which sample rate to record guitars vocals etc

P

powpowmeow

New member
i have been using 24 bit / 96 khz for a while.
was wondering if there is any advantage to using different sampling rates?
what do you guys use?
any reason to use 192khz instead of 96khz, for example? i realize there is more headroom etc

thanks!
 
There's no reason to use any sample rate other than 44.1 KHz for music, or 48 KHz for video projects where 48 is the standard. Anything higher than these sample rates just wastes hard disk space and lowers your track count, with zero audible benefit.

--Ethan
 
TimOBrien said:
Agree with Ethan.
Click to expand...


Ethan Winer said:
There's no reason to use any sample rate other than 44.1 KHz for music, or 48 KHz for video projects where 48 is the standard. Anything higher than these sample rates just wastes hard disk space and lowers your track count, with zero audible benefit.

--Ethan
Click to expand...

TimOBrien said:
Unless you have a multi-million dollar studio, 24bit/44.1khz is all you need.
Click to expand...

Not strictly the same thing, is it?
 
Even in a multi-million dollar studio, anything over 48k is pointless. It simply gives you the ability to record frequencies that aren't there, the mics wont pick up, and we can't hear. That ability costs hard drive space, track count andplugin instaces. Totally not worth it.
 
Some instruments do produce harmonics above 20K. Electric guitars and human voices aren't among them (maybe some of the higher portions of the noise blast we call S or T?). No, we can't hear those frequencies, and there's no good reason to record them except...

...if you intend to "varispeed" the track down a significant amount. Something like a cymbal (whose harmonic series easily reaches into the supersonic range) slowed down to half time will sound pretty dull if it was recorded originally at 44.1K because (after the speed change) it has a super steep LPF somewhere between 10-11K, and is missing the top octave that you'd expect to hear. Record it at 88.2K, though, and it will sound much more natural after you slow it down. It's a corner case, probably more for noise artists and sound effects people, but it's the one good argument for higher sample rates that I've come up with.
 
Farview said:
Even in a multi-million dollar studio, anything over 48k is pointless. It simply gives you the ability to record frequencies that aren't there, the mics wont pick up, and we can't hear. That ability costs hard drive space, track count andplugin instaces. Totally not worth it.
Click to expand...

Exactly. And few speakers play back ultrasonics either.

ashcat_lt said:
if you intend to "varispeed" the track down a significant amount. Something like a cymbal (whose harmonic series easily reaches into the supersonic range) slowed down to half time will sound pretty dull if it was recorded originally at 44.1K
Click to expand...

Good point, though as you said this is a very special case and certainly doesn't apply to "normal" recording. Even if you have a million dollar studio.

I'll further argue that capturing ultrasonic content we can't hear is not only not needed, but can cause audible IM distortion at loud volumes in your ears.

--Ethan
 
So you are worried about a natural sound on cymbals that are pitched down an octave?

I suppose you are right, but you had to look pretty far for a scenario that would require it. It also assumes the mics dont have a rolloff below 30k.
 
I'm not worried about anything, but I do a lot of extreme stretching/speed changes, and one of the issues ive always had has been the fact that after you drop something a couple octaves it loses anything you might call treble. I usually just add distortion to "replace" those higher harmonics.

I did fairly recently record myself washing silverware in a stainless steel pot at 88.2K and was amazed at how completely normal it sounded when I slowed it down to half time. Like, you'd have never known! Except there were some voices in the background which made things pretty obvious.

I think the natural dropoff of most mics is going to be significantly less severe than any decent anti-aliasing filter.
 
if 44.1 is an arbitrary music industry resolution, (as opposed to 48khz, somewhat less arbitrary) why record at double of 44.1 (88.2k)?

Just curious.

Regarding topic, we do a lot of extreme pitch and time shifting. it appears recording at 96khz may be useful. Hopefully our lack of net worth will not cause a rip in the fabric of the universe when we record at 96khz
 
It's not going to help much on voice and especially on electric guitar, except that the "air" around it might sound more natural after pitch shifting.

I prefer to avoid sample rate conversion as much as possible in normal practice. 44.1K is plenty and is easier on everybody. The field recorder I used for the sink experiment only went up to 88.2K. But there's nothing special about any of them except for the AA filter cutoff.

The big problem with ultrasonics is that you never know for sure what's up there until you pitch it down! That applies to every aspect of the process from the ambient noise and acoustics in the room to mic response to, well, everything else in the chain. There are a lot of things in our world which are specifically designed so that the noise and distortion they make is just a bit higher than anybody can hear.
 
Farview said:
Even in a multi-million dollar studio, anything over 48k is pointless. It simply gives you the ability to record frequencies that aren't there, the mics wont pick up, and we can't hear. That ability costs hard drive space, track count andplugin instaces. Totally not worth it.
Click to expand...

~~


More to the point than hitting sampling rates above 48kHz is the bit depth.

The difference between 44.1kHz and 48kHz is clearly audible to a moderately good listener with little or no ear damage. Sampling rates above this tend to be a little excessive. However the difference between 16-bit and 24-bit sampling is "night and day" - even at 44.1kHz. They increased availability of dynamic range afforded by those extra bits throws the noise floor miles further down. Even with a cheap interface like a 48kHz "gaming" sound card or entry-level M-Audio/AVID Fast Track interface and similarly entry-level Behringer C1 or C2 microphone, the difference is like the difference between AM and FM radio.

It's also sensible to record using settings that are likely to be available on friends' equipment. Showing-off and working at 192kHz, 32-bit floating point is all well and good, until you discover your musical partner only has 48kHz, 24-bit.

Switching between 16-bit/24-bit makes a clearly audible difference even with cheap microphones and "domestic" sound cards; the difference between 48kHz and 96kHz is rather less profound, even in top-flight pro studios.

Also - 192kHz require 4 times as much storage-per-second as 48kHz. Far better to record at 48kHz and put the money saved on disc space towards better microphones. The microphone is the key, key factor to good acoustic recording - anything lost by a poor microphone cannot be retrieved later.
 
I disagree with radiophonic's first paragraph. These things might have been true in the past, but low pass filters on 44.1k no longer reach down far enough into the audible spectrum to make any real difference, certainly not an obvious one. Assuming your converters are less that 7-10 years old.

Even the difference between 16 and 24 bit isnt really night and day anymore. When there was a difference, it eas the design of the converters that caused the difference, not simply the sample rate.

This is just out of date info.
 
You must log in or register to reply here.
Back
Top