[superjawes]

...what I've read by neuroscience definitely suggests strongly that the amount of information isn't at risk of overloading the ability of neurons to process...

This is not what is being said.

Any audible waveform can be represented by a combination of sinusoidal waves at various frequencies. In order to adequately reproduce any given frequency f_0, you must sample at a frequency of 2*f_0 or higher. Once you're above 2*f_0, you have all of the information necessary to reproduce the original frequency.

The highest frequencies that humans can hear are somewhere between 20 and 22 kHz. So in order to capture all of the audible frequencies, your minimum sampling frequency needs to be 44 kHz. Sampling at a higher rate captures information from frequencies that are inaudible to humans, but it has no effect on how accurate the audible frequencies are.

Basically higher sampling rates increase the total information in the audio file, but they do not change the audible information.

 

[totuus]

I suppose one could listen to two sound files from a record company and compare them. Listen to the CD/44.1-16 file and the so-called studio master 192-24 file and see if you can detect any difference. Simple enough to do.  If there's no appreciable difference, then that doubles the tunes I can have!

 

[RRod]

  Sure, biological systems have limits, but we are able to make out the differences in information depth in moving from lower resolution images to 1K --> 2K -->4K images. The law of diminishing returns certainly applies strongly at some point. 
 
For sound, I see your points re: filtering etc., but that's not what I'm wondering about. It's the issue of, say, a lower information-density sound file not allowing the resolution of two tones that are close, but a higher resolution file doing just that. Isn't that significantly an issue of information density? 
 
Perhaps the point here is that the change in musicality in progressing from 44.1/16 to 96+ and 24 bit, is much less than apparent from simply looking at the sheer amount of information presented?

 
Well, audio is inherently "easier" than video, so it shouldn't be surprising that it took us longer to get visually perfect video.
 
The multiple tones thing isn't an issue. CD will play back the sum of, say, a sine wave at 100Hz and another at 100.01Hz just fine (note that musicians are never exactly, mathematically in tune with one another; they only need to be audibly so
 
On your last point: people assume that, since the change from, say, 8/22.05 to 16/44.1 is stark, that the change up to 24/96 must also be. This just isn't true, because biology gets involved.

 

[totuus]

Is there any benefit to sound quality from buying 44.1/16 files rather than buying the CD and ripping it?

 

[RRod]

  Is there any benefit to sound quality from buying 44.1/16 files rather than buying the CD and ripping it?

 
No benefit to sound quality. It's all about your preference for archiving.

 

[totuus]

This has been a very useful discussion, much appreciated!

If you purchase files then, RRod, I assume you are sticking with 44.1/16 for the main?
 
thanks again, now I get twice as much music for the same cost!

 

[lamode]

   
No benefit to sound quality. It's all about your preference for archiving.

 
 
CD makes a poor archival format, in my opinion, because if you lose your digital server, you must re-rip the entire collection which is very time consuming as soon as you have a few hundred CDs or more. Much better to archive on another hard disk - that way you can restore your entire collection in minutes rather than days, or weeks, and the tags, cue files, images, etc will all be intact.
 
It seems that the only real benefit to buying on CD is to have a tangible product in your hand, which can be nice, especially when you hand someone a gift for Christmas, for example. Or in the event that a download simply isn't available for that music yet.

 

[RRod]

  This has been a very useful discussion, much appreciated!

If you purchase files then, RRod, I assume you are sticking with 44.1/16 for the main?
 
thanks again, now I get twice as much music for the same cost!

 
Most certainly. Next step for me is a proper listening room with surround setup, but that's years down the road of money saving.

 

[RRod]

   
 
CD makes a poor archival format, in my opinion, because if you lose your digital server, you must re-rip the entire collection which is very time consuming as soon as you have a few hundred CDs or more. Much better to archive on another hard disk - that way you can restore your entire collection in minutes rather than days, or weeks, and the tags, cue files, images, etc will all be intact.
 
It seems that the only real benefit to buying on CD is to have a tangible product in your hand, which can be nice, especially when you hand someone a gift for Christmas, for example. Or in the event that a download simply isn't available for that music yet.

 
CDs are my backup-backup archival plan , and making a backup of my rips is as simple as an rsync command. But in the event that I somehow lose both my main and backup drives, I at least have the CDs which, so far, haven't died on me. Plus liner notes are, sometimes, interesting, and all for the same cost as a download if you go used.

 

[lamode]

  If true sound complexity entering my ears is composed of an enormous amount of information, then am I not getting more information from the 96kHz/24 bit representation of the analogue tape, than the 44.1kHz/16 bit representation? Assuming the tape has the best accumulation of the original sound experience in that hall. If one of those bits, very crudely put, "represents the bounce of the tympany and horns and strings etc. off the left middle wall of the hall, and so on nearly ad infinitum to become the final combination of 96kHz times 24 bits per sample, isn't this simply more information reaching me than the 44.1kHz/16bit representation could be? Am I not hearing more of the analogue tape this way? And so finally, is this not a truer representation of the original sound? (If more information in sound reproduction can be called "truer", which I suggest possibly it can.)

 

 
It doesn't work that way, exactly. No matter how many hundreds or even thousands of sound sources and reflections etc are combined, at the end of the day the air pressure at your ear drum is still at one value at any one moment in time, and all that is required to capture the auditory experience is to capture that waveform as accurately as possible. Modern digital standards like 24/96 can capture this information so accurately that the deviations are simply inaudible (-100dB or better). To put it another way, the background noise in a quiet listening room will be 50dB higher than any slight deviations in the waveform! Distortion, especially unmusical kinds of distortion such as IM distortion, from poor quality playback devices can be audible and give an unnatural quality to the music, so this is where there is still room for improvement but it is not a limitation of the format, but of the playback devices.

 

[totuus]

  . . . standards like 24/96 can capture this information so accurately that the deviations are simply inaudible (-100dB or better). To put it another way, the background noise in a quiet listening room will be 50dB higher than any slight deviations in the waveform! Distortion, especially unmusical kinds of distortion such as IM distortion, from poor quality playback devices can be audible and give an unnatural quality to the music, so this is where there is still room for improvement but it is not a limitation of the format, but of the playback devices.

 
OK, it's coming clearer now. So although the information density is higher, and some very tiny "wanted" sounds (timbre, spatial stuff, subtle sound info) would be present by increasing sampling freq/bit depth, (i) they would be progressively smaller and smaller, quite possibly inaudible, and definitely/obviously over-borne by the amplitudes of other, unwanted inputs, as well as distortion, true noise, hardware inabilities, etc., 
 
quite revelatory, many recording companies market right up at the edge, being technically able to state that if you have good enough equipment and hearing . . .
 
good ol' CDs I guess!
 
and the other focus of this discussion has been to look at what can be changed in recording, such as microphones, etc.

 

[kraken2109]

   
OK, it's coming clearer now. So although the information density is higher, and some very tiny "wanted" sounds (timbre, spatial stuff, subtle sound info) would be present by increasing sampling freq/bit depth, (i) they would be progressively smaller and smaller, quite possibly inaudible, and definitely/obviously over-borne by the amplitudes of other, unwanted inputs, as well as distortion, true noise, hardware inabilities, etc., 
 
quite revelatory, many recording companies market right up at the edge, being technically able to state that if you have good enough equipment and hearing . . .
 
good ol' CDs I guess!
 
and the other focus of this discussion has been to look at what can be changed in recording, such as microphones, etc.

The average pro microphone is a lot better than most speakers. Professional recordings are rarely limited by microphones. 

 

[bigshot]

  So although the information density is higher, and some very tiny "wanted" sounds (timbre, spatial stuff, subtle sound info) would be present by increasing sampling freq/bit depth, (i) they would be progressively smaller and smaller, quite possibly inaudible, and definitely/obviously over-borne by the amplitudes of other, unwanted inputs, as well as distortion, true noise, hardware inabilities, etc.

 
Two samples are all you need to *perfectly* reproduce any frequency. You could use a thousand extra samples to create a frequency or just two and it's exactly the same.
 
CDs are capable of reproducing all frequencies perfectly up to 20kHz, the upper edge of human hearing. Higher sampling rates can reproduce sound beyond our ability to hear.
 
There are no timbres, spacial stuff or subtle sound info that aren't expressed in frequencies. If you can reproduce the frequencies perfectly, all the rest is reproduced perfectly.

 

[lamode]

   
Two samples are all you need to *perfectly* reproduce any frequency. You could use a thousand extra samples to create a frequency or just two and it's exactly the same.
 
CDs are capable of reproducing all frequencies perfectly up to 20kHz, the upper edge of human hearing. Higher sampling rates can reproduce sound beyond our ability to hear.
 
There are no timbres, spacial stuff or subtle sound info that aren't expressed in frequencies. If you can reproduce the frequencies perfectly, all the rest is reproduced perfectly.

 
Actually there is a problem with digital sampling, close to f/2. You will get interference patterns - in other words you will get the right frequency reproduced but with varying amplitude. So I wouldn't call that "perfectly"!!
 
Example:

 

[Don Hills]

 ...  So what the listener is getting, and the brain is converting to the experience, is a complex waveform absolutely beyond imagination. Or, in another sense, as neuroscientists now classify, the amount of information that the listener’s brain is processing is monumental.  ...

 
Er... no. Your brain processes only a very small part of what enters your ears.
Read this: (You'll need MS PowerPoint or a PPT reader program)
http://www.aes.org/sections/pnw/ppt/jj/highlevelnobg.ppt
 
It's from this page:
http://www.aes.org/sections/pnw/ppt.htm