Why bother mixing in 24 bit or float when it's just going to get bounced back down to 16 bit? Because the mixing and mastering process itself can be lossy, so it's better to work at a high resolution.

Almost every financial system on the planet uses 4 decimal points internally, even though every dollar amount is rounded to 2. If you're doing many calculations, you always work at a higher precision than what you need on the output, otherwise you end up with inaccuracies.

I still have yet to see anyone prove that they can hear the difference between 44/48 and 88/96, but the precision of your final result is an entirely separate matter from the precision of your calculations.

___________________
My party schedule:
2009-02-21 - DJ Attention @ I'm So Popular
2009-06-18 - DJ Annoying @ People Need To Know Where I'll Be
2012-11-32 - DJ Insufferable ɸ Or At Least the Stalkers I Complain About
2048-06-66 - Spastic & Whocares ¶ Although I'm Actually Flattered
9999-45-81 - Tweaker Gimp ☼ I Probably Won't Even Go To This But I Have To Make Sure I Fill Up All The Available Space Here

If you have both options available, it is always better to just use the higher rate.

However, if only 44 hz is available, don't go panicking and buying more things.

This guy says the device chain in a PA should begin with a 96kHz wave file. Computer Music Tutorial also asserts that, even when converting down to 44kHz, it's best to start with as high a resolution as possible.

___________________

Now with extra singles!
my old stuff, not quite up to snuff - but I still dig it - UPDATED 9/23/2012

This seems to come up every now and then.

According to Bob Katz' mastering book, working at higher bit rates improves the end result when coming back to 16 bit.

Working at higher sample rates doesn't improve things at all when coming back to 44.1 kHz. It'll only use up more hard drive space, RAM and CPU.

Of course, it gives you the luxury of running off a higher resolution version if it's going to go on a Blu-Ray disc or something at some point (if you've recorded external instruments).

quote:

Originally posted by EddieZilker This guy says the device chain in a PA should begin with a 96kHz wave file. Computer Music Tutorial also asserts that, even when converting down to 44kHz, it's best to start with as high a resolution as possible.

Thanks for that video, that was actually really enjoyable to watch

___________________
www.soundcloud.com/timestretcher

if you export from your daw and convert directly to mp3 without any mastering or anything i dont see how it matters, the daw works in 32bit anyway dont they? of course if u plan to send it to mastering or something like that i see the point but seriously do any of us produce at that level anyway? if youre recording external synths or vocals etc id probably work with 24bit though. im not sure about this but i havent had any problem producing in the box at 44,1 16bit but ive had problems with the oposite. unless you find a good reason to do elsehow id stick to the standards. imo. most of you dont have monitoring (soundcard+monitors) being able to reproduce the microdynamics in 24bit anyway. elitist thinking never get things done.

edit: what im trying to say is that if you are working with software only i dont really see any point in exporting the final result as anything else than 44,1 16bit unless you plan to master it. your soundcard settings should be as good as possible for the best possible monitoring while producing. IE 24bit 96kHz is pretty common. This doesnt mean its any point in rendering at that rate. I dont really know how to explain what I really mean.

edit2: lol. in the end its all about how you produce. if you record stuff and do external mastering etc etc (like a real oldschool band or something), definitly work in as high bitrate and samplerate as possible. But if you, like me, only works ITB with software, and are a noob like most of us, you SHOULD actually export as 44,1 16bit, but monitor at 24bit 96kHz if you have the possibility in your soundcard. send a label a 24bit 96kHz wav for release and they might fuck up the mp3 convertion, it has happened to two of my releases and I know many others had the same problem. Submit 44,1 16bit wav to labels is my advice.

Last edited by Zak McKracken on Mar-30-2011 at 16:51

quote:

Originally posted by derail Working at higher sample rates doesn't improve things at all when coming back to 44.1 kHz. It'll only use up more hard drive space, RAM and CPU.

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.

___________________
My party schedule:
2009-02-21 - DJ Attention @ I'm So Popular
2009-06-18 - DJ Annoying @ People Need To Know Where I'll Be
2012-11-32 - DJ Insufferable ɸ Or At Least the Stalkers I Complain About
2048-06-66 - Spastic & Whocares ¶ Although I'm Actually Flattered
9999-45-81 - Tweaker Gimp ☼ I Probably Won't Even Go To This But I Have To Make Sure I Fill Up All The Available Space Here

Without adding more to what has already been said (in favour of using a higher resolution), might I just add that without using a sufficient low-pass filter before resampling down to your output sample rate, there is no point in using a higher resolution. Check out Voxengo's r8brain PRO if you're really serious about resampling, especially its LinPhase and ultra-steep modes.

___________________
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 derail This seems to come up every now and then. According to Bob Katz' mastering book, working at higher bit rates improves the end result when coming back to 16 bit. Working at higher sample rates doesn't improve things at all when coming back to 44.1 kHz. It'll only use up more hard drive space, RAM and CPU. Of course, it gives you the luxury of running off a higher resolution version if it's going to go on a Blu-Ray disc or something at some point (if you've recorded external instruments).

I'd agree with this. Sorry but I don't see the point of sampling at 96khz. Especially when going back to 44.1 since they don't divide evenly and I understand this causes issues.

24 bit I sort of understand though because it means you have more headroom if you decide to perform some dynamics proccesses later.

___________________
New Mix: March 2010 Promo
Soundcloud|Facebook

Tried working in 96khz but the only differences i noticed were a bigger file size for my renders and higher cpu load

___________________
hearthis
soundcloud
youtube

quote:

Originally posted by evo8 Tried working in 96khz but the only differences i noticed were a bigger file size for my renders and higher cpu load

my point exactly BUT you should split the discussion into two:

A: during work/monitoring
B: export/rendering and mastering

during A you should have as high as possible without having problems with CPU load etc. if you record external gear this should also be as high as possible, and only then will project-file size be a problem. this is simply your soundcard settings. 24bit 96kHz for example.

when exporting (B) you need to decide whether you plan to send it for mastering or not. incase not you should render as 44,1kHz 16bit as that is what labels, digital download websites, promo CDs etc etc all wants. sending on a different format than standard will in some cases fuck up the whole release and mp3 coding etc etc.

if you plan to send it to mastering you should render your track at whatever the selected mastering company wants it on. if you plan to master yourself in a different software than your sequencer then just render at highest usable format, like 48kHz 24bit or something. then you render another new file with your mastering program at 44,1kHz 16bit.

my point is this: never have your finished product as anything else than 44.1 16bit wav as this is the industry standard. if you send around your 24bit 48kHz (or whatever) all kinds of things might happen to it - it will fuck up. on the other hand you should monitor at the best possible settings.

if you plan on getting your track onto a SACD or surround soundtrack or something like that, you can render two versions. but never send the other version to labels unless they specifically ask for it.

to sum up:
monitoring/recording: 24bit 96kHz
rendering/final: 16bit 44.1kHz