advertisement


Blind ABX test shows difference between 44.1 and 88.2

AndyU

pfm Member
According to this AES paper expert listeners can discriminate between files recorded at 88.2kHz and files recorded at 44.1kHz in a blind ABX test. Anyone know anything more about this experiment?

Sixteen expert listeners were asked to compare 3 versions (44.1kHz, 88.2kHz and the 88.2kHz version down-sampled to 44.1kHz) of 5 musical excerpts in a blind ABX task. Overall, participants were able to discriminate between files recorded at 88.2kHz and their 44.1kHz down-sampled version. Furthermore, for the orchestral excerpt, they were able to discriminate between files recorded at 88.2kHz and files recorded at 44.1kHz.
 
It is customary for insecure subjectivists to trawl the web looking for the merest scrap of evidence to support their claims, as a spoiler for having to accept the obvious.

Therefore I shall introduce a spoiler from the other side of the the fence - namely that the files have been down-sampled, by a process we cannot know presumably without either purchasing the AES paper or contacting the authors.

Therefore, sadly, the test must be viewed with suspicion until it can be proven that the downsampling is transparent.
At best this test compares 88.2 with a down-sampled version - it doesn't compare two sampling rates in isolation but also the sample rate reduction process.

Nevertheless it warms the heart to see former sceptics firmly embracing ABX as a valid method for comparison, and holding up the apparent 'results' to support their arguments. This is truly a mark of progress.

You see, nobody can win at this game ;)
Nice try though!
 
Therefore I shall introduce a spoiler from the other side of the the fence - namely that the files have been down-sampled, by a process we cannot know presumably without either purchasing the AES paper or contacting the authors.

Unfortunately every 16bit 44.1khz file you own has been downsampled from the original.

Why not avoid the possibility of corruption?
 
Unfortunately every 16bit 44.1khz file you own has been downsampled from the original.

Why not avoid the possibility of corruption?

That's true Mike, but it's a different argument.
There is the argument which addresses the question of degradation through down-sampling or recording at lower sampling rates, against which is the commercial reality of having to crunch studio output down to fit a domestic format. The latter is done in order to fit the domestic medium . The issue of sound quality is a separate argument. I'm all for extensive testing of these things but certain basic house rules need applying to the test. Otherwise it's all too easy to question or dismiss the results, as here.

FWIW, going by the title of the test, the comparison should have been between the same source signal encoded at 88.2 and 44.1, then compared.
Otherwise the test is comparing sample rates and sample rate convertors - and some are better than others.

In the spirit of not engaging in cyclical arguments I'll leave it there and let the nerds who enjoy these games commence battle ;)
 
Therefore I shall introduce a spoiler from the other side of the the fence - namely that the files have been down-sampled, by a process we cannot know presumably without either purchasing the AES paper or contacting the authors.
If you examine the abstract you find,

AES said:
Overall, participants were able to discriminate between files recorded at 88.2kHz and their 44.1kHz down-sampled version.
So the downsampling is audible, but the native 88k2 and 44k1 recordings were indistinguishable except,

AES said:
Furthermore, for the orchestral excerpt, they were able to discriminate between files recorded at 88.2kHz and files recorded at 44.1kHz.
The problem here is that running an AD at two different rates is a bit like downsampling....

And then we have the uncertainty of a replay chain running at two different rates.

So this paper appears to suggest that for most material 88k2 and 44k1 are equivalently capable, but that downsampling and other DSP may be audible. Which is disappointingly orthodox.

Paul
 
No, the paper does not suggest that at all. You're reading things into it to support your belief it all sounds the same.
 
namely that the files have been down-sampled, by a process we cannot know

I read the paper ages go. It is not fully clear, but it suggests that they used Pyramix for the SRC. At that time, the sample rate convertor of Pyramix was proven not-blameless.


At best this test compares 88.2 with a down-sampled version - it doesn't compare two sampling rates in isolation but also the sample rate reduction process.

Disagree. What they did is the only correct way, only the implementation leaves something to be desired.

All ADC chips sample at a high rate, and perform internal downconversion to the target rate. If you were to compare a recording at 88.2kHz with a recording of the same source and with the same ADC set to 44.1kHz, then you are in fact comparing the potential effects of the sample rate and of the ADC's final downconversion stage.

Now ADC chips are built for economy, and just about all of them use half-band FIR anti-aliasing filters. This means that all of them violate the sampling theorem: the half-band nature, combined with the nature of music, ensures that the band between, say, 20kHz and 22kHz is filled with aliasing components, which upon playback with a standard half-band filtered DAC will guarantee the excitation of this DAC's pre-ringing. The audibility of this can be argued, but what cannot be argued is that this is an unclean, undesirable starting point for comparing just the effects of sampling rate.

Therefore one has to take the 88.2kHz output of the ADC (or better still, 176.4k) and then use best-known software-based practices for getting to 44.1k (88.2k) while totally avoiding any aliasing.

And for replay one then has to convert all back to 88.2kHz, again with best practices (but this is easy), so as to suppress any difference originating from the DAC running in two different modes.


So what we have in this paper is a comparison between 88.2kHz native from the ADC, 44.1k native from the ADC, and 44.1k generated in a DAW, but now we know that:

1) just about all ADC chips are compromised when doing the final conversion to target rate.

2) the DAW they used was very likely compromised when doing the final conversion to target rate. (In fact most 'pro' DAW SRCs at that time ranged between suspect and utter crap.)

So the results of this paper can only be tagged 'inconclusive'. It needn't have been this way, if only the authors would have had a deeper understanding of the mechanics of digital recording and the performance parameters of the gear they employed. Alas.
 
But if you propose that the difference they heard was due to the inadequacies of the downsampling in the kit they used, it is still arguably beneficial for a consumer to have a copy of the 88.2 master, because it won't have been subject to this poor downsampling. If downsampling isn't transparent in practice, an undownsampled master is a worthwhile option for those who might want it. Interesting too that there is a suggestion that orchestral music reveals such differences more readily. (Interesting to me at least, because I mostly listen to classical music).
 
No, the paper does not suggest that at all. You're reading things into it to support your belief it all sounds the same.
Well, unless you can make an argument, I'd say you're making stuff up.

See Werner's post.

Paul
 
Some forum posters have been pointing this out for over a decade. Not all digital processing is without a footprint.

It's been clear to me for a while that the best digital files I have are the ones I record and playback on the same converter at it's native sampling rate.

KISS as the old adage goes.
 
Robert - what really puts things into perspective for me is that in an earlier post you considered 320k Spotify higher-res than vinyl. Seeing as vinyl can sound rather special - this tells me that a good 320k stream is all that's necessary for the "Audiophile" - the rest is just a marketing numbers game.
 
What's supposed to be the point of all this? Both sides just shout and shout, no-one listens. No one agrees. So, again, what's the point?
You might find it just as useful to 'debate' where a circle begins and ends.....
 
Robert - what really puts things into perspective for me is that in an earlier post you considered 320k Spotify higher-res than vinyl. Seeing as vinyl can sound rather special - this tells me that a good 320k stream is all that's necessary for the "Audiophile" - the rest is just a marketing numbers game.

+1

Chris
 
Robert - what really puts things into perspective for me is that in an earlier post you considered 320k Spotify higher-res than vinyl. Seeing as vinyl can sound rather special - this tells me that a good 320k stream is all that's necessary for the "Audiophile" - the rest is just a marketing numbers game.

But is there any correlation between resolution and "sounding rather special"?
 
But is there any correlation between resolution and "sounding rather special"?

The point I'm making is that if all the musical information can be served up using a lower res format like vinyl - then using a good DAC on the end of Spotify (perhaps a Metrum Octave or Gold Note DAC-7) should achieve similar results.
 
Robert - what really puts things into perspective for me is that in an earlier post you considered 320k Spotify higher-res than vinyl. Seeing as vinyl can sound rather special - this tells me that a good 320k stream is all that's necessary for the "Audiophile" - the rest is just a marketing numbers game.
Yup, humans won't hear anything different due to a higher res than 320k, assuming the same master, though for some there are psychological benefits.
 
The point I'm making is that if all the musical information can be served up using a lower res format like vinyl - then using a good DAC on the end of Spotify (perhaps a Metrum Octave or Gold Note DAC-7) should achieve similar results.

Sure, but that is stipulating that resolution is the relevant measure of quality. There could be other reasons why people consider the sound of vinyl "special".

The only way to find out is to record the output of a vinyl record player, encode the result into a mp3, and ABX the result with the original.
 
Some forum posters have been pointing this out for over a decade. Not all digital processing is without a footprint.

It's been clear to me for a while that the best digital files I have are the ones I record and playback on the same converter at it's native sampling rate.

KISS as the old adage goes.

Makes sense, kinda.

Could you explain what you mean a little more? Are you saying that ADCs and DACs have a "native" bitrate and that using 44.1 native DACs will sound best for CD for example? Or am I misunderstanding you?
 


advertisement


Back
Top