Skip to main content

For isn'tance when Digital Cameras hit 11.4mega pixels they surpassed film in equivalent resolution...

Know what the audio one is?

Topic Tags

Comments

anonymous Mon, 07/28/2003 - 05:37

I would recommend people go by DPI (dots per inch) instead of mega pixels as mega pixels is just a set number, the result can be scaled up to the point where you can really see the pixels but when printing, you'd have to print that on a sheet of paper many feet wide.

As for audio, I personally believe that 1-bit/2.8MHz format surpasses any analog recording and can reproduce realistic sounding sounds... it's all arguable though, either way the true resolution is infinite, just like in the visual

RecorderMan Mon, 07/28/2003 - 10:02

Originally posted by sheet:
The ear is very much a digital listening device, not analog, and the AVERAGE ear has an equivalency of 24bit/48kHz.

I got that from my bro, Dr. Handy, PhD.

the AVERAGE ear...that's the crux...our ears aren't average (but I have been known to EQ the shit out of the wrong channel sometime after 16hours...and think I'm hearing something..so what do I know-LOL) ;) :p :p:

sdevino Mon, 07/28/2003 - 16:10

Physics and nyquist say you can capture all the frequency information you can hear as long as the sample rate is at least 2x what ever your hearing limit is. Bit depth required depends on the listening environment. Currently 20 to 24 bit digitaizers have more usable dynamic range than any electrical signal path so that is more than enough for an electronic based system

The problem is no one is sure if physics and nyquist are all there is too it.

Steve

sserendipity Wed, 07/30/2003 - 10:51

>The problem is no one is sure if physics and >nyquist are all there is too it.

Actually, yes they are sure, and yes there is more to it than the nyquist theorem states.

The nyquist theorem only holds true for perfect sine waves. Any complexity in the signal, otherwise considered as frequencies above nyquist limit, are not reproduced.

However, the human ear is not designed for listening to perfect sine waves - it's designed to listen to real sounds - waves of all shapes and kinds. Unlike a digital signal processor, it doesn't do this by fourier transformation of the signal - ie it doesn't split up a signal into it's component frequencies and analyze them separately. The whole ear is involved in listening to every signal, (to paint it broadly).

Consequently, you can discern the difference between wave shapes that, according to fourier transformation, constitiute frequencies well above the upper end of your listening frequency range.

There's plenty more going on, of course - both in your ear, and in the gear. For example, the closer one approaches the maximum frequency that a sample rate can produce, the more one is beholden to the D/A converters that are producing the sound. Depending on the quality and flavour of the signal process, the shape of the signal will change. Since the original transient information (the complexity I mentioned above) is lost, it's up to the designer to decide how the signal should be shaped - and whether the gear actually does the intended...

anonymous Wed, 08/06/2003 - 13:30

Not that this is quite on the subject but...

I find it interesting how engineers always try to put something that is an 'experiencial' thing into purely technical terms.

For example, I don't know how many times I've heard church goers complain about not likeing the 'loudness' of drums in church. when you explain to them that an average Church organ cranking away is many times louder than a drum set being played from 30 feet away they just stare at you like you're crazy. you can show them on an SPL meter and they still won't believe you.

The question's answer is that it's pretty hard to just spit out a number. I've heard exceptional sounding old 16-bit converters that blow away the cheap 24-bit converters that are floating around today.

It reminds me of when I was working house sound at a 2600 seat auditorium here in Lincoln. The designer was the venerable Paul Venneclausen (sp?) The hall sounded great but the sound system was awful. We had a meeting with all the contractors to discuss the problem. One in particular was that there was no speakers aimed at the balcony. Paul's idea is that the hall sounded so good they didn't need them. The reality was that we had over 900 written complaints that patrons in the balcony couldn't understand what was being said, sung, etc... The design engineer was angrily pounding the table saying, "I've got an SPL meter reading that says the sound in the balcony is 3db hotter than the floor. Those people don't know what they're talking about when they say they can't hear!" I just stared at him and plopped the box of 900 letters on the table in front of him and told him that these people didn't agree with him and his meter be damned.

All that to say that I don't think the question has a good technical answer. Hmmm sounds like I'm ranting but I'm not really.

ejm

sdevino Wed, 08/06/2003 - 16:00

Originally posted by Jonathan El-Bizri:
>The problem is no one is sure if physics and >nyquist are all there is too it.

Actually, yes they are sure, and yes there is more to it than the nyquist theorem states.

The nyquist theorem only holds true for perfect sine waves. Any complexity in the signal, otherwise considered as frequencies above nyquist limit, are not reproduced.

This is just false. Nyquist holds for all frequency content within the passband. Where did you find information that differs?

However, the human ear is not designed for listening to perfect sine waves - it's designed to listen to real sounds - waves of all shapes and kinds. Unlike a digital signal processor, it doesn't do this by fourier transformation of the signal - ie it doesn't split up a signal into it's component frequencies and analyze them separately. The whole ear is involved in listening to every signal, (to paint it broadly).

Nyquist is not concerned with fourier transform either. It is usually defined in the time domain. A Perfect reconstruction filter is represented by a step response with sinx/x rise time. I think you need to do some experiments to show yourself how all this works. Some of us have been working with Nyquist and VERy complex real waveforms for a very long time.


Consequently, you can discern the difference between wave shapes that, according to fourier transformation, constitiute frequencies well above the upper end of your listening frequency range.

This may or may not prove to be true in the long run but as of right now there is no reliable data that shows this to be true. Ask George Massemburg and the AES committee he chairs that is researching this if need be.


There's plenty more going on, of course - both in your ear, and in the gear. For example, the closer one approaches the maximum frequency that a sample rate can produce, the more one is beholden to the D/A converters that are producing the sound. Depending on the quality and flavour of the signal process, the shape of the signal will change. Since the original transient information (the complexity I mentioned above) is lost, it's up to the designer to decide how the signal should be shaped - and whether the gear actually does the intended...

The problem is that the original detail is not lost. You clearly need to do some more research or sit with someone that actually understands digital audio sampling.

I know this from having worked with and designed very advanced DSP based audio measurement and synthesis equipment for over 20 years.

RecorderMan Wed, 08/06/2003 - 19:48

Originally posted by Alécio Costa - Brazil:
Sa(x) = sen x /x

Oh My God.. How many centuries ago I had to calculate that..lol
:)

DAMN you Alicio...I believe in the theory that your brain can only hold so much. So when it gets full, if another fact comes in..one gets pushed out in it's place (remember the "married wuth children episode"?). Anyway, for the last several years I've been trying to forget asd much technical mumbo jumbo as I can (there's alot to get rid of). It's much more relavant if you run your own shop; but the gigs I like to do come with assistanst and techs. The more I can forget of that the more I have to create and feel in soundland....AND YOU REMINDED ME OF MATH......what was it I forgot? ????? :eek: :c: