Why 24/192 is a bad idea?

[amirm]

The graph is an illustration of what IMD is, and how it produces both higher and lower frequency products, and that's all that it is. I never implied anywhere that it was anything but.

Why not have a disclaimer above it saying so that the levels you have there are not realistic whatsoever? And not proceed to put test files at the same levels? What if audiophiles created contrived tests like that? You would be up in arms. Yet here we are.

As we shouldn't base the format on assumed equipment limitations, the format should not include ultrasonics in the event they contribute to audible IMD (as they contribute nothing useful anyway).

How many times will you ignore my answer to this? Why not let the listener/customer take out the ultrasonics? He can take out as little or as much as needed to adapt it to his system/ears. Since his system is different than someone else's, how would taking it out at production time accommodate both? And folks like you can continue to get the decimated version as exists today. Everyone wins.

It would only have to occur once, wouldn't it?

Occurring once was enough to base an entire paper to talk people into ultrasonics doing harm? And that being a hypothetical and not one that you have experienced? is this how we write a thesis these days?

And why do you think we can build perfectly functioning systems that reproduce 22Khz but somehow are crippled when it comes to making it work for 33 Khz?

Again, I'd love to hear from folks trying out the IMD test samples I posted in the article on their playback rigs. For convenience sake, they're here:
http://people.xiph.org/~xiphmont/demo/neil-young.html#toc_1ch

I looked at some of them. It is the same misleading point translated into audio files. The two tone test for example seems to be at -12 dbfs. Please tell me why I should care about 33Khz tones a -12 dbfs. And also, prove to me that if I hear something, it is not due to aliasing in my audio pipeline/DAC instead of IMD.

 

[xiphmont]

Normally I say it is good to discuss a topic from educational point of view. But neither one of you work in the recording industry, or in the field of creating and playing digital audio (i.e. DACs, and related signal processing). A better paper was written on the topic long before you all got the idea and does justice to the topic.

Ahem:

Please hold yourself to a higher standard and open yourself to discussion and not confrontation or the provoking of hostility.

Amir, we've consciously refrained from any dick waving whatsoever up to this point. Follow your own rules, or lose all credibility.

Monty
Xiph.Org

 

[xiphmont]

I looked at some of them. It is the same misleading point translated into audio files. The two tone test for example seems to be at -12 dbfs. Please tell me why I should care about 33Khz tones a -12 dbfs.

So you're uninterested. Don't reply then. [You also only looked at one file.]

And also, prove to me that if I hear something, it is not due to aliasing in my audio pipeline/DAC instead of IMD.

Aliasing extending more than Fs/4? Technically, I suppose it's a possibility... do you have any reason to believe it's not vanishingly unlikely? If you do, what the hell, I'll add aliasing test files. Why not?

 

[amirm]

The recording world decided that quite long ago, when there was little conclusive data and 'playing it safe' was defensible. The modern generation working in home studios has not decided on high rate--- though some aspire to it solely out of a cargo-cult mentality and the gobbledygook spouted by audiophiles.

Home studios? You have statistics on how much of the music we consume comes from there and what percentage of them are produced at 16/44.1? I settle for any data. Please show that.

I'm concerned with the very fundamentals of how digital audio works, and undoing some common, persistent myths.

That's cool. Please show proper research and analysis as Bob has done. Don't put up ultrasonics near clipping and say, "look, there is distortion." That is not how to advance the discussion. You may not have noticed but I am far, far closer to your position than the 192/24. But you make me uncomfortable in that type of tactic you are using and lack of proper and deep analysis. Please step back and realize that. If you can't have me be on your side, you have no prayer of having anyone else.

The experimental record disagrees with you. Believing otherwise fervently does not make it true.

The belief I have is in analyzing the problem as Bob has done. We can compute the perceptual levels of the system noise+distortion and compare that to threshold of hearing. I read that you don't know why that works. But if you had read Bob's other research papers, you would have seen that to be proper. Therefore, we have evidence that 16 bit channel even with dither is not transparent to our source relative to what we can hear. Since it doesn't cost anything to go up to 20 bits, I say that is where we need to be. Pretty simple.

I have no interest in your business plan, especially if it benefits even tangentially from misleading people.

I said, "Seeing how we all have systems that can do better, your insistence that we shouldn't fails to make the point on technical and business grounds.". The business is not mine. It is that of music creation and production which is producing high resolution masters which are getting released to consumers. You want to call it mine, fine. But it is the reality of the world and you have provided no reason anyone would or should change that.

The article was title '24-bit / 192kHz music downloads and why they don't make sense', not '24-bit / 192kHz music downloads and why they should be outlawed'.

Why was the article about 24/192? The only public statements from Neil Young/Jobs was along these lines: http://allthingsd.com/20120131/neil-young-and-the-sound-of-music/?mod=googlenews

"It’s not that digital is bad or inferior, it’s that the way it’s being used isn’t doing justice to the art,” Young said. “The MP3 only has 5 percent of the data present in the original recording. … The convenience of the digital age has forced people to choose between quality and convenience, but they shouldn’t have to make that choice.”

There is no reference to 24/192 Khz uncompressed files. He is railing against compressed music distribution. You say you had some conversations in some user group where 24/192 was talked about. That is cause to go and campaign against it when the man's central point in public has been against compressed music at below 16/44.1 quality? Don't you think you are distorting his public views by making it about 24/192?

That's exactly the matter at hand. Not good enough. Transparent.

And you can't show transparency at 16/44.1 unless you can counter Bob's analysis. So far, I have not seen you do that.

Ahem:
Does this standard apply to you as well, Amir, or not? You can't seem to resist constant baiting, and it's tiresome.

We are not similarly situated. You have written an article with a sensational headline. That article, and its author, you, have been put up as a topic to be discussed. So we are discussing it. Your credentials are to be evaluated against other people who have come up with different conclusions than yours. If we were discussing someone else's paper and I started to talk about you personally, then sure you can complain. But here, for good or bad, you are the topic .

And it is not like you are showing due respect. Just in that last post you claimed I didn't know what an interpolator was. It is like saying they don't make hamburgers at Burger King .

 

[xiphmont]

Why not have a disclaimer above it saying so that the levels you have there are not realistic whatsoever? And not proceed to put test files at the same levels? What if audiophiles created contrived tests like that? You would be up in arms. Yet here we are.

Something like taking MRIs of test subjects and drawing conclusions based on the colored patterns? You're right, no one sane would ever present that as evidence of ultrasonic hearing.

I present the IMD test patterns for a particular purpose: people can try them out on their own systems to see if they get a clean slate. Most modern systems do in fact, at least for normal listening levels; that means these folks can try out ultrasonics listening tests knowing that IMD is not contaminating the results.

I'm going for in-home reproducibility so that folks can judge _accurately_ for themselves with all the available information.

How many times will you ignore my answer to this?

I did not ignore you. I disagree. There's a difference.

Why not let the listener/customer take out the ultrasonics? He can take out as little or as much as needed to adapt it to his system/ears. Since his system is different than someone else's, how would taking it out at production time accommodate both? And folks like you can continue to get the decimated version as exists today.

I don't care about your business model. Sell ultrasonics all you want. I care about fighting misinformation. Ultrasonics are useless. We've gone in a circle again.

Occurring once was enough to base an entire paper to talk people into ultrasonics doing harm?

If we could find a single individual out of a million that heard true ultrasonics in recordings (and verify it adequately of course), that one person would be sufficient for me to say "OK, we need the higher sampling rates."

Similarly, if one out of 1000 recordings is hurt by IMD... isn't that enough reason to say 'throw the useless ultrasonics out'?

And why do you think we can build perfectly functioning systems that reproduce 22Khz but somehow are crippled when it comes to making it work for 33 Khz?

No one ever suggested that. You're putting your own fallacies into the argument.

Monty
Xiph.Org

 

[amirm]

Something like taking MRIs of test subjects and drawing conclusions based on the colored patterns? You're right, no one sane would ever present that as evidence of ultrasonic hearing.

So you think you gain credibility by discussing the topic like the other side does? That doesn't sanction their approach in that manner? This is why I don't agree with your stance because you deviated from what is proper analysis and stepped into creating sensational headlines based on highly exaggerated simulations and graphics.

I present the IMD test patterns for a particular purpose: people can try them out on their own systems to see if they get a clean slate.

Please excuse me for not seeing this as you stated here. This is what that section starts as:

"192kHz considered harmful

192kHz digital music files offer no benefits. They're not quite neutral either; practical fidelity is slightly worse. The ultrasonics are a liability during playback."

And in just a few lines, the graph is put forward:

99% of the people reading that graph would have no ideal what those levels are and what the scale means. But they do see all those spikes and run home worried. Precisely what you aimed. And as misleading as it gets. It is not like the graph has 2X the normal levels. or even 10X. It is far, far more than that.

I'm going for in-home reproducibility so that folks can judge _accurately_ for themselves with all the available information.

There is no information. You have have synthetic test files. If this is such an important issue, why not provide real recordings at 192 Khz/24-bit that show the problem but not at 16/44.1? Isn't that your thesis? Have you ever heard such a test case? Ever?

I did not ignore you. I disagree. There's a difference.

Didn't ask you why you ignored me. But why ignoring the answer I give you. So again, why can't I be in control of removing ultrasonics/IMD and you download 16/44.1 as you wish? I think this is probably the tenth time I have asked and no answer has come about.

I don't care about your business model. Sell ultrasonics all you want. I care about fighting misinformation. Ultrasonics are useless. We've gone in a circle again.

Well, as long as you evade the answers given, that is where we find ourselves. Once again, high resolution downloads are not subject to loudness wars. That is what I am selling. Real improvement in audio quality over CD and MP3s. I do not get the feeling that fidelity improvements of that sort matter to you since you reviewed Meyer and Moran paper in your article but completely ignored this important point.

So I am doing the same thing with your report. I want to make sure people don't forget the real benefits of high res audio that are demonstrable and real and undisputed.

If we could find a single individual out of a million that heard true ultrasonics in recordings (and verify it adequately of course), that one person would be sufficient for me to say "OK, we need the higher sampling rates."

Unnecessary as long as we have real benefits below 20 Khz in high resolution downloads.

Similarly, if one out of 1000 recordings is hurt by IMD... isn't that enough reason to say 'throw the useless ultrasonics out'?

No because I can do that selectively. You would be doing that to every title.

No one ever suggested that. You're putting your own fallacies into the argument.

Ah, we are getting some place . So if I can make a system perform as well at 33 Khz as I can at 22 Khz, what is the problem then with higher sampling rate/bandwidth? Surely you are not advocating that I cut out 20 Khz because it can create IMD too. Are you?

 

[xiphmont]

Home studios? You have statistics on how much of the music we consume comes from there and what percentage of them are produced at 16/44.1?

I have only anecdotes from email contacts. It's not data, so I have nothing much to offer here.

That's cool. Please show proper research and analysis as Bob has done.

The fact that we're having this discussion at all indicates I've provided sufficient information to debate. I'll continue adding to it.

Don't put up ultrasonics near clipping and say, "look, there is distortion."

The fact that it is near 0dBFS is irrelevant as digital shows the least noise and distortion near clipping. You're thinking in power-amplifier terms (and I think we can reasonably expect most people to have mastered their volume knob). This is a perfectly valid synthetic test signal for system analysis.

That is not how to advance the discussion. You may not have noticed but I am far, far closer to your position than the 192/24. But you make me uncomfortable in that type of tactic you are using and lack of proper and deep analysis. Please step back and realize that. If you can't have me be on your side, you have no prayer of having anyone else.

And yet that feels more like a threat than any offer of advice, Amir.

The belief I have is in analyzing the problem as Bob has done.

[...]

We can compute the perceptual levels of the system noise+distortion and compare that to threshold of hearing. I read that you don't know why that works.

No, I said that the analysis Bob presented looks correct on its face, but doesn't stand up based only on the details provided.

But if you had read Bob's other research papers, you would have seen that to be proper.

I have read Bob's other research papers and found them... overly tinged with the fantastic. I was worried about linking Coding2 at all because several others are severely problematic (and disagree with Coding2 for that matter). But I judged it was a good piece of writing and presented many of the same concepts very well, even if I disagreed with specifics, and had problems with the other papers.

But, I'll be a good sport, I'll go back and look.

Therefore, we have evidence that 16 bit channel even with dither is not transparent to our source relative to what we can hear.

No, you have a theory that experimentation didn't support.

Why was the article about 24/192? The only public statements from Neil Young/Jobs was along these lines: http://allthingsd.com/20120131/neil-young-and-the-sound-of-music/?mod=googlenews

Fair question.

When Neil's group contacted me, the first thing they wanted to talk about was 24/192. I replied I probably wasn't the person they wanted to talk to, because I'm a skeptic on the subject. They replied that they didn't believe in high-res either, but they couldn't sell to audiophiles unless their product was high res. So then they wanted to know what other hooks could serve the same marketing purpose. We went back and forth on that a bit, also discussed FLAC, and the conversation eventually petered out.

When the Neil / Steve story hit apple insider, wired, etc, a significant amount of the commentary was about 24/192. Most of the justifications for absolutely needing high-res had the basics wrong. This was the real impetus, and I wrote an article. End of story.

Don't you think you are distorting his public views by making it about 24/192?

Not in the slightest.

And you can't show transparency at 16/44.1 unless you can counter Bob's analysis. So far, I have not seen you do that.

Analysis isn't data.

The data doesn't agree with his analysis that slightly more frequency extension is needed.

His analysis of bit depth states that 14 bits shaped is enough; the experimental data is generally on 16 bits shaped, but it shows 16 bits shaped is transparent.

All data agrees that 20 bits undithered is transparent, but I don't see any data demonstrating it's necessary.

We are not similarly situated. You have written an article with a sensational headline. That article, and its author, you, have been put up as a topic to be discussed.

_You_ put me up for discussion. I should think discussing me is off-topic.

So we are discussing it. Your credentials are to be evaluated against other people who have come up with different conclusions than yours.

Oh, so the correctness of my presentation is proportional somehow to my credentials? I assert it is not.

But here, for good or bad, you are the topic .

That's solidly off limits in professional discourse. Knock it off right now.

And it is not like you are showing due respect. Just in that last post you claimed I didn't know what an interpolator was. It is like saying they don't make hamburgers at Burger King .

I withdrew my statement, and said you were right. You're seriously trying to use that against me?

 

[amirm]

No, I said that the analysis Bob presented looks correct on its face, but doesn't stand up based only on the details provided.

Let's have you make that case. Stage is yours.

 

[xiphmont]

So you think you gain credibility by discussing the topic like the other side does?

God, the endless baiting via logical fallacy. I said nothing of the sort.

You suggested audiophiles would be 'up in arms' and I pointed out only that audiophiles have embraced data that makes no sense so long as it supports the audiophile argument.

also: OFF TOPIC

99% of the people reading that graph would have no ideal what those levels are and what the scale means. But they do see all those spikes and run home worried. Precisely what you aimed.

You're ascribing false motivations, and it's also OFF TOPIC. Please stop.

So again, why can't I be in control of removing ultrasonics/IMD and you download 16/44.1 as you wish? I think this is probably the tenth time I have asked and no answer has come about.

Because you are in control and I can't [and wouldn't] change that.

The injection of logical fallacies has become too painful to continue, and it's OFF TOPIC
anyway. if anyone would like to discuss audio, I'm here. I recall Orb and DonH50 had said something interesting about quantization that I wanted to simulate so that I had graphs for later...

 

[arnyk]

There's nothing to PM you about. The files sound the same to me, and for the ones that are different those differences are about 48 dB down. What are these files supposed to show, and how do they prove that bit depth affects more than the noise floor?

My results are the same. I was under the impression that the samples were about reconstruction filter pre/post echo. If they are, they prove my point because they don't show any artifacts as large as what the article their provider cited as being audible, not by a country mile!

 

[arnyk]

If you want to advocate something worse than the above, then it is up to you to prove inaudibility. I am not here to do your homework for you to counter Bob's paper .

Stuart has a long track record of advocacy of technical overkill. He profited from the DVD-A boondoggle by selling licenses for MLP.

Asking for a proof of inaudibility is a demand for a logical conundrum - you are in essence asking for a proof of a negative hypothesis.

I've done my homework and so have others - here are Stuart's demands:

58kHz sampling rate, and
· 14-bit representation with appropriate noise shaping, or
· 20-bit representation in a flat noise floor, i.e. a ‘rectangular’ channel"

IME the only thing on that list that makes good sense is using noise shaping with 14-16 bit sampling.

58 KHz is significant overkill based on listening tests, but it is still way shy of 96 KHz or 192 KHz so thanks for that!

 

[MarinJim]

When Neil's group contacted me, the first thing they wanted to talk about was 24/192. I replied I probably wasn't the person they wanted to talk to, because I'm a skeptic on the subject. They replied that they didn't believe in high-res either, but they couldn't sell to audiophiles unless their product was high res. So then they wanted to know what other hooks could serve the same

I think this is very important, and very relevant to the original topic. Hi-res is new to the average consumer/audiophile and in a society that worships "I have to have that before my neighbor does". Companies are jumping on the hi-res bandwagon not because it clearly offers a supreme, near perfect recorded format and is greatly superior to the CD , but because the average consumer and/or audiophile can't tell the difference, but buy into it because they fear they will not be with the "in crowd". Just saying.

 

[arnyk]

I think this is very important, and very relevant to the original topic. Hi-res is new to the average consumer/audiophile and in a society that worships "I have to have that before my neighbor does". Companies are jumping on the hi-res bandwagon not because it clearly offers a supreme, near perfect recorded format and is greatly superior to the CD , but because the average consumer and/or audiophile can't tell the difference, but buy into it because they fear they will not be with the "in crowd". Just saying.

I think it is also due to the fact that quantitative improvements (word size, sample rate) are easier to understand than qualitative (actual sound quality) improvements.

 

[DonH50]

I had been tossing 'rounding' and 'truncation' into the same boat as they're equivalent save a small DC offset. Both should behave roughly identically with respect to distortion.

Not when I do it, but I would rather not step into this mess. There's a thread somewhere* where I show the difference when I took a 24-bit signal and compared FFTs using rounding and truncation to reduce to 16-bit data. Truncation added harmonic spurs; rounding did not. The spurs were at very low level. Whether you would see spurs with non-ideal signals I am not sure and do not really care. The plots are for ideal DACs without dither (but dither does not reduce distortion).


* http://www.whatsbestforum.com/showt...24-and-32-bit-conversion-LSB-extension-errors and look at post number 6.

 

[arnyk]

Not when I do it, but I would rather not step into this mess. There's a thread somewhere* where I show the difference when I took a 24-bit signal and compared FFTs using rounding and truncation to reduce to 16-bit data. Truncation added harmonic spurs; rounding did not. The spurs were at very low level. Whether you would see spurs with non-ideal signals I am not sure and do not really care. The plots are for ideal DACs without dither (but dither does not reduce distortion).

A DC offset is unconditional and can easily be disqualified as being dither. Rounding is conditional and just may be random and uncorrelated enough to count as dither.

 

[MarinJim]

I think it is also due to the fact that quantitative improvements (word size, sample rate) are easier to understand than qualitative (actual sound quality) improvements.

In simplistic terms, I find the reverse true. It is very easy for me to understand if hi-res sounds "better". All I have to do is listen to it. If I can hear a marked improvement great, if not, then so what. Wait a second, I just spent $24 on a HD Track download and it sounds maybe just a little bit "better" than my 16/44.1 CD? Can I get a refund?

 

[Mike Lavigne]

i've been facinated by this thread from an entertainment perspective; while admitting that much of the technical stuff is over my head.

the question of the thread seems very simple to me. all one needs to do is visit a pro audio studio and hear what various digital bit depths and sample rates do to an analog signal. when the native source is consistent then the result is preditable. i'm not claiming scientific/technical support for my listening perceptions. simply easily heard repeated results.

 

[Lee]

i've been facinated by this thread from an entertainment perspective; while admitting that much of the technical stuff is over my head.

the question of the thread seems very simple to me. all one needs to do is visit a pro audio studio and hear what various digital bit depths and sample rates do to an analog signal. when the native source is consistent then the result is preditable. i'm not claiming scientific/technical support for my listening perceptions. simply easily heard repeated results.

You don't even have to do that Mike. Just get on HDTracks and download files at various bit rates for playback. As long as you have a DAC that doesn't automatically upsample/downsample you can hear the difference for yourself.

We do lots of recordings at 16/44 and 24/176 using a split mic feed into two Sound Devices boxes. The 16/44 doesn't even begin to capture the tonality of a violin or cello or the dimensions of the soundstage.

 

[amirm]

Stuart has a long track record of advocacy of technical overkill.

What track record is that? In this paper, he is attempting to convince people that they just need to go a bit over 44.1Khz sampling. And that with the right signal processing, we don't even need 16 bits of sample resolution! This is not a sign of someone who is trying to sell people "more is better." He goes through extensive analysis to prove his points in this paper.

He profited from the DVD-A boondoggle by selling licenses for MLP.

Actually, the main reason MLP/lossless compression was necessary in DVD-A was needed would go away if people listened to him trying to keep them from going nuts with sampling rate/resolution:

"The author has felt strongly for some time, that we are on the threshold of the most fantastic opportunity
in audio. It comes from two directions. First, psychoacoustic theory and audio engineering may have
progressed to the point where we know how to define a recording system that can be truly transparent as
far as the human listener is concerned. Second, we will soon see the evolution of a high-density audio
format, related to DVD, that has, if it is used wisely, the data capacity to achieve this goal."

Translation: let's analyze what matters in audio and what doesn't. Don't blindly increase specs.

"We would like to approach transparency in each of the measures of audio given earlier. Obviously, we
could ensure transparency by over-engineering each aspect (assuming that we know how to), but this
will increase the data rate of the audio description in the channel."

So clearly he is pleading that we do not over-engineer the format. He continues later:

"I realise that by expressing the requirement of transparent audio transmission – I am nailing a flag to the
mast and lay myself open to all kinds of attack! However, this analysis has been based on the best
understanding to date on this question and we should exceed this requirement only when there is no
detrimental cost to doing so."

Well, the condition he made at the end became true. Storage cost became trivial for digital audio even at highest sampling rates and resolution. The pros led the charge, convincing themselves that they must maintain and carry high sampling rates from start to finish. Now lately, the we have done away with another factor: cost of stamping out physical goods. By moving to digital distribution, content owners have zeroed out the production costs/retailing. They give the bits to an etailer and wait for their checks. So they are not worried anymore whether it costs more to distribute those bits or not. The e-tailer sees this as a differentiation and ability to make money in a field that is essentially profitless for the distributor (think of how many online music e-tailers have survived). And the customer, based on real or perceived value is willing to buy. And based on analysis of said content, we know that it is in most cases better mastered in the clearly audible frequency range and not subjected to loudness compression. So it is all working for the right reasons.

I have been part of trying to bring about transition to digital distribution for more than a decade, for music and movies. I always thought that when the bottom fell out of making money with compressed music and rampant piracy, that distributing high-res audio would be a small but useful niche that should exist. Same frankly for video. There are a few who care about quality to the N'th level beyond the general public and they should form a useful market that should be saved. My head is down these days when I hear a song and go to buy it and that the only version is a compressed MP3 and AAC. I don't like my choices taken away that way.

And for what? Because Monty thinks there is "harm" in this? And to show said harm he uses dual tones at near clipping and not one sample of music to demonstrate the issue? How is he the unbiased reporter in this?

Asking for a proof of inaudibility is a demand for a logical conundrum - you are in essence asking for a proof of a negative hypothesis.

No, it is easily done as Bob did. If I said tape has no noise, you would quote the 70 db signal to noise ratio and say, "here it is." That is what Bob is doing. He is computing the noise floor of the system and contrasting it with our ability to hear it. Can it be off by a bit? Sure. But we better be because when setting a format we don't want to cut it so close for the sake of it. Some headroom is good and "free" these days.

I've done my homework and so have others - here are Stuart's demands:

I don't think people have done their homework sadly. And I am talking about both sides. Not realizing that high-res music is not subject to loudness compression is a super important factor. Yet if you are not looking at this at a comprehensive level, you miss it. Even on technical ground, I recall when you and I discussed Monty's article, you said Bob's paper was marketing material and never published at AES. How could we discuss a topic where *the* seminal paper is not read or known about? Any such challenge must start with where Bob went wrong and whether you have something fresh to write about it.

58kHz sampling rate, and
· 14-bit representation with appropriate noise shaping, or
· 20-bit representation in a flat noise floor, i.e. a ‘rectangular’ channel"

IME the only thing on that list that makes good sense is using noise shaping with 14-16 bit sampling.

58 KHz is significant overkill based on listening tests, but it is still way shy of 96 KHz or 192 KHz so thanks for that!

The use of 58 Khz is not necessary to give you more bandwidth for tones you hear but rather to provide more bandwidth to park the quantization noise and to make it easier to build DACs that have less in-band ripple. It is no like he is saying we should multiply this by 4 as Monty keeps sensationalizing the issue to be. He is not even saying 2X. He is adding about 50% to it. If I were designing something, that much safety margin would be perfectly acceptable if it doesn't increase costs which this does not. It actually saves money since there is less need for oversampling and such then.

Noise-shaping unfortunately has not taken off in creating music. People either truncate the bits or subject it to dither. The concept of noise-shaping is very foreign to people. My copy of Adobe Audition audio editing app for example defaults to no noise-shaping. So the base assumption should be that noise shaping is not used. If it is not used, then 14 to 16 is not enough if we are to demonstrate inaudibility. So the solution is to not put them in the situation to choose. Get the original bits from them and be done with it. As I keep saying, if I then want smaller files, I can create them using noise shaping or whatever.

 

[MarinJim]

You don't even have to do that Mike. Just get on HDTracks and download files at various bit rates for playback. As long as you have a DAC that doesn't automatically upsample/downsample you can hear the difference for yourself.

We do lots of recordings at 16/44 and 24/176 using a split mic feed into two Sound Devices boxes. The 16/44 doesn't even begin to capture the tonality of a violin or cello or the dimensions of the soundstage.

Then you need a better DAC/CDP!