CD Quality Is Not High-Res Audio

[asiufy]

follow up:

You did have to be way more careful in the days before non-destructive editing and a gazillion virtual tracks.

Exactly, and that was my point. The accessibility of digital tools, and their ease of use, made them available to millions, but how many of those millions are actually capable of doing good work on them?
So we end up blaming the tools for the job (ProTools), while the culprit is only those who operate them...


alexandre

 

[Julf]

So we end up blaming the tools for the job (ProTools), while the culprit is only those who operate them...

I still think Autotune should be classed as a Weapon of Mass Destruction...

 

[Phelonious Ponk]

I still think Autotune should be classed as a Weapon of Mass Destruction...

You'll get no argument from me on that one, but the same technology that produced the abomination Autotune, has created a pedal I can run my mic line through at a solo gig. Step on it and it will generate a couple of harmony parts to the melody I'm singing, and even generate minute, random, adjustible timing errors to make it sound human. Does it? Not really, but it's still pretty cool.

Tim

 

[rbbert]

I don't think dither is part of the Shannon-Nyquist theorem, nor is it clear that the mathematically perfect reconstructions actually occur in practice, either in A>D or D>A. It's certainly possible that 24 bit (as opposed to 16 bit) allows existing hardware and software to more closely approximate the mathematical ideal. I'm not sure exactly how that can be tested; anecdotally it seems to be true, but that's hardly convincing evidence one way or the other.

 

[Julf]

I don't think dither is part of the Shannon-Nyquist theorem

No, dither isn't. Dither is a noise shaping / shifting tool.

nor is it clear that the mathematically perfect reconstructions actually occur in practice, either in A>D or D>A.

It is not clear to me that they don't occur in practice. What makes you think they don't?

It's certainly possible that 24 bit (as opposed to 16 bit) allows existing hardware and software to more closely approximate the mathematical ideal.

Why would that be?

I'm not sure exactly how that can be tested

Actually very simply - measure the output of a DAC and compare it to the input signal.

Equally simple - take an oscilloscope and look at the output of a DAC - do you see any stair steps?

 

[Groucho]

I don't think dither is part of the Shannon-Nyquist theorem, nor is it clear that the mathematically perfect reconstructions actually occur in practice, either in A>D or D>A. It's certainly possible that 24 bit (as opposed to 16 bit) allows existing hardware and software to more closely approximate the mathematical ideal. I'm not sure exactly how that can be tested; anecdotally it seems to be true, but that's hardly convincing evidence one way or the other.

If it's the mathematical aspect of bit reduction from 24 to 16, I see no problem in testing the results entirely in software. If the question is one of the fidelity of the hardware at 24 bits or 16, then I would expect that a loopback test from a DAC and back into an ADC using various test signals, subjecting the results to batteries of analyses or run through a program like Diffmaker would be a pretty good test.

How about doing the loopback 10 times, say, in series, and asking a Golden Eared audiophile to do a DBT with the original? Unless the DAC and ADC's supposedly imperfect reconstructions were cancelling out perfectly, you would expect to get a fairly good idea of their fidelity.

(Q: Do A->D conversions rely on spurious analogue noise in the signal for dithering? If so, might the above experiment be compromised if the ADC/DAC combination was too good..?!)

 

[microstrip]

in my limited experience, totally agree. Great CDs can sound great. FIM, MA Recordings, Analogue Productions, Verve Masters, Blue Note Remasters, Channel, Harmonia Mundi...some truly beautiful mastering jobs out there. On more modern stuff, I have enjoyed the superb remasters by Bob Ludwig on CD of the Police and Sting albums.

Loyd,

Happy to know you enjoy such good recordings - I also enjoy many of them, Harmonia Mundi is one of my preferred labels. When I am listening to them the format (for me) does not matter also. At those moments I just enjoy the music.

But when I listen to similar quality well mastered recordings in other formats I really appreciate and enjoy the small difference.

Perhaps I will find in the future that with a better CD player or other type of playing equipment the CDs will sound even better than I experience now - twenty years ago I could not expect them to sound so good as they do today. I am prepared to revise my opinion at that time!

All in my humble and subjective opinion.

 

[LL21]

Loyd,

Happy to know you enjoy such good recordings - I also enjoy many of them, Harmonia Mundi is one of my preferred labels. When I am listening to them the format (for me) does not matter also. At those moments I just enjoy the music.

But when I listen to similar quality well mastered recordings in other formats I really appreciate and enjoy the small difference.

Perhaps I will find in the future that with a better CD player or other type of playing equipment the CDs will sound even better than I experience now - twenty years ago I could not expect them to sound so good as they do today. I am prepared to revise my opinion at that time!

All in my humble and subjective opinion.

Hi Microstrip...I do not doubt you at all. In fact, I am confident that (done right) hi-res must be indeed be much better. But in the same way I have committed myself to concentrate all my resources into the best digital (and avoid TT, Tape, etc)...I have also committed myself to one major upgrade and wait until the next big move on ANY of my components. For now, with the limited availability of true hi-res, the abundance of low-cost CD and very very good quality CD remasters...I am extremely happy where I am with Zanden and indeed consider myself blessed indeed to be able to enjoy it every day.

 

[rbbert]

It is not clear to me that they don't occur in practice. What makes you think they don't?

Actually very simply - measure the output of a DAC and compare it to the input signal.

Equally simple - take an oscilloscope and look at the output of a DAC - do you see any stair steps?

Too many people (me sometimes included) complain about the sound of even a very good A>D>A process for me not to wonder whether the practice actually matches the theory.

How will you compare the input to the output of a DAC? By ear, I assume? Check with Bruce B., or any mastering engineer; I don't think there is any transparent recording system or medium. Nothing to do with stairsteps, perhaps partly with quantisation errors and/or distortion, perhaps analog circuitry.

 

[rbbert]

If it's the mathematical aspect of bit reduction from 24 to 16, I see no problem in testing the results entirely in software. If the question is one of the fidelity of the hardware at 24 bits or 16, then I would expect that a loopback test from a DAC and back into an ADC using various test signals, subjecting the results to batteries of analyses or run through a program like Diffmaker would be a pretty good test.

How about doing the loopback 10 times, say, in series, and asking a Golden Eared audiophile to do a DBT with the original? Unless the DAC and ADC's supposedly imperfect reconstructions were cancelling out perfectly, you would expect to get a fairly good idea of their fidelity.

(Q: Do A->D conversions rely on spurious analogue noise in the signal for dithering? If so, might the above experiment be compromised if the ADC/DAC combination was too good..?!)

I took a 24/88.2 transfer of an analog master (to avoid the problems of a lower noise floor with a native 24 bit recording), dithered it to 16 bits using mbit+, saved the file, then expanded it to 24 bits again using mbit+, inverted it and then combined it with the original 24 bit file. As far as I can tell, they null perfectly.

 

[Bruce B]

As far as I can tell, they null perfectly.

Yeah, but do they sound differently?

 

[Phelonious Ponk]

Anybody compared Bruce's files yet?

Tim

 

[garylkoh]

Yes.

 

[NorthStar]

Anybody compared Bruce's files yet?

Tim

Yes.

And...?

 

[garylkoh]

And...?

Since I trust my ears, and there are some here who tell me that I shouldn't and should just believe that the measurements show that I won't be able to hear the difference, I'm waiting for someone else to respond first.

 

[NorthStar]

Wow! ...Where is your sense of freedom Gary? ...And autonomy too?

* If you cannot make your own mind without being influenced or misguided by others, then I'd say that you still have some frontiers to penetrate further and for the better of a deeper understanding. :b
What you share musically (your passion) with all of us, is more important than all the rest.

 

[Julf]

How will you compare the input to the output of a DAC? By ear, I assume?

By ear (ABX) is probably the most relevant way, but you could also match the waveforms using an autocorrelator function, and look at the difference signal.

I don't think there is any transparent recording system or medium. Nothing to do with stairsteps, perhaps partly with quantisation errors and/or distortion, perhaps analog circuitry.

Agree that there is no 100% transparent medium, but the question is if that contribution is significant enough to be audible when added to the contributions from the source and all the other steps in the chain? Yes, everything matters, but not to an equal degree...

 

[Groucho]

...what I did capture is low ambient room noise, which I feel brings out the best of what 24-bit can do.

Are you sure you're not thinking of the signals that lossy compressors have a hard time encoding? I always find that it's poor compression that shows up on ambient noise, rather than bit depth.

If something is random-ish, but not quite random, I think we're good at following that non-randomness with our ears. Lossy compressors, on the other hand, attempt to make decisions on which components to discard or encode based on small segments of the waveform. From segment to segment, the algorithm makes the best decisions it can, but with a noise-like (but not quite random) waveform, the compressor has to work hard, discarding a lot of information. The boundaries between segments can be abrupt and our ears pick up on this.

 

[mep]

"Lossy compressors?" What you talkin bout Willis?

 

[JackD201]

Seems somebody is mixed up between compression and compression.