religion but engineering. I agree with your statement that the content should be fact based, except you should get your facts straight first. --212.123.26.150 (talk) 13:35, 5 March 2010 (UTC) Maybe you should read this: http://en.wikipedia.org/wiki/Philosophic_burden_of_proof#The_fallacy_of_demanding_negative_proof --212.123.26.150 (talk) 13:41, 5 March 2010 (UTC)

Intro

There are a few important points that should be conveyed in the intro without going into too much detail. For starters, there is controversy over the audibility of the differences that cables have. It is also very important to say that basic electrical parameters such as resistance, capacitance, inductance, and conductance will have an effect on the signal, but that they are by no means the only influence. Many people falsely believe that these factors are the only things that influence the signal when science clearly predicts a variety of other phenomena. It is not appropriate to include a long winded explanation of all of these phenomena in the intro, but it is important to let people know that the problem is more complex than what a simple lumped circuit analysis will tell you. —Preceding unsigned comment added by Apblake (talk • contribs)

The comment above appears to be from a purveyor of seven thousand-dollar audio cables. —Preceding unsigned comment added by 24.6.157.14 (talk) 21:32, 26 October 2007 (UTC)

It would be nice if he could elaborate about these other effects and what is the relative influence of these effects to the audio signal that the cables carry.--80.24.18.223 (talk) 13:38, 27 May 2008 (UTC)

Tribolectic effect? Rofl.

Guys, the triboelectric effect is the effect whereby dissimilar materials in contact will pick up a charge - like wool and amber. But it's static electricity, so it will drain instantly out of any cable that's just sitting there. The only way it's going to make a difference is if A) the cable is so poorly constructed that dissimilar materials are actually rattling around inside it B) the cable is physically being moved around while you're listening, causing these dissimilar materials to come in and out of contact. You would get a pop/crackling noise. Unless I'm missing something really obvious, this is a howlingly ridiculous claim, and I'm rather shocked that it's been in the lede for several months... <eleland/talkedits> 18:59, 15 June 2008 (UTC)

science or art

The article currently mostly speculates, and discusses artist type tests. It would be more useufl to base it on sound engineering principles to show that

1. cable choice does affect sound
2. to an extent that is utterly insignificant, unless extremely long runs are used

86.4.152.167 (talk) 12:21, 26 January 2009 (UTC)

This article is a mess

It looks like any facts have been removed and replaced with "audiophiles claim that high end speaker cables are awesome".

I removed the WSJ article reference by the way because: a) it wasn't double blind, and b) the probability of it occuring by chance is 10%. Not exactly stunning.

Missing section

This page has no section on line-level analogue interconnect. In parts it looks like it's all about speaker cable (which would make it a badly-named article), but it includes sections on digital interconnect and mains cabling. What's it really about? --ToobMug (talk) 02:17, 5 January 2010 (UTC)

HDMI Standard

The article states that HDMI uses error correction. Error correction is only used in the data portion of the HDMI signal. The video and audio portions do not have error correction. You can read through 5.4.4 of the 1.3 spec for video and 7.7 for audio.

http://www.hdmi.org/download/HDMI_Spec_1.3_GM1.pdf —Preceding unsigned comment added by 68.104.183.6 (talk) 04:09, 3 January 2011 (UTC)

I've looked at your linked doc. At section 3 it says

"In order to transmit audio and auxiliary data across the TMDS channels, HDMI uses a packet structure. In order to attain the higher reliability required of audio and control data, this data is protected with a BCH error correction code and is encoded using a special error reduction coding to produce the 10-bit word that is transmitted. "

Then section 5.1.1 says "The packet data consists of audio and auxiliary data and associated error correction codes."

Table 5.1 show that the so-called "data island" contains "Packet Data - Audio Samples - InfoFrames " and has "TERC4 Coding (4 bits converted to 10 bits) " which is an error-correction coding (see glossary in the beginning of the specs).

Section 7.7 says only "The behavior of the Sink after detecting an error is implementation-dependent. However, Sinks should be designed to prevent loud spurious noises from being generated due to errors. Sample repetition and interpolation are well known concealment techniques and are recommended. " Which is not clear for me (but is supposedly so for the experts), however it says there's error detection, not the contrary.

Overall, it seems to me that for the audio part there is error correction, though I'd like to find more concrete statements of it. (Also see http://arstechnica.com/civis/viewtopic.php?f=6&t=275120&p=6271817 )

Hoemaco (talk) 12:01, 2 June 2011 (UTC)

However, that statement

"Cables carrying digital signals, such as S/PDIF and HDMI, are effectively immune to signal degradation for the short lengths used in consumer audio." in itself might not be exactly true. One one hand, it might very well happen that (either in household or office or outdoor use) one might want to use a cable several (or even a few) meters long; or a cable that is not-very-well-made (contact fault etc) or an unshielded cable in an environment with large EM noise (near a transformer or electric machines) and if the SNR goes too low, digital stuff won't help anymore. (See USB - also can't do even a few meters). Hoemaco (talk) 12:06, 2 June 2011 (UTC)