Shouldn't we upgrade the 20-20 audible range ?!

[IPunchCholla]

I don't think it needs to be repeated as it was a Foobar2000 ABX DBT. It's a real, trustworthy result, it's what everyone wants to see.

To answer the question of whether a single individual can hear differences caused by changes above 20khz under the conditions of the test, it answers it, assuming honesty on the test takers part. (It is trivially easy to game foobar by simply doing lots of tests).

These tests are limited strictly to the conditions of the test. Any inference past that is unwarranted and also needs to be tested.

It doesn't necessarily even mean someone is hearing above 20khz even as other mechanisms have not been ruled out.

To answer the question of an expanding the typical audio range (no on has said it is an absolute barrier, just typical) it actually has no bearing. The questions to be answered there is what portion of the music listening population would benefit from it and what would be the cost?

I also don’t understand what you are even arguing about. Is someone out there preventing you from buying or listening to higher bandwidth higher bit depth recordings? Do you want to force everyone to listen to higher bandwidth/bit depth recordings? Or are you just arguing for the sake of argument? Which is functionally trolling even if that is not your intent.

 

[krabapple]

It is one single unverified and un-replicated data point. It doesn't in itself prove much.

Two HA folk -- googlebot and Igor -- claim to have done it, but both have questions as to what the cause is.
Wombat did too, after a fashion.
How one resamples, and how one converts bit depths, can be made to matter. Playback setup too (see discussion of IM on that thread).
Whether they do generally, in the practices of the 'real world', is another thing.

 

[SigbergAudio]

That sounds like cherry picking.

We can't cherry pick when it's only one result.

Why do you think we don't have more proof?

I also wonder why people are so hung up in trying to prove these extremely small differences, why not focus on the more low hanging fruit where it's plain as day for everyone that differences exist?

Power conditioners and cables and DACs for tons of money, dump it all in a room with no acoustic treatment and pair it with speakers that are clearly nowhere close to linear (while at the same time demanding linearity above 20khz). Everyone are free to carry out this hobby as they please of course, but it's somewhat baffling to me.

 

[pablolie]

It is one single unverified and un-replicated data point. It doesn't in itself prove much.

I greatly mistrust any test that lacks control... it's too easy to cheat with many of these experiments.

 

[danadam]

There may be more recent summaries of what's known about this, but this, from Snopes in 2001, still gets referenced a lot. The verdict is "Undetermined'

CD Length​

Were CDs designed to be long enough to hold Beethoven's 9th Symphony?​

FWIW
«... Sony's top-
engineer, Toshitada Doi, who, when I told
him about my doubts about Beethoven,
said, "Of course you are right, but it was a
good story, wasn't it?"»

The Truth About Tape Recorders

I love my R2R deck. It used to belong to my grandfather and finally had it serviced to restore its recording capability last year. It's fun to play with from time to time.

www.audiosciencereview.com

 

[Mort]

I also wonder why people are so hung up in trying to prove these extremely small differences,

Sealioning

 

[welwynnick]

It is one single unverified and un-replicated data point. It doesn't in itself prove much.

Indeed. Your picking (and interpretation) of that single data point and making your case based on it is indeed cherry picking.

We can't cherry pick when it's only one result.

It's a valid and useful test, and trying to brush it aside with no good reasons apart from you don't like the result most certainly is cherry picking, and is very naughty.

I also wonder why people are so hung up in trying to prove these extremely small differences, why not focus on the more low hanging fruit where it's plain as day for everyone that differences exist?

On the other hand, I'm wondering why there's so much vehement opposition to the ideal that CD isn't transparent. The very idea seems quite intolerable to many, as if it was an affront to decent, moral, law-abiding society. There are no babies being killed or pensions being stolen. It's only HiFi. It might actually be a good thing - better understanding of audio and audibility - moving forwards, advancing the state of the art, that sort of thing. I think I know what the underlying reason is, and it's a bit hard to digest, so I'm going to hold fire for a moment.

Power conditioners and cables and DACs for tons of money, dump it all in a room with no acoustic treatment and pair it with speakers that are clearly nowhere close to linear (while at the same time demanding linearity above 20khz). Everyone are free to carry out this hobby as they please of course, but it's somewhat baffling to me.

Confession time. I've got a power conditioner, but it's been gathering dust for a long time. Stupid mistake.
Otherwise I totally agree. Audiofools are idiots. Amir has done a great job on that. Science rules.

 

[SigbergAudio]

It's a valid and useful test, and trying to brush it aside with no good reasons apart from you don't like the result most certainly is cherry picking, and is very naughty.

But that's not what's happening though.

On the other hand, I'm wondering why there's so much vehement opposition to the ideal that CD isn't transparent. The very idea seems quite intolerable to many, as if it was an affront to decent, moral, law-abiding society. There are no babies being killed or pensions being stolen. It's only HiFi. It might actually be a good thing - better understanding of audio and audibility - moving forwards, advancing the state of the art, that sort of thing. I think I know what the underlying reason is, and it's a bit hard to digest, so I'm going to hold fire for a moment.

It's not so much that babies are puppies are killed (thankfully), it's more that we've had the discussion 50 times already for the last twenty years or so. And hi-res has taken money from people for many years now claiming superior sound quality, and it's dangerously close to taking money for high-end cables, since so far no one has proved that anyone can hear the difference between hi-res and CD.

So after a decade or so it gets a bit annoying, that's all.

My only point was that it would be more fruitful to work towards practically better, not just theoretically better.

 

[Julf]

It's a valid and useful test, and trying to brush it aside with no good reasons apart from you don't like the result most certainly is cherry picking, and is very naughty.

If you say so. On the other hand, I would argue that brushing aside hundreds of useful tests and half a century of research might be seen as somewhat naughty too.

On the other hand, I'm wondering why there's so much vehement opposition to the ideal that CD isn't transparent.

I am wondering why there is so much vehement opposition to the ideal that the earth is flat and cows can fly.

The very idea seems quite intolerable to many, as if it was an affront to decent, moral, law-abiding society. There are no babies being killed or pensions being stolen. It's only HiFi. It might actually be a good thing - better understanding of audio and audibility - moving forwards, advancing the state of the art, that sort of thing. I think I know what the underlying reason is, and it's a bit hard to digest, so I'm going to hold fire for a moment.

Can you maybe try to digest the idea that it might simply be because science has concluded that CD as a format is transparent?

Science rules.

Sop why do you act the opposite?

 

[welwynnick]

It's not so much that babies are puppies are killed (thankfully), it's more that we've had the discussion 50 times already for the last twenty years or so. And hi-res has taken money from people for many years now claiming superior sound quality, and it's dangerously close to taking money for high-end cables, since so far no one has proved that anyone can hear the difference between hi-res and CD.

I appreciate your comments. It's clear that the difference between CD and Hi Res is vanishingly small, usually inaudible except to trained ears, and not worth the extra cost, but I don't think it's fair to say that no one has proved that anyone can hear the difference between hi-res and CD. I just gave an example on HA - where they shoot down such discussion even more forcefully than here. There are many others.

 

[danadam]

Just to pick nits, for fun, the lower limit of CD is actually not 0 Hz, but 0.0002 Hz (as the longest waveform a CD can reproduce is 5400 seconds (90 minutes) long.

What if you play it on repeat?
(assuming the repeat is gapless, which I don't know if it is)

 

[SigbergAudio]

I appreciate your comments. It's clear that the difference between CD and Hi Res is vanishingly small, usually inaudible except to trained ears, and not worth the extra cost, but I don't think it's fair to say that no one has proved that anyone can hear the difference between hi-res and CD. I just gave an example on HA - where they shoot down such discussion even more forcefully than here. There are many others.

I was actually not aware of any consistent / repeatable tests proving this, I'd be interested to see some references.

 

[welwynnick]

If you say so. On the other hand, I would argue that brushing aside hundreds of useful tests and half a century of research might be seen as somewhat naughty too.

I have a very specific answer to that.
An ABX test is looking for a difference - a small difference in audio fidelity.
A positive test is good proof that a difference does exist.
However, a negative test doesn't prove a difference doesn't exist, only that none was found.
You can't prove that something doesn't exist.
You can only prove that something (a small audible difference) does exist.
Therefore use the best possible tests to find the smallest difference that can be detected.
The limit of audibility should be set by the best positive result, rather than the worst negative result.

 

[krabapple]

It's a valid and useful test, and trying to brush it aside with no good reasons apart from you don't like the result most certainly is cherry picking, and is very naughty.

Indeed: it shows you the sort of hoops you have to jump through to set up and pass such a test. And how many questions those hoops can raise.

It's a very far cry from simply saying '96/24 sounds better to me than 44/16'. Which many seem to think suffices.

Did you read the whole thread?

On the other hand, I'm wondering why there's so much vehement opposition to the ideal that CD isn't transparent.

'isn't transparent' or 'isn't transparent in rare cases'?

It's transparent to you, of course.

 

[krabapple]

I have a very specific answer to that.
An ABX test is looking for a difference - a small difference in audio fidelity.
A positive test is good proof that a difference does exist.

Be careful what difference you are talking about. And what difference was detected.

And ''can be heard' in condition A does not mean 'is heard' in condition B.

Which is why googlebot was good to note that he couldn't hear the differences between his custom-made samples using either his usual Canton loudspeakers or his Grado headphones.

And again, who is 'setting the limits of audibility' at 20Hz-20kHz? Who are you actually talking about? Digital production -- and consumer product options - moved past 44/16 decades ago . On the production end there were reasons involving ease of filtering, added headroom for live recording, and avoiding quantization artefacts. On the consumer end...well, not as much reason, mainly LOOK AT THE BIGGER NUMBERRZZZZ!.

 

[solderdude]

It's one or the other.
Yes, the title is misleading, and in the ASR thread many objected that it was a test of filters, instead of DACs.

I go along with that, and took it instead "as proof that some people can hear beyond 20kHz".
Since he got a good win rate in a good number of tests, the result has a high degree of confidence.
I don't think it needs to be repeated as it was a Foobar2000 ABX DBT. It's a real, trustworthy result, it's what everyone wants to see.

Its what you want to see and misinterpret.
All that was shown was that a certain small excerpt with a tell was AB'ed successfully and it only differed in treble extension 20kHz to 22kHz and phase.
It was a difference between different reconstruction filters.
It does say nothing about Hi-Res files versus CD quality. It also is not proof of CD quality not being transparent.

Detecting a possible artifact in a recording when using a different filter of a non-hi-res recording is NOT proof that CD quality is not audible transparent or not.

If you want to see another one, here's another Foobar2000 ABX DBT on HydrogenAudio by googlebot, who compared 24/96 and 16/44 tracks. He identified them 16/16, which is high confidence. Two other participants went over 95% confidence.

Are those 44.1/16 down-sampled from the 96/24 files and what re-sampler was used ?
What equipment was used is another question.
Were the files analyzed ?

 

[welwynnick]

googlebot, who did this test and reported it on HA, actually changed two variables, sample rate AND bit depth. Which is one issue. But he knows how to frame claims properly:

"I do not hear a difference over my regular Canton speakers, which top at about 20kHz! Neither over my Grado headphones. Since both act as a mechanical low pass, I think they just cause the same artifacts in the pass band as the digital low-passing."

"Until today I have mostly laughed at high resolution apologists. But that ABX result somewhat changed my perspective. Maybe those, who say that you do need excellent speakers to hear a difference are actually right? On the other hand this is just one very artificial and synthetic sample. The results can very likely not be extrapolated to real world recordings. I don't know. What do you all think?"

It's good to read the entire thread after that.

Which is why googlebot was good to note that he couldn't hear the differences between his custom-made samples using either his usual Canton loudspeakers or his Grado headphones.

I highlighted that particular thread because of the things you quoted.
There was the paradox about not being able to hear above 17kHz, yet only hearing the difference between 16/44 and 24/192 with the 50kHz Elac speakers.
I'm going to make a plausible assumption here, and guess that googlebot established his hearing limit using increasing sinusoidal tones. Anyone would, wouldn't they?

Therefore he can hear above 17kHz - but not simple sine waves - only complex signals.

This is part of what I had in mind way back on page 43:

Therefore when one asks .............."what bandwidth is necessary for transparency?"
the answer is assumed to be ......."what bandwidth is necessary for transparency of sinusoidal signals?"
but I think it should instead be .."what bandwidth is necessary for transparency of signals?"

Incidentally, regarding the TWO changes to the samples, this all suggests that the difference is down to the bandwidth rather than the dynamic range.

 

[krabapple]

I highlighted that particular thread because of the things you quoted.
There was the paradox about not being able to hear above 17kHz, yet only hearing the difference between 16/44 and 24/192 with the 50kHz Elac speakers.
I'm going to make a plausible assumption here, and guess that googlebot established his hearing limit using increasing sinusoidal tones. Anyone would, wouldn't they?

If you want to know the frequency limit of your hearing of pure sinusoidal tones, sure. The idea that they are *less* discriminating as test signals, than 'complex tones' is rather unorthodox, though I can imagine settings where that matters.

Therefore he can hear above 17kHz - but not simple sine waves - only complex signals.

I suggest more investigation would be needed to claim that for sure. Which is the direction the thread was going in.

This is part of what I had in mind way back on page 43:



Incidentally, regarding the TWO changes to the samples, this all suggests that the difference is down to the bandwidth rather than the dynamic range.

It suggests that that at least two experiments need to be done. And of course, there is scientific literature on both. No one would claim that the HA reports constitute full scientific rigor.

Again, more fundamentally, who is 'setting the limits of audibility' at 20Hz-20kHz? Who are you actually talking about? This is the hill you are fighting on so desperately, after all.

 

[welwynnick]

Are those 44.1/16 down-sampled from the 96/24 files and what re-sampler was used ?
What equipment was used is another question.
Were the files analyzed ?

He started with a 24/192 file and used a SourceForge Sound eXchange (SoX) SRC with linear phase filters and sloped TPDF dither to generate the 16/44.1
He seemed to be pretty conscientious about his methodology.
Arnie Kruger, no less, looked at his test and said:

I don't see anything obvioiusly wrong with the samples.

Googlebot said he was a Hi Res sceptic.

 

[krabapple]

Are those 44.1/16 down-sampled from the 96/24 files and what re-sampler was used ?
What equipment was used is another question.
Were the files analyzed ?

Actually it's more complex than that. googlebot's starting sample was 192/24.

The high resolution contestant (24/96) was produced using the following settings:

sox Sample_192000Hz_24bit.wav 24-96000.wav rate -v 96000 dither -s

That is the setting recommended by Sox' developers for SRC with large target rates. Basically linear phase filters and sloped TPDF dither.

For the Redbook contestant, the 24bit/192khz track was first down-sampled to 16bit/441000Hz:

sox Sample_192000Hz_24bit.wav -b 16 16-44100.wav rate -v 44100 dither -s

And then up-sampled to 24bit/96kHz, again at the highest quality settings:

sox 16-44100.wav -b 24 16-44100-24-96000.wav rate -v 96000 dither -s

Everyone discussing that HA thread should read the whole thread, carefully, to the end. And the connected thread, with IgorC's initial report.