Digital recording query

From Edison cylinders to pre-amps to ProTools: talk about it here.
User avatar
Xenu
Sellout
Posts: 2209
Joined: Mon Apr 07, 2003 8:15 pm

Digital recording query

Postby Xenu » Thu Apr 06, 2006 6:22 pm

Malc writes on sh.tv:

But you are forgetting that sounds begin at some point in time. If you have 20 people banging on bottles with their house keys do you think 44.1kHz will capture all the "attacks" going on that the human ear can detect? How about 96kHz? From what I gather even 192kHz may not capture what the human ear can detect in terms of consecutive attacks. It's not all about reproducing a fourier transform of a bunch of continuous sine-waves......


This actually brings up a point I'm a bit confused by. I always thought that digital recording was a continuous process, i.e. that "sampling" is a bit of a misnomer; the digital recorder isn't checking in 44100 times a second, making a little "note" of what's happening, and jotting it down, but is instead taking a continuous waveform and converting it into 44100 values per second. Malc seems to be of the position that the former happens--that the digital recorder checks in 44100 times a second, and thus can't possibly "hit" all of these attacks because they might happen on one of those "off-samples" or something.

I could, however, be totally wrong, and I'd like the actual technical people here to address this briefly before I say anything else.
-------------
"Fuckin' Koreans" - Reno 911

User avatar
lukpac
Top Dog and Sellout
Posts: 4585
Joined: Wed Apr 02, 2003 11:51 pm
Location: Madison, WI
Contact:

Postby lukpac » Thu Apr 06, 2006 7:54 pm

What he keeps forgetting is that if that "attack" was so short as to fall between two samples, you'd never hear it anyway as it would be over 22 kHz. Unless he thinks you could differentiate between multiple "attacks" initiated within 1/44,000 of a second, which I'm skeptical about.
"I know because it is impossible for a tape to hold the compression levels of these treble boosted MFSL's like Something/Anything. The metal particulate on the tape would shatter and all you'd hear is distortion if even that." - VD

User avatar
lukpac
Top Dog and Sellout
Posts: 4585
Joined: Wed Apr 02, 2003 11:51 pm
Location: Madison, WI
Contact:

Postby lukpac » Thu Apr 06, 2006 8:04 pm

And actually...

Take any mono CD. Slide one of the channels by one sample. Do your ears detect two different sources on transients, or a single source, possible with an odd phase?

That's what I thought.
"I know because it is impossible for a tape to hold the compression levels of these treble boosted MFSL's like Something/Anything. The metal particulate on the tape would shatter and all you'd hear is distortion if even that." - VD

User avatar
Rspaight
Posts: 4384
Joined: Wed Apr 30, 2003 10:48 am
Location: The Reality-Based Community
Contact:

Postby Rspaight » Thu Apr 06, 2006 8:18 pm

Oh, no! 192K might not even be enough!

I will settle for nothing less than 128-bit, 4MHz recordings. I want my DAC to interfere with shortwave radio transmissions.

Ryan
RQOTW: "I'll make sure that our future is defined not by the letters ACLU, but by the letters USA." -- Mitt Romney

User avatar
Xenu
Sellout
Posts: 2209
Joined: Mon Apr 07, 2003 8:15 pm

Postby Xenu » Thu Apr 06, 2006 8:28 pm

Why do I do this to myself over and over?

All of you are banned, right?
-------------

"Fuckin' Koreans" - Reno 911

User avatar
Xenu
Sellout
Posts: 2209
Joined: Mon Apr 07, 2003 8:15 pm

Postby Xenu » Thu Apr 06, 2006 8:36 pm

Because, of course, my deep dark secret is that unlike Krab and ThomH, I really don't specialize in this area, and being the sole person in that particular argument allows me to get bashed over the head with things like tri-delta sinosoidal waves that follow a fourier series, as I really don't know if that makes any sense. It's like evolution debates...I'm fine until it really gets down to the microbiological level.
-------------

"Fuckin' Koreans" - Reno 911

User avatar
lukpac
Top Dog and Sellout
Posts: 4585
Joined: Wed Apr 02, 2003 11:51 pm
Location: Madison, WI
Contact:

Postby lukpac » Thu Apr 06, 2006 9:09 pm

Xenu wrote:All of you are banned, right?


"Suspended" technically.

If I wanted to I could probably get back easy enough. That's a headache I don't need though. Much easier to post via proxy if necessary.
"I know because it is impossible for a tape to hold the compression levels of these treble boosted MFSL's like Something/Anything. The metal particulate on the tape would shatter and all you'd hear is distortion if even that." - VD

User avatar
lukpac
Top Dog and Sellout
Posts: 4585
Joined: Wed Apr 02, 2003 11:51 pm
Location: Madison, WI
Contact:

Postby lukpac » Thu Apr 06, 2006 9:34 pm

Suppose the time interval between the two events is in the order of 22.68 microseconds - in that case they won't be encoded accurately with 44.1kHz sampling.


If they are that close you won't be able to differentiate the two.

But, you may say, we can't hear anything above around 20kHz so big deal.... unfortunatly that only applies to continuous waveforms.


Say WHAT?

Beyond this, our host has pointed to his mastering of the Creedence albums for SACD vs. CD as an easy way to hear the difference generated by the difference in sampling rates and signal format (redbook vs. DSD). He claims, and others have backed him up with their observations, that the sense of ambience in at least one of the tracks is truncated by the lower sampling rate. By that I mean the sense of the spacial depth of reverberation as the sounds fade to silence, in other words, the signal decay envelope is reproduced differently and is clearly audible by anyone. Steve is certainly not the only one to make this type of observation, and you can easily verify this phenomenon for yourself if you doubt.


You can also verify a Redbook conversion of the DSD doesn't have this problem, meaning the problem was with the original Redbook transfer.

I should stop now.
"I know because it is impossible for a tape to hold the compression levels of these treble boosted MFSL's like Something/Anything. The metal particulate on the tape would shatter and all you'd hear is distortion if even that." - VD

User avatar
lukpac
Top Dog and Sellout
Posts: 4585
Joined: Wed Apr 02, 2003 11:51 pm
Location: Madison, WI
Contact:

Postby lukpac » Thu Apr 06, 2006 9:38 pm

AHHHHHHHHHHHHHHHHHHHHHHH

David, It seems that you are taking the view that the human ear is sensitive only to amplitude-response variation across a limited frequency range (20Hz -20 KHz). I would contend that human hearing also detects time-domain distortions or timing errors that cannot be measured directly in the frequency domain. Isn't that why jitter measurements are conducted nowadays (for one example) in addition to the standard frequency-response measurements when digital equipment is being designed and tested? Jitter has been demonstrated to be audible to very low thresholds. There are probably other time-domain distortions that are audible as well - perhaps quantization errors introduced by digital sampling?


How is frequency not "time domain"?
"I know because it is impossible for a tape to hold the compression levels of these treble boosted MFSL's like Something/Anything. The metal particulate on the tape would shatter and all you'd hear is distortion if even that." - VD

User avatar
Xenu
Sellout
Posts: 2209
Joined: Mon Apr 07, 2003 8:15 pm

Postby Xenu » Thu Apr 06, 2006 10:35 pm

I'm done. Someone else can do it.
-------------

"Fuckin' Koreans" - Reno 911

User avatar
lukpac
Top Dog and Sellout
Posts: 4585
Joined: Wed Apr 02, 2003 11:51 pm
Location: Madison, WI
Contact:

Postby lukpac » Thu Apr 06, 2006 11:57 pm

This is why a 24/96 recorded 1kHz tone at 0dB sounds better than a 32 kbps MP3 1kHz tone at 0db. They'd both measure the same on a frequency analysis (barring any impure harmonics).


If they sounded different, they wouldn't measure the same. DUH.
"I know because it is impossible for a tape to hold the compression levels of these treble boosted MFSL's like Something/Anything. The metal particulate on the tape would shatter and all you'd hear is distortion if even that." - VD

User avatar
Rspaight
Posts: 4384
Joined: Wed Apr 30, 2003 10:48 am
Location: The Reality-Based Community
Contact:

Postby Rspaight » Fri Apr 07, 2006 8:08 am

I realize that "0dB" in the above context refers to "no attenuation", but it still sounds odd to talk about two 0dB signals sounding different.

Ryan
RQOTW: "I'll make sure that our future is defined not by the letters ACLU, but by the letters USA." -- Mitt Romney

User avatar
krabapple
Posts: 1615
Joined: Wed Apr 16, 2003 4:19 pm

Postby krabapple » Fri Apr 07, 2006 11:41 am

Xenu wrote:Because, of course, my deep dark secret is that unlike Krab and ThomH, I really don't specialize in this area, and being the sole person in that particular argument allows me to get bashed over the head with things like tri-delta sinosoidal waves that follow a fourier series, as I really don't know if that makes any sense. It's like evolution debates...I'm fine until it really gets down to the microbiological level.


Specialize? No. I'm a biologist by trade. Reading the technicalia of audio is just a hobby...and I certainly do NOT still understand it all. But I know where to look for people who do...

Dan Lavry posted this back in january, on his prosoundweb board, in reasponse to yet another claim that 'more samples are better'.

I wish I had a quarter for each time I said it:
Your common sense makes you think of "more dots is better" or now "thin slices". But that "street's common sense" is leading you into the wrong conclusion. The slices need to be thin enough, but only to a point. This is Nyquist theory, and it is one of the main "legs of all modern technology. Say you want to have 20KHz audio. In theory, you need to take more then 40000 "slices" (40KHz sampling). In practice you may need to raise it a bit. The "reconnecting" ot the dots into a PERFECT wave (like th original analog one) is done by an analog filter. A perfect filter will make a perfect wave, at sample time and ALSO BETWEEN THE SAMPLES. That is why that "stuff" works.

The misunderstanding or lack of awarness by non technical people regarding Nyquist is the main reason behind that 192KHz hype. The electronics engineers understand Nyquist. The marketing types often do not. They took thier "street sense" into costly IC fabrication and "new wave of 192KHz gear", wrongly selling "more resolution". The sad thing - once they invested in that flawed concept, they refuse to take it back and acknowladge they simply took on what is over their heads...
"I recommend that you delete the Rancid Snakepit" - Grant

User avatar
lukpac
Top Dog and Sellout
Posts: 4585
Joined: Wed Apr 02, 2003 11:51 pm
Location: Madison, WI
Contact:

Postby lukpac » Fri Apr 07, 2006 12:20 pm

Now the theory is the decay after a transient could start in less than 1/44,000 of a second.

These people don't have a clue.
"I know because it is impossible for a tape to hold the compression levels of these treble boosted MFSL's like Something/Anything. The metal particulate on the tape would shatter and all you'd hear is distortion if even that." - VD

User avatar
Rspaight
Posts: 4384
Joined: Wed Apr 30, 2003 10:48 am
Location: The Reality-Based Community
Contact:

Postby Rspaight » Fri Apr 07, 2006 1:01 pm

What I don't understand is how, if this "Fortunate Son" BS were true (which we've proven it isn't), why those 22KHz+ decays were faithfully recorded on the 35-year-old analog tape that topped out around 16K or so at best...

Ryan
RQOTW: "I'll make sure that our future is defined not by the letters ACLU, but by the letters USA." -- Mitt Romney