When is hi-rez overkill?

[ssully]

An ABX with Foobar on my PC would actually reveal significant differences between CD and Hi-res at a very high level of confidence This is because depending on the sample rate, in my setup there are different and very noticable colorings of the sound. It's a basic system: Realtek AL888 onboard sound card and cheap Logitech 2.1 speakers.

Even assuming these reportedly audible difference really WOULD hold up under blind conditions (and it's not really sufficient to just claim they must; measurment evidence or blind testing would still be needed to support your claim) it is just as likely evidence of coloration in your setup, as for inherent audible difference between redbook.

Despite turning off all DSP effects and such, there's still some manipulation that makes the test useless in this particular case, on my gear that is.

Can you describe what happens, in more detail, at various SRs? The laptop I'm typing this on has an AL888 series integrated card on its motherboard. I use it to output digital audio at various sample rates via S/PDIF. There's no obvious coloration at different rates, but I would not expect there to be since it's just outputting data. Your setup, I gather, involves D/A conversion in your computer for analog output to your Logitechs.

Which illustrates what I'm trying to explain about both (ABX/DBT and "home listening") methods: 1) There are hidden factors that can affect any listening test. 2) Lack of predefined hypothesis makes it more difficult to identify and eliminate the relevant factors. This is not only common sense, it is also common scientific methodology sense. But please bear in mind that it's not an argument against ABX tests itself but rather an observation on how they should be conducted.

The ABX only tells you if A and B sound different. It doesn't tell you WHY, nor is it meant to. And again, why do you say there is no 'predefined hypothesis' in an ABX when there is clearly a null hypothesis?

Regarding the theoretical framework, I truly appreciate your efforts to explain the different aspects of applied signal theory. But I'm afraid I still have this hunch that you may be missing a couple of points. Firstly, the attempts to increase resolution and timing with for example dither seems to me very much like substituting the unpredictable and not possible to sample with something predictable and possible to handle within the model.

Sorry, this bizarre view of dither just shows you need to better acquaint yourself with digital audio theory.

Secondly, I suspect that the Nyquist theorem holds if and only if the time function is invariable (same frequency content and amplitudes) during some time interval (I've even seen some mentionings about this in Wikipedia for the underlying maths, although I cannot remember where).

See above.

Putting these two things together, I end up with the feeling that in order to make digital sampling work we filter away the finest and unforeseeable details in the sound. In other words, we assume that music only contains what can be handled with our samples and then filter away whatever content that could eventually be beyond it's scope.

This is an overelaborate way of saying the Nyquist works because by definition it requires removing out-of-band signals.

Whether or not this hypothetic content is relevant or not for the listener is naturally a different story. As I said from the beginning, this part is about the theoretical framework.

Which you seem to be unclear about. I can suggest two good books to start you off -- Nika Aldrich's 'Digital Audio Explained' and Ken Pohlmann's massive 'Principles of Digital Audio'

I'm of course not expecting to convince anyone about anything with this layman's talk of mine. As a matter of fact, I would be happy enough if I have made you understand my point of view on the Shannon-Nyqvist based signal theory.

And you need to understand that Shannon-Nyquist is probably one of the most well-tested theorems in existence.

As with the ABX method above, I'm not at all questioning the theorem itself. It's merely about some of the conclusions we draw from it.

Yes, if a difference is or is not revealed by an ABX, we should be ready to offer plausible reasons why. This is standard scientific practice too.
The test itself does not tell you the reason. That usually requires other tests, like measuring differences in the outputs of the devices used in the test.

 

[ssully]

Well you should consider the entire chain and measure it thoroughly for this type of test. The closest I have come to a simple conclusion to this type of problem is to basically measure the DAC/ADC loopback on the soundcard I am using. Measure it in every possible setting you can - most likely you will want to stick with ASIO and avoid direct sound.

Or WASAPI, if your OS is Windows Vista.

 

[ssully]

Do you think that any of your questions could be satisfactory answered here?

If ArnoldLayne wants his technical answers in a forum, versus from a book, I'd say he should post his ideas to www.hydrogenaudio.org. Similar 'skeptical' views about Shannon-Nyquist have been posted there before and dealt with in some depth, by members with expertise in digital signal processing.

 

[ArnoldLayne]

Even assuming these reportedly audible difference really WOULD hold up under blind conditions (and it's not really sufficient to just claim they must; measurment evidence or blind testing would still be needed to support your claim) it is just as likely evidence of coloration in your setup, as for inherent audible difference between redbook.

...

As a matter of fact, I was trying to explain that the coloring would invalidate any positive test result. Just as unexpected downsampling could invalidate a negative result. My point was that it can be difficult to identify some of these factors. And yes, D/A conversion together with uncontrolled up- or downsampling are possible explanations for what I hear on my PC.

The ABX only tells you if A and B sound different. It doesn't tell you WHY, nor is it meant to. And again, why do you say there is no 'predefined hypothesis' in an ABX when there is clearly a null hypothesis?

Read me again, please. I'm discussing the quality of the hypothesis and it's connection to theory.

Which you seem to be unclear about. I can suggest two good books to start you off -- Nika Aldrich's 'Digital Audio Explained' and Ken Pohlmann's massive 'Principles of Digital Audio'

I have skimmed through the available parts of both these in Google books. I haven't found any sign of that they treat the case when the sampled function's characteristics are not invariable for a relevant timespan. Are you saying that they actually discuss this case? Do they include information on for example transients and wavelets? If you confirm this, I'll place my order inmediately.

Cheers
AL

 

[ssully]

Perhaps that's because 'transients' and 'wavelets' and 'filtering away the finest and most unforseeable details' aren't really the problems for Shannon-Nyquist that you imagine them to be, requiring separate treatment?

Anyway, if you're not willing to actually read a book about how digital audio works, then I would suggest again posting your theories to hydrogenaudio.org, for free peer review. Immediately.

 

[simonwagstaff]

While you guys have been reading books and all that stuff I have been downloading a new torrent of the Dead 4/28/71 from 1st gen reels. Just sort of a warm up, there will be direct transfers of most of the run coming on line in the next few days. sound? Fantastic, very 3-d, very smooth midrange. I have been waiting for a while for some nice 24 bit transfers to come up from the master reels and it looks like it is starting to happen. I am happy to sit and listen to them and appreciate the sound, regardless of whether you think the format adds anything or not. What is nice is that a DVD will hold 2:20 of music in 24/96. I do not believe that hi rez is overkill for 7.5 ips reel to reel sources.

"If you get confused, listen to the music play"

 

[G E]

This thread has a life of its own... thought this one had been put to bed.... Measurements and theories!

If a beautiful woman appeared in your bedroom would you take her measurements or would you smile and enjoy the music?

 

[ssully]

While you guys have been reading books and all that stuff I have been downloading a new torrent of the Dead 4/28/71 from 1st gen reels. Just sort of a warm up, there will be direct transfers of most of the run coming on line in the next few days. sound? Fantastic, very 3-d, very smooth midrange. I have been waiting for a while for some nice 24 bit transfers to come up from the master reels and it looks like it is starting to happen. I am happy to sit and listen to them and appreciate the sound, regardless of whether you think the format adds anything or not. What is nice is that a DVD will hold 2:20 of music in 24/96. I do not believe that hi rez is overkill for 7.5 ips reel to reel sources.

"If you get confused, listen to the music play"

And while you're doing that, I'm playing the music myself, on a real instrument. PWNED.

 

[ssully]

This thread has a life of its own... thought this one had been put to bed.... Measurements and theories!

If a beautiful woman appeared in your bedroom would you take her measurements or would you smile and enjoy the music?

Audio isn't even nearly *sex*, for me. Sorry if it is, for you.

But even then, are you saying no one should ever ask 'why' or 'how' sex works the way it does?

 

[daved64]

This thread has a life of its own... thought this one had been put to bed.... Measurements and theories!

If a beautiful woman appeared in your bedroom would you take her measurements or would you smile and enjoy the music?

This thread is here so that those that wish to talk about the science behind the sound are free to do so, or to try and debunk myths etc....and to keep other threads from derailing. Feel free to ignore it if it's not up your alley.

 

[BananaSlug]

This thread is here so that those that wish to talk about the science behind the sound are free to do so, or to try and debunk myths etc....and to keep other threads from derailing. Feel free to ignore it if it's not up your alley.

does that mean we can, or can't derail this thread towards our ultimate goal: If we could write a song that caused listeners to orgasm and issue it only in hi-rez DVD-A, fully watermarked, could we not save the format and get rich in the process?

 

[daved64]

If we could write a song that caused listeners to orgasm and issue it only in hi-rez DVD-A

Sadly, such a song exists, but Be Bop Deluxe's Adventures In A Yorkshire Landscape will probably never be issued on DVD-A.

 

[Almen]

While you guys have been reading books and all that stuff I have been downloading a new torrent

You don't have to sit by and watch the torrent download. You could do other stuff while the computer works - such as reading a book.

Or, as I am going to do tonight, rehearse with my band.

 

[simonwagstaff]

I let them run overnight so I can sleep while they download...



I have the utmost respect for people who can continue to make music through their adult life. I was in my share of bands in high school and college and then the time it took to keep up with a wind instrument just got away from me. At some point I decided to listen to great musicians rather than be a below average one myself though they are not mutually exclusive. I think the point is more the following: does mathematics and testing using our current paradigms tell us everything that needs to be known about a recording format? I think it does not, but that is just my opinion. there is still some magic in sound and music that needs to be reproduced.

 

[Almen]

does mathematics and testing using our current paradigms tell us everything that needs to be known about a recording format?

Well, what do you need to know? You don't even want to sit down and read a book, so do you really want to know?

Before you can answer your own question, you will have to learn the "current paradigms".

 

[simonwagstaff]

Hmm, you are very sure of yourself, and a little nasty. Who says I haven't read any of the books? I respect your opinion, you don't hear a difference, you don't think hi-rez audio is worth it. My own experience, which seems to carry no credence, tells me otherwise. Reading books is good, listening is even better.

Nothing to be accomplished here. I won't change your mind and I don't see the need to. You won't change mine either.

Adios!

 

[Almen]

I respect your opinion, you don't hear a difference, you don't think hi-rez audio is worth it.

Except that that is not my opinion. I would appreciate if you could refrain from making up your opponents' points of view that are not correct, as you have been prone to in this thread.

You don't like hi-rez, you think it is no better than hi-bit rate MP3

O.K., we can add a cheap solid state amp to the Miller Lite, greasy hamburgers and limp french fries

 

[Eggplant]

you don't hear a difference

This whole kerfuffle is not even about a "difference".
It's about an improvement -- a demonstrable and susbstantial one.

Studies show subjects tend to overlook or ignore flaws in the media they're used to, while heaping unrelenting criticism on small but unfamiliar drawbacks of the new media. The audiophile community lambasted early CD's for their harshness and lack of low-level detail, etc. And what about vinyl's "imperfections" -- surface noise, inner groove distortion, rumble, groove wear, RIAA EQ error, dynamic range limitation, playing time limitation, IM distortion restrictions affecting mixes . . . ? Those were all perfectly acceptable. And if you just pressed only 50-gram virgin vinyl with Direct Metal Mastering and limited pressings to only 5000 copies, they'd be so much better.

It's a good idea to keep in mind that the audiophile-nonsense world has a vested interest in expensive, limited-edition things like 50-gram vinyl discs, because that's what they're selling -- specialty. That high-quality CD's could be pressed for a dime and bring better quality to the masses surely was quite annoying. They would have to come up with a new format now. And come up with it, they did. (Several, in fact.)

The first CD's were bad, but not because of any inherent format limitations. They sucked for a usual reason -- sloppy rush releasing combined with lack of experience with a new technology. Later CD's, properly mastered, sound very good indeed.

The question is not whether there's a difference between CD's and hi-rez.
That could mean slightly better, slightly worse, or some combination.
Nobody upgrades their system and buys 100's of new discs for that.
It needs to be clearly, demonstrably better -- and better enough to justify the cost, inconvenience of a new format, and its inevitable new drawbacks.

Do SACD's sound "different" to me? Sometimes, yes. A little, maybe.
Does that make them better? No. Worse? Not that I've noticed.
Have I been presented any real evidence that they're substantially better?
Not at all.

Even in audiophilia, extraordinary claims still require extraordinary evidence.

 

[simonwagstaff]

Opponent?

Sheesh.

Adios!

 

[Key]

Okay how about this for a hypothesis? If you were to record a song and mix it in 96kHz 24-bit. Take the final master and in a very specific manner convert it to 44.1 16-bit and back again to 96kHz 24-bit that the unaltered 96kHz master will be indistinguishable (unable to ABX) from the resampled copy.

I don't know either way but if I wanted to know that is where I would start. I can make the recording if you like. When it comes down to it the only thing we will be testing is if it makes a difference on my recordings but it's a decent model that can be repeated and keeps the variables down to a minimum. The testers will need a decent 96kHz soundcard, a PC, and Foobar though.