Take the 2-minute tour ×
Sound Design Stack Exchange is a question and answer site for sound engineers, producers, editors, and enthusiasts. It's 100% free, no registration required.

Do I maximize sampling rate beyond 44100 Hz, or do I do something else?

And at what point do humans stop noticing any increase in sound quality?

share|improve this question
add comment

migrated from avp.stackexchange.com Jan 24 at 12:01

This question came from our site for engineers, producers, editors, and enthusiasts spanning the fields of video, and media creation.

1 Answer

up vote 3 down vote accepted

Well, it depends on what you're doing.

The primary limitations to the sound quality of any sound recorder are

  • (if you're directly recording an acoustic source) microphones
  • preamps
  • ADC

Only the third component here has anything to do with samplerate and bit depth. If the microphones are cheap electrets and the preamps noisy, going above 16 bit 44.1 kHz can't really improve the sound quality. Yet even in this case, the overall sound quality of a project will be rather better if you do at least not need to switch back and forth, so choose what fits best any media already present or to the desired target format (e.g. usually 44.1 for CDs, 48 for video).
Only when the microphones and preamps are good (or you're directly recording a line signal), the ADC becomes really important. But again, it's primarily a hardware issue: a good ADC at 16 bit 44.1 kHz can sound splendid, while a bad ADC won't sound ok even if you try to run it at 32 bit 192 kHz (which is generally not possible at bad ADCs anyway).

Still, there is one reason why it's good to record at high samplerates/bit depths (I'll call it SRnBD for now): Your signal will likely be processed digitally in some way. Now, even though it's DSP, such processing does mainly affect the "analog" qualities of a signal. These can be modelled nicely on a high-SRnBD signal, but not so well by reduced one: additional artifacts may be introduced. For instance, applying a compressor to a 16-bit signal will amplify the already present quantization errors/noise. Pulling a 44.1 kHz signal through a digital overdrive processor will produce audible aliasing frequencies. And so on.

I think 24 bit 96 kHz is the limit to where it can really make any significant difference at all anymore.

share|improve this answer
    
Oh interesting - I see. Do a lot of recorders compress the signal as they're recording? And is Gibbs overshoot one of those artifacts? –  InquilineKea Sep 19 '11 at 18:24
    
Some recorders won't even have a decent signal path as the incoming audio's digitised - my own handy recorder, the Zoom H2, is a great little thing, but even if you set to record at 24 bit the SNR and quality of the captured audio (as qualitatively measured by users) is no different from 44.1 kHz, 16 bit capture. Also, the gain switch on it adjusts the gain POST-digitisation... Meaning any clip just gets reduced (or boosted) by 10 dB. I tend to leave that switch at zero! Bear it in mind when shopping around for players - do extensive research on potential devices to identify any shortcomings. –  Christopher Woods Sep 19 '11 at 19:17
    
@Christopher Hang on... since when does the H2 have a digital input? What you mean is, I guess, when the incoming audio comes from another device's DAC. But that does in no way grant any quality benefit. It's just what I said: you pass the signal through the H2's microphone preamp and ADC, and because these aren't so great it's not gonna yield noteworthy quality boosts even if you select ridiculously high digital rates. –  leftaroundabout Sep 20 '11 at 8:22
    
@leftaroundabout I meant exactly what I said -- the gain adjustment was placed in the circuit flow post-ADC from the inbuilt mics. External audio sources plugged into the Mic/Line In socket likely pass through a similarly designed signal path so will experience the same problem. I simply said "post-digitisation" to describe how it's the already-digitised signal that's being gain adjusted (and therefore simply scaled, including any distortion) as opposed to an analogue, limiter-style adjustment pre-ADC which could perhaps cope with a little overmodulation going in. –  Christopher Woods Oct 13 '11 at 3:39
    
Ah, now I understand. Seriously, the gain control is post-ADC – are you sure? Sure it is digitally controlled in the small mobile devices these days, but that control should in any properly designed device still steer an anolog VCA or the ADC's sensitivity itself, not just perform a digital multiplication. –  leftaroundabout Oct 13 '11 at 9:35
show 1 more comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.