[shep]

  I think I just got trashed Should have known better than to say stuff like that without a PHD to back it up!

[Daygloworange]

Not at all Shep. It's just that those kind of analogies that started at the advent of digital still abound, and this influences people's perceptions, and they try and connect the dots (no pun intended) between the image of bits and bites, instead of a more beautiful sinuous, continuous, uninterrupted waveform that analog represents.

The old illustrations they use to use were of a S shaped sine wave, and then of the same waveform at a 3 bit sample rate. The 3 bit sample looked like 3 steps one would use to do "step aerobics".

It was a gross oversimplification of what is actually going on. But it resonated with "audiophiles" everywhere, and became the mantra of audiophiles everywhere, and has become something of an urban legend.

Cheers

[Russell Dawkins]

On the other hand, and playing devil's advocate for a second, two very well respected engineers - Tim de Paravicini and Tony Faulkner - both argue the case for higher sample rates, rather than longer wordlengths. Tony has gone on record to say he would rather record at 16/176.4 than 24/44.1.

Tim de P. says essentially the same thing when he says (in a nutshell) that 16/44.1 is good 'til 3.5kHz, 88.2 is good 'till 7kHz and 176.4 is good 'till 14kHz. Sounds about right to me, but I haven't done any careful comparisons.

I do know that a recording where the fundamentals are about right seems to transcend most of these considerations. By fundamentals, I mean good music, good musicians, good or appropriate mics well placed with decent mic preamps and good mixing (a biggie).

On yet another hand (I have three) I still remember that day 14 years ago when I had spent about 12 hours listening critically to digital recordings in the process of finding appropriate EQs to make all these different sessions sound the same. I was working on what was considered to be the best sounding DAW at the time, at least in North America - Sonic Solutions. Finally I just couldn't listen any more and I went to the kitchen for a snack and, to relax, I put on a cassette I had made from a record - analog all the way. Even though it was only a lowly cassette, what an instant sense of ease and relief!

[Daygloworange]

Tim de P. says essentially the same thing when he says (in a nutshell) that 16/44.1 is good 'til 3.5kHz, 88.2 is good 'till 7kHz and 176.4 is good 'till 14kHz. Sounds about right to me, but I haven't done any careful comparisons.

Russell,

If you don't elaborate on that, people are gonna get the wrong idea! 

Cheers 

[Russell Dawkins]

well, again in a nutshell, Tim seems to think PCM digital (CDs and DVD-As, not DSD/SACD) works well up to 3.5 kHz for 44.1 kHz sampling frequency, 7kHz for 88.2 kHz  and 14 kHz for a sample frequency of 176.4 kHz. I personally don't think that what happens above 14 kHz has much effect on what I experience while listening, so it would seem that 176.4 is as high a sampling frequency as I need to go, although theoretically it would make sense to get a little buffer and sample at 352.8 or so but realistically, for me, that's for the future.

If I had my choice, and I don't, I would record at 20/176.4 kHz.

To make it clear, my understanding is that Tim thinks that "downband spuriae" (I think that's the term) limit the clean handling of signal to a theoretical 3.5 kHz at the usual sampling rate and that's the reason for the tiring nature of CD sound.

[shep]

"It was a gross oversimplification of what is actually going on. But it resonated with "audiophiles" everywhere, and became the mantra of audiophiles everywhere, and has become something of an urban legend."    Ahh! I feel better now. Buying into an urban legend sounds classier than plain ignorance!

"On yet another hand (I have three) I still remember that day 14 years ago when I had spent about 12 hours listening critically to digital recordings in the process of finding appropriate EQs to make all these different sessions sound the same. I was working on what was considered to be the best sounding DAW at the time, at least in North America - Sonic Solutions. Finally I just couldn't listen any more and I went to the kitchen for a snack and, to relax, I put on a cassette I had made from a record - analog all the way. Even though it was only a lowly cassette, what an instant sense of ease and relief!"

I do like that slight of hand  The third hand is the one to watch! Have I just been vindicated? Time will tell. Pleasesomebody with a great system listento the "Aras" cd and tellme if I'm right...or if I'm right.

[Russell Dawkins]

"On yet another hand (I have three) I still remember that day 14 years ago when I had spent about 12 hours listening critically to digital recordings in the process of finding appropriate EQs to make all these different sessions sound the same. I was working on what was considered to be the best sounding DAW at the time, at least in North America - Sonic Solutions. Finally I just couldn't listen any more and I went to the kitchen for a snack and, to relax, I put on a cassette I had made from a record - analog all the way. Even though it was only a lowly cassette, what an instant sense of ease and relief!"

I do like that slight of hand  The third hand is the one to watch! Have I just been vindicated? Time will tell. Pleasesomebody with a great system listento the "Aras" cd and tellme if I'm right...or if I'm right.

A footnote: I made my first post in this thread without realizing you were talking not only about engineering methods, but even more perhaps about the analog/digital question.

It is probably highly significant that at the time of the episode mentioned above about the cassette I was not yet aware of the virtual necessity for the use of "dither" in the process, I was to discover that two years later, in 1995.

I was working on a choral recording of mine that had just been edited on Sonic Solutions by a friend who owned it (they were extremely expensive at the time). After a few hours I began to get more and more negative about the sound, thinking it really wasn't any better than most recordings out there, if I were to be honest and impartial (hard to do with your own stuff). It didn't seem to have any of the attributes I was working to get. Then I started to get tiny dropouts, sounding like grains of sand falling onto a table. I thought maybe I had some other technical problem, so I checked the master from which this edit was made and - lo! everything I had identified as a flaw was not there. This included finer texture to the sound - think cotton vs silk - , more specific imaging and sense of "place" within the soundstage, more sense of hall. And no dropouts.

By sheer coincidence, that night I read an article by John Atkinson in Stereophile about a recording he was editing on his Sonic Solutions and mentioned, point by point, exactly the characteristics I listed, but attributed it to the fact that he had not used dither at the end of the process. So I asked and found out that my friend had not used dither.

Needless to say from then on I always made sure that dither was used on my projects, although I believe it was years before the use of dither became standard with my friend.

I believe that many CDs made before 1998 were compromised by being edited or manipulated digitally without dither being properly applied at the end of the process. You see, dither essentially means adding a small amount of noise to the signal during conversion from high resolution to CD resolution for creation of the master.

This may be the biggest factor influencing analog/digital opinions in the earlier days of CDs.

I believe that 16/44 "done right" can sound glorious, but digital needs to be handled as carefully as analog.

[shep]

"A footnote: I made my first post in this thread without realizing you were talking not only about engineering methods, but even more perhaps about the analog/digital question"
You're right Russel (and polite) I was running two parallel lines of thought...carried away.

[Dan Banquer]

Here's the real issue, http://www.audioholics.com/techtips/specsformats/CurrentFormatTrends.php

To all who are reading this thread read the above link very carefully. In addition there are plenty more links like this on the net.
               d.b.

[shep]

Thankyou Dan for that link. I was flailing around a bit. I think that really points the finger in the right direction and makes a lot of sense. Reminds me of the whole issue of vinyl pressing. Some things never change...