Neil Young Announcement - Blu-Ray is the way

[ArnoldLayne]

I'm a layman in this field but looking at Nyquist's formula I still find some questionmarks. Is it really that extendable beyond a perfect signal where the components have different frequencies but the same amplitude and phase? The latter two are actually not present at all in the theorem. So from a cientific point of view it is really nothing more than a simplification when it comes to representing real life music. Correct me if I'm wrong but I have never seen any proof that Nyquist holds when taking into account amplitude and phase. I'm not saying it doesn`t. But there seems to be a couple of big Q.E.D.s missing before assuming that a CD can store whatever a human ear can perceive.

Regarding the first part of the N.Y. archives, to me it seems to be one of those editions where one skips through two thirds of the tracks. I might get it anyway, or even more likely the next part, because I just learned how to rip music from Blu-ray to DVD-Audio. It will be perfect o compile my own "greatest hits". A little like "Decade", which I already have in vinyl and CD already, but now in hi-rez

 

[ssully]

I'm a layman in this field but looking at Nyquist's formula I still find some questionmarks. Is it really that extendable beyond a perfect signal where the components have different frequencies but the same amplitude and phase? The latter two are actually not present at all in the theorem.

The fact is, if Shannon-Nyquist still had such a 'question mark', telecommunications today wouldn't work.

When you sample , you are effectively sampling the amplitude and phase of the signal at a certain rate. And a 'complex' signal is merely a superposition of 'pure' signals, as per Fourier, it does not entail new requirements. You only need 3 nonoverlapping samples, spaced close enough together (the Nyquist sampling rate, see below) to capture and later reconstruct the correct amplitude and phase and frequency of the wave. (Bit depth -- the number of bits per sample -- in PCM audio determines the 'fine detail' of the shape of the wave -- the precision with which the real-world sampler represents each sample)



The Shannon-Nyquist theorem demonstrates the sampling rate needed to accurately sample the signal, but it only 'works' for a BAND-LIMITED signal, that is, one where the are upper (and lower, though that is effectively DC) frequency bounds. And that's fine because our ears are band-limited receptors themselves; they have an upper limit too. So we set the upper bound at the 'limit' of human hearing (20 kHz, though some rare people can actually hear a few kHz above that), apply a filter to perfectly remove all the sound above that, and then we can apply S-N: sampling at just over twice the upper band limit. In theory that perfectly captures the band-limited signal in digital form. In practice it's unwise to shave the sampling rate that close to the Nyquist frequency, due to the aliasing/imaging artifacts I mentioned above. Real-world filters aren't perfect.

 

[ArnoldLayne]

The fact is, if Shannon-Nyquist still had such a 'question mark', telecommunications today wouldn't work.

When you sample , you are effectively sampling the amplitude and phase of the signal at a certain rate. And a 'complex' signal is merely a superposition of 'pure' signals, as per Fourier, it does not entail new requirements ...

I was actually thinking of the Fourier transformation as the week point since a.f.a.i.k. it doesn't handle randomness that well. I'm no expert in telecommunications but I imagine there are certain rules and relationships between frequency, amplitude and/or phase, which are not present in music. Anyway, maybe this thread is not the place to expose my scepticism any further, even if the matter is interesting and I appreciate the clearness of your explanations.

 

[dobyblue]

Neil hated the first CDs. But he has embraced DVD-A and now Blu-Ray. To me, the reason is simple. The higher the sampling rate and bit rate used for a transfer, the closer one comes to the original wave form, correct? If so, the 24 bit, 192 kHz Blu-Ray discs should sound great, even if only in stereo. I think he just wants his fans to get the best possible recording. I'm all for it.

Agreed.

I am pretty sure human interaction with sound goes beyond what our ears are limited to hearing. Aren't there tests which show this to be the case as per brain activity with frequencies well above 20kHz?

 

[ssully]

Agreed.

I am pretty sure human interaction with sound goes beyond what our ears are limited to hearing. Aren't there tests which show this to be the case as per brain activity with frequencies well above 20kHz?

I'm pretty sure that Pyramix marketing graph is meaningless in terms of what's audible and what is delivered at output in practice. It's certainly been debunked before. It's in the same class as graphs showing how square waves look better and better when you raise the sample rate. Except, our ears are bandlimited filters, they don't pass 'square waves' to our brains.

Oohashi et al's peculiar 'hypersonic effect' setup and results , to which you refer, haven't exactly been embraced as definitive either. Perhaps if someone else could replicate them, it would help. (The one attempt I know of, failed.) It's also sketchy to go from an effect on blood flow in the brain, to positing a perceptible audible difference. fMRI researchers always have to be wary of making such cause/effect leaps. Lastly, Oohashi et al's playback setup and tst material was peculiar and specifically purpose-built enough that his results seem unlikely to account for what various 'audiophiles', using all sorts of gear, are reporting.

These caveats are distinct btw from the well-known effect of supplying ultrasonic frequency very near or in contact with the skin -- bone conduction effect. That too seems unlikely to account for the reports based on being seated appreciably away from loudspeakers that probably aren't very good at outputting ultrahigh frequencies anyway, much less from recordings that proabably don't contain much in the first place.
Yet the 'audiophile' raves keep on coming.

The idea that longtime noise-lover Neil Young can sense ultrahigh frequencies, accounting for his preference for HDCD, then DVD-A , then Blu-Ray, is particularly amusing.

 

[ArnoldLayne]

Once upon a time, in a very distant youthhood, I was a guitar player. As such, I was of course aware that tube amps sounded much fatter and warmer than the transistor Yamaha I could afford. One day I came across an article that explained why: tubes and transistors amplify odd and even harmonics in different ways. Aha, one might think, here we have yet another listening experience totally within audible frequencies and no problemo to represent that on a CD. But that is of course not the point I wished to make. Let's take a closer look at the adjective... "warm" it was. It doesn't relate to frequencies, volumes nor phases. Rather it is a word we also use for example to describe temperatures or characterize relations. So it's not only about what we hear but also what we feel, one might think. So then, where goes the limit for music's capability to provoke feelings? Me, myself and maybe some others too say "far beyond mainstream digital formats". There has effectively been little or even no sustainable evidence that we are right. Which on the other hand is not necessarily a solid proof of that we are wrong. Once upon a time, in a very distant medievalhood, there were these two beliefs that the earth was either round or flat. There was also some cientific basis for each statement. Two paradigma, so to say, at their defense. As we all know, it took a third and totally different approach to find the very truth. It might be the same with all these format wars, maybe we should try to look beyond both ABX testing and psychophysics one of these days.

With the above in mind, it is for me not a total surprise that Neil Young chooses to use high-resolution formats. In the need of hearing aid or not, his music often expresses strong opinions and feelings. In other words, he's an artist with a message. Then why not choose a format that maybe, just maybe improves communication with his audience? That is, beyond notes and lyrics, instruments and rythms. Then again, I don't know the guy. Maybe it's just an attempt to draw attention to back-catalog material in times of slumping record sales? Either way, he's also one of the few artists that has brought us fresh stuff too (Greendale) in high resolution. So at least we should recognise his involvement in the field, left with the doubt whether it is despite or because of his age.

 

[ssully]

Would love to know where you got the idea that there was 'scientific evidence' for flat and round earth models, back in the *Middle Ages*. Especially since the ancient Greeks had proved it was round. Flat earth models were not based on scientific thinking, they were based on religion or mythology -- and were already discredited by the Middle Ages.

corroboration for my crazy ideas here:
http://en.wikipedia.org/wiki/Flat_Earth

Anyway: a preponderance of evidence on one side versus another means that both views are not equally likely to be right.

 

[ArnoldLayne]

Would love to know where you got the idea that there was 'scientific evidence' for flat and round earth models, back in the *Middle Ages*. Especially since the ancient Greeks had proved it was round. Flat earth models were not based on scientific thinking, they were based on religion or mythology -- and were already discredited by the Middle Ages.

corroboration for my crazy ideas here:
http://en.wikipedia.org/wiki/Flat_Earth

Anyway: a preponderance of evidence on one side versus another means that both views are not equally likely to be right.

Well, I do not find that article - nor the related one on the myth - particularly illustrative. All in all they do not deny that there still was at least some doubting regarding the shape of the earth in the early medievals. And as for scientific thinking, the very same articles suggest that the "flat earth" idea based on observations (a truly cientific apporach) goes back further than the Flintstones.

Anyway, I intended use the flat/round earth thing merely to point out that science is good but gets better if it's pluralistic. We have been for a decade now defending or attacking the Nyquist theorem via ABX or psychophysics, respectivelly. Very often the debates arise in forums and forum sections dedicated to high-resolution formats, whilst a.f.a.i.k hi-rez fans mostly refrain from commenting on CD, MP3 an Ipods on their home ground. It is quite natural that in a thread like this about a 192/24 edition there will appear people who believe this format is the best (or know it like I do ). IMHO, at least here and now we can afford considering the remote possibility of existing a grain of truth in that. It doesn`t have to go non-scientific. For example, I found the discussion here before on Nyquist more interesting than any return to ABX & psychophysics would ever be for me. I only wish I had the mathematical skills to follow up the hunches I have.

 

[ssully]

Science is 'pluralistic' in that there is no 'Chinese science' versus 'Belgian science' versus 'American science' -- the body of knowledge is accepted and shared globally, because of the way it was arrived at. This is one of the things that makes it distinct from the realm of religious or folkloric 'facts', which is where 'hi rez is better' claims pretty much fall. The other thing is that the knowledge is provisional --it's always *open* to revision.

Someone could revise the current model by providing really, really good evidence that hi rez really matters audibly at the consumer end. Still waiting for that.

 

[dobyblue]

I don't know whether it's because CD needs to use a few of its 16 bits for noise shaping or not because I'm not that technically savvy, but I do find based on years of critical listening that vinyl and SACD and DVD-A are more pleasant and visceral 2-channel listening experiences to me than CD.

I don't think 24/192 is necessary, but I do find based on all my listening experiences that there is something to be said for Bob Stuart's findings and they simply show that CD is not quite "perfect" enough.

CONCLUSIONS
This article has reviewed the issues surrounding the transmission of high-resolution digital audio. It is suggested that a channel that attains audible transparency will be equivalent to a PCM channel that uses:
•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
This conclusion has the following obvious implications:
•The CD channel with 44.1kHz 16-bit coding (even with noise shaping to extend the resolution) is inadequate
•Even 48kHz sampling is not quite high enough
•Sampling at 88.2kHz or 96kHz is too high, and therefore wasteful of data
•The use of sampling rates above 96kHz to convey a wider audio bandwidth cannot currently be
justified
On the assumption that the industry will chose sample-rates based on 44.1kHz or 48kHz (i.e. 88.2kHz and 96kHz), we have looked at options for improving coding efficiency at these rates.
Noise shaping combined with a new pre-/de-emphasis characteristic for 96kHz (88.2kHz) applications can result in an effective addition of between 2 and 7 bits to the channel. In other words, at these sampling rates a 16-bit channel should be sufficient..
This coding scheme compares very well with other methods of reducing the data rate, offering a very low implementation cost, assured transparency and compatibility with existing systems. The author and other members of the ARA strongly urge its standardisation.
The paper discusses a lossless coding scheme that provides significant savings in peak data rate at both 48kHz and 96kHz. The savings made in the high-rate channels are sufficient to allow more than five channels to be carried in a 6.144Mb/s stream and/or to leave room for video on a DVD audio carrier. Masking-based lossy schemes and bitstream coding are rejected on a number of grounds.

What part of Nyquist takes into account the bit depth? If 44.1kHz alone will give you frequency response that equals human hearing, would you conclude that a 1-bit/44.1kHz signal alone would give you all the human ear requires?

 

[ArnoldLayne]

That is actually the beauty of pulse code modulation: frequency and amplitude are separated and dealt with in different ways. From my limited mathematical knowledge, I think critics of the model rather has to address the assumptions behind the Nyqvist theorem. E.g. to what extent it holds for random (and audible) pitch and phase changes. I'm truly interested in information about that, whether or not it supports my preferences towards hi-rez audio.

What lies behind the choice of 58KHz in that article? Is it merely a question of "harmless" filtering or is it about audible content?

 

[D. Lundberg]

I don't know whether it's because CD needs to use a few of its 16 bits for noise shaping or not because I'm not that technically savvy,

Noise shaping is done to lower the noise floor in critical areas. It does not subtract bits from the representation.

but I do find based on years of critical listening that vinyl and SACD and DVD-A are more pleasant and visceral 2-channel listening experiences to me than CD.

That probably has more to do with differences in mastering than format.
With identical mastering CD, SACD and DVD-A will sound pretty much the same (under normal conditions). Vinyl will usually sound a bit different (but still good) due to its technical limitations.

Mastering with no consideration for sound quality and the "loudness war" are the real problems with CD.
The sad thing is that if another format replaces CD as the "mainstream format", it will probably get the same treatment.

What part of Nyquist takes into account the bit depth? If 44.1kHz alone will give you frequency response that equals human hearing, would you conclude that a 1-bit/44.1kHz signal alone would give you all the human ear requires?

A 1 bit/44.1kHz signal will have the same frequency range as a 24bit/44.1kHz signal. The difference is in the dynamic range.

Sample rate -> Frequency range
Bit depth -> Dynamic range

From my limited mathematical knowledge, I think critics of the model rather has to address the assumptions behind the Nyqvist theorem. E.g. to what extent it holds for random (and audible) pitch and phase changes. I'm truly interested in information about that, whether or not it supports my preferences towards hi-rez audio.

Of course it holds for random pitch and phase changes, as long as the signal is band limited and sampled faster than twice its highest frequency.

Read this: http://www.lavryengineering.com/documents/Sampling_Theory.pdf

 

[dobyblue]

Noise shaping is done to lower the noise floor in critical areas. It does not subtract bits from the representation.

So why does Stuart's conclusions list two different bit-depths as recommendations based on whether noise shaping or flat noise floor are used? I don't understand that part.

That probably has more to do with differences in mastering than format.

With identical mastering CD, SACD and DVD-A will sound pretty much the same (under normal conditions). Vinyl will usually sound a bit different (but still good) due to its technical limitations.

Usually when I compare a title it's one that is available in both formats from the same re-mastering, like the REM or Seal CD+DVD-A series and the Depeche Mode hybrid SACD/CD remasters. I still find the SACD/DVD-A formats sound closer to analog, which I think we feel differently about because where you find vinyl can "still sound good" I find it is usually the most enjoyable and realistic performance. The reason I love SACD and DVD-A is because of the multi-channel capabilities...but for 2-channel listening vinyl is the tits and given that there are hundreds of thousands more titles available it's a no brainer for now.

I guess I would need to undergo a controlled dbt to see if indeed it's all in my mind, but until that time I just don't find CD to be transparent to high resolution PCM.

Mastering with no consideration for sound quality and the "loudness war" are the real problems with CD.
The sad thing is that if another format replaces CD as the "mainstream format", it will probably get the same treatment.

I don't believe this, because I think most high resolution masters have been given better treatment since 1999 than CD has. If they cock up the digital copy, that's okay with me.

What lies behind the choice of 58KHz in that article? Is it merely a question of "harmless" filtering or is it about audible content?

I believe it's audible content...and because 58kHz isn't a reality you have to go above it.

The full white paper is on Meridian's website here - http://www.meridian-audio.com/w_paper/Coding2.PDF

 

[D. Lundberg]

So why does Stuart's conclusions list two different bit-depths as recommendations based on whether noise shaping or flat noise floor are used? I don't understand that part.

It's because you can get almost equivalent performance in the critical areas with noise shaping and fewer bits. 14 or 16 bits with appropriate noise shaped dither is pretty much equivalent to 20 bits in practice.
Not that 16 bits is horrible without noise shaping. There are very few circumstances where you need more than 16 bits for playback.

Consider that the dynamic range of vinyl (or tape) would be equivalent to somewhere between 8 and 12 bits in digital audio.

Usually when I compare a title it's one that is available in both formats from the same re-mastering, like the REM or Seal CD+DVD-A series and the Depeche Mode hybrid SACD/CD remasters.

The problem is that the mastering is almost always different between the formats, even between the layers of hybrid discs.
And sometimes the CD-layer of a remastered disc isn't remastered at all, but taken from an earlier release.

There are also cases where the CD-layer is intentionally worse mastered than the SACD-layer!

I still find the SACD/DVD-A formats sound closer to analog, which I think we feel differently about because where you find vinyl can "still sound good" I find it is usually the most enjoyable and realistic performance.

I like vinyl a lot, but I'm also aware of the technical limitations of the format. And it could actually be argued that it is the flaws in the format that makes it so enjoyable to listen to!
Like the high amounts of background noise (providing an ambiance similar to real rooms) and even order harmonic distortion (providing "warmth"), lack of channel separation (providing a artificially wider sound stage), dynamic compression (providing the appearance of more detail), etc.

From a "Hi-Fi" or technical perspective it is obviously inferior to the digital formats, but that doesn't mean it's less enjoyable to listen to.

I guess I would need to undergo a controlled dbt to see if indeed it's all in my mind, but until that time I just don't find CD to be transparent to high resolution PCM.

I don't think it's all in your mind, but I don't think the differences you hear are necessarily caused by the formats either. There are so many variables (mastering, levels, hardware, etc.) you need to eliminate to make a fair comparison that very few people have done it.

And in the few documented comparisons done so far (that I know of at least) no person has been able to tell the formats apart (under normal listening conditions).

Like in this one: http://www.bostonaudiosociety.org/explanation.htm

Multi-channel audio and (often) superior mastering are very good reasons for favoring DVD-A, SACD and BD-A over CD.
There may also be other reasons, but so far there is very little evidence for them.

I don't believe this, because I think most high resolution masters have been given better treatment since 1999 than CD has. If they cock up the digital copy, that's okay with me.

But "Hi-Res" audio has been a niche market mostly targeting audiophiles so far.
What do you think will happen when people (artists, producers, etc.) whose only interest as far as audio quality is concerned is to make it as loud as possible start using a new format?

And there are actually quite a few cases of "loudness mastering" on DVD-A and SACD.

 

[Guy Robinson]

My major reason for going with SACD or DVD-A or Blu-Ray is the surround aspect. They are all better than lossy DTS or Dolby.

 

[ArnoldLayne]

...

Of course it holds for random pitch and phase changes, as long as the signal is band limited and sampled faster than twice its highest frequency.

Read this: http://www.lavryengineering.com/documents/Sampling_Theory.pdf

Thanks for the link, a very interesting article. But I'm afraid it doesn't address the thing I have in mind. From my limited knowledge, I will try to express myself more clearly:

Firstly, let's establish that sampling is fixed in time, e.g. it happens at t1, t2, t3 etc. whilst analog sound is not. That is, the peaks of the waves may or may not coincide with t1, t2, t3 etc.

We begin studying a sine wave that for convenience 1) has a frequency equal to half the sampling frequency; 2) an amplitude of 1; 3) its peaks coincides with the exact moments of sampling t1, t2, t3 etc. I have no difficulty to understand that it is possible to reconstruct the wave from the samples, as we have identified amplitude and frequency and also know it is a sine wave.

Then we move on to another case using the same wave as before but altering point 3) so that the peaks don't coincide with the exact moments of sampling. Let's say that the samples measure the wave when it reaches 80% of its amplitude. At t1, t2, t3 etc we obtain samples 0.8, -0.8, 0.8 etc. As far as I can see it is no longer possible to reconstruct the wave. Frequency is still correct but amplitude will be underestimated.

I guess (and hope) this is somehow taken into account in the signal theory used when developing our consumer electronics. For example, would 44.1, 48 or 96KHz PCM gives us sufficient sample points to estimate the true amplitude of a 20KHz wave even if its maximum value is not present among the samples? If so, what are the maths behind?

Many thanks in advance
AL

 

[ssully]

I don't know whether it's because CD needs to use a few of its 16 bits for noise shaping or not because I'm not that technically savvy, but I do find based on years of critical listening that vinyl and SACD and DVD-A are more pleasant and visceral 2-channel listening experiences to me than CD.

I don't think 24/192 is necessary, but I do find based on all my listening experiences that there is something to be said for Bob Stuart's findings and they simply show that CD is not quite "perfect" enough.


What part of Nyquist takes into account the bit depth? If 44.1kHz alone will give you frequency response that equals human hearing, would you conclude that a 1-bit/44.1kHz signal alone would give you all the human ear requires?

Bob Stuart (CEO of Meridian....which makes and markets hi-rez technology) alas has never presented any lsitening test data to demonstrate the theoretical 'inadequacy' of Redbook translates to audible inadequacy. And I've read pretty much every one of his hi-rez advocacy papers.

One can create pathological conditions where the limits of Redbook are audible -- raising the playback level very high while listening to 'digital silence' for example -- but for normal listening, there is simply no good evidence that Redbook isn't 'enough' for home audio.

Now, if the play back chain actually delivers, say, 14bit resolution versus 16 bit, then one could find problems more easily. But that's because of the recording, the CD player, or the downstream gear, rather than Redbook (16./44.1) itself. In which case the 'advantage' of hirez is not inherent, it's simple an implementation advantages...it is EASIER to implement hi-rez 'perfectly' than Redbook.

Your question about 1 bit suggests you need to study up on how digital is actually delivered these days.

 

[ssully]

That is actually the beauty of pulse code modulation: frequency and amplitude are separated and dealt with in different ways. From my limited mathematical knowledge, I think critics of the model rather has to address the assumptions behind the Nyqvist theorem. E.g. to what extent it holds for random (and audible) pitch and phase changes.

Nope. Nyquist is not in question, sorry. Seriously, if your doubts were credible and it didn't 'work' for 'random pitch and phase changes', then digital telecommunications wouldn't have worked for the last few decades.

I'm truly interested in information about that, whether or not it supports my preferences towards hi-rez audio.

The bulk of evidence to date suggests that preference for 'hi rez; consumer productgs actually boils down to difference in mastering (which is a *gross* difference, compared to format difference) or placebo-like effects. Would you be willing to accept that explanation?

When fair blind comparisons are done, the 'big differences' seem to shrink radically.

What lies behind the choice of 58KHz in that article? Is it merely a question of "harmless" filtering or is it about audible content?

58 kHz accounts for all potential (but real) implementation issues involved in achieving transparent filtering, as well as theoretical 'need' for more bandwidth. See high end DAC designer Dan Lavry's paper for more corroboration:

http://www.lavryengineering.com/documents/Sampling_Theory.pdf

"Nyquist pointed out that the sampling rate needs only to exceed twice the signal bandwidth.What is the audio bandwidth? Research shows that musical instruments may produce energy above 20 KHz, but there is little sound energy at above 40KHz. Most microphones do not pick up sound at much over 20KHz. Human hearing rarely exceeds 20KHz, and certainly does not reach 40KHz. The above suggests that 88.2 or 96KHz would be overkill. In fact all the objections regarding audio sampling at 44.1KHz, (including the arguments relating to preringing of an FIR filter) are long gone by increasing sampling to about 60KHz."

 

[ssully]

I guess I would need to undergo a controlled dbt to see if indeed it's all in my mind,

Yes, you most certainly would.

 

[dobyblue]

Yes, you most certainly would.

...and have you?

You question Stuart's findings but within the same thread post Lavry's findings which also agree that 58kHz would be more desirable for transparency than 44.1kHz? Or are you saying that Lavry's findings which corroborate Stuart's conclusions are also not supported by listening tests outside of the pathological conditions you've described? Which listening tests specifically?