as discussed already it wont make a noticable difference if your working ITB.

however...if your recording 16 tracks of audio from a band - ie 12 mics on a drum kit plus guitars vox etc then having it at double the resolution makes a significant difference.


but it doubles your file sizes... which can be an issue.

when i'm doing ITB stuff i'll always render in 24 bit and dither to 16 bit when burning to CD after mastering.

___________________
www.soundcloud.com/iainsanderson

So my take is that working at higher bit rates (as high as possible) is more important than working at higher sample rates.

That said, if you have the disk space and CPU (which is cheap these days in relative terms) then work as high as possible. Now bear in mind on a pair of PC speakers, working at 96 or 192 is just a futile exercise.

So all things being equal (you have the kit to take advantage of it) working at higher sample rates, is better than lower sample rates.

The main advantage as pointed out, is being able to hear more during the monitoring process, and therefore engineer, mix and produce better.

I've worked with one very well known score mix engineer who will only work at 96k - any thing less and it's a no go. One time our digital desks were being buggy an would not stay in 96k mode, he literally walked out of the mix session and took the gig to another studio.

In terms of recording at higher sample rates, there is evidence to suggest that you are able to capture high harmonics that would otherwise not be recorded. While this does not matter for lower register frequency sounds like vocals or basses, high strings have been shown to benefit from this effect, even if the maximum upper threshold of hearing frequency (20khz) is one quarter below the sample rate 96khz.

But, it will also greatly depend on your soundcard as to whether this this higher sample rate can be taken advantage of; even tiny clock variations at these higher sample rates can mean huge inaccuracies meaning the introducing of artifacts, truncation of waveforms and aliasing.

If you have something like a lavry, metric halo, lynx or prism then you're on the right track but anything below that grade, it's really not worth it all things considered.

Also, beyond a certain point (IMO 96khz) the ratio of file size to perceived improvement is not worth it.

IMO, work at 96khz of you have a pro studio grade soundcard and monitors, otherwise it's just a waste of CPU and disk space.

ive been wondering though, my soundcard doesnt have settings for bitrate. will the DAW use whatever it can when monitoring? like 32 float typically?

quote:

Originally posted by Senator Clay Davis ive been wondering though, my soundcard doesnt have settings for bitrate. will the DAW use whatever it can when monitoring? like 32 float typically?

You're in logic so unless you've changed it from the default setting you should be working in 32bit float.

Your AF4 can do a max of 24bit/96k which is more than enough. Bear in the mind the conversion from a digital signal is done in the AF4 so your DAW provides the mixing environment & headroom (at 32 bit float it's above 1500db) and with 24bit it's close 115db of usable headroom at output conversion - you're never going to get close to that unless you're doing it wrong and like blood dripping from your ears.

quote:

Originally posted by DJ RANN Your AF4 can do a max of 24bit/96k which is more than enough.

but will it do this automatically? the bitrate that is, i know i can adjust the samplerate in the AF console, but the bitrate? will logic just setup the best possible with the available device??

I wonder if certain plug-ins incorporate an 'upsampling' feature where a 22.05 kHz low-pass filter isn't used... To my knowledge so far no such (mastering) plug-ins exist (I haven't noticed a distinct difference between working at 44.1 kHz), but basically the only benefit to using a higher sample rate is if you have content (whether it be existed or generated) past 22.05 kHz.

Think about it: say you boosted the treble at 16 kHz with a wide Q...what happens past 22.05 kHz when working at 44.1 kHz? The plug-in's design would be made so as to roll off around the nyquist frequency with a steep low-pass filter to avoid aliasing, but certain plug-ins could be designed to work more efficiently without this when working at 96 kHz.

___________________
Put an end to the loudness war. Don't limit or compress your mixdown until mastering; leave the master channel alone.

Contact me for mastering.

quote:

Originally posted by DigiNut On what do you base this claim? A higher sample rate, like a higher bit rate, more closely emulates the analog signal; a lower bit rate will make a sine wave look more like a square wave, and a lower sample rate will make it look more like a triangle. Mixing at a higher sample rate allows you to keep those higher harmonics, minimizing any aliasing distortion during the mixing process itself. You'll still get a certain amount of that distortion when down-converting, but that's the same as dithering from 32-bit to 16-bit.

As I said, I based it on Bob Katz's book "Mastering Audio". That was the conclusion he reached. Working at higher bit rates made for an improved sound when going back to 44.1kHz/16 bit, but working at higher sample rates didn't. I'll dig it out and see what his main point is. Higher sample rates will give you more frequency content above 22 kHz, but this may actually degrade the sound if you aren't using appropriate filters to remove this content when down sampling.

I have a really old computer in any case, so I'd need a strong case that it's going to make a noticeable difference in quality in the final 44.1kHz/16 bit file. I suppose it's easy enough to try out. I'll give it a go.

quote:

Originally posted by Senator Clay Davis but will it do this automatically? the bitrate that is, i know i can adjust the samplerate in the AF console, but the bitrate? will logic just setup the best possible with the available device??

Yes, it's running at 24bit at all times. the sample rate is user selectable so what you're actually hearing from your speakers is 24bit/44.1k conversion quality.

So the digital signal gets to the AF4, and then is converted to analogue using 24bit, 44.1k conversion (unless you change the sample rate).

The internal processing of your daw (32BFP)is just internal digital headroom.

quote:

Originally posted by derail As I said, I based it on Bob Katz's book "Mastering Audio". That was the conclusion he reached. Working at higher bit rates made for an improved sound when going back to 44.1kHz/16 bit, but working at higher sample rates didn't. I'll dig it out and see what his main point is. Higher sample rates will give you more frequency content above 22 kHz, but this may actually degrade the sound if you aren't using appropriate filters to remove this content when down sampling.

As far as I remember it's because bit rate is based on log (not linear) so there is an advantage in recording your original material with 24-bit resolution, as it increases the available dynamic range and also translates into greater headroom, reducing the risk of overloads/ transient clipping.

Samplerate is directly linear, so if you're only going to play back your track at 96k on a 96k system, then great, it will technically sound better but it's doubtfull you can really tell the difference.

If however it's ever going to end up as 44.1k, then my advice if you really want to go to a higher sample rate is to go with 88.2k as it's an even division, not an approximation which can result in waveform truncation.

Ther is an argument to say that higher sample rates are better because if you have 30 tracks of audio, all at 44.1k then they have a give digital inaccuracy over the true analogue source.....but you're mixing 30 of them, which means you're mixing 30 times that inaccuracy together in to your final one track. So if you were to lessen that digital inaccuracy by going to a higher sample rate which was higher quality or a more accurate representation of the source, then combine 30 of them together it's a substantially less inaccuracy than before.

But to be honest it really doesn't make a shade of difference to 99% of us, unless we're recording or using a lot of analogue outboard during the engineering process.

quote:

Originally posted by Atlantis-AR I wonder if certain plug-ins incorporate an 'upsampling' feature where a 22.05 kHz low-pass filter isn't used... To my knowledge so far no such (mastering) plug-ins exist (I haven't noticed a distinct difference between working at 44.1 kHz), but basically the only benefit to using a higher sample rate is if you have content (whether it be existed or generated) past 22.05 kHz. Think about it: say you boosted the treble at 16 kHz with a wide Q...what happens past 22.05 kHz when working at 44.1 kHz? The plug-in's design would be made so as to roll off around the nyquist frequency with a steep low-pass filter to avoid aliasing, but certain plug-ins could be designed to work more efficiently without this when working at 96 kHz.

I believe that many synths do 2x oversampling, so many mastering plugins probably would also, or at least include a mode for this. In that case, there would still be an antialiasing filter, but at twice the nyquist frequency, which should help eliminate unwanted reflections back down into the freqency spectrum

___________________
New Mix: March 2010 Promo
Soundcloud|Facebook

quote:

Originally posted by kitphillips I believe that many synths do 2x oversampling, so many mastering plugins probably would also, or at least include a mode for this. In that case, there would still be an antialiasing filter, but at twice the nyquist frequency, which should help eliminate unwanted reflections back down into the freqency spectrum

I think some of the UAD plugs upsample internally, im sure other plugs do also

___________________
hearthis
soundcloud
youtube