96 or 44.1 kHz?

bEAN23 · Post by **bEAN23** » Sat Jan 26, 2008 3:25 am

Hi all,
I have found the answer for whether I should be saving my samples and projects in 24/32bit (32 of course), but I couldnt find the answer to whether I should save these in 96 or 44.1 kHz. The main purpose of the files should be loops for liveset. Does saving in 96 kHz make any difference in quality and is it worth saving the samples like this and having my soundcard set for 96kHz in the settings?
Thanks for your answers.

divisional · Post by **divisional** » Sat Jan 26, 2008 3:31 am

bEAN23 wrote:Hi all,
I have found the answer for whether I should be saving my samples and projects in 24/32bit (32 of course), but I couldnt find the answer to whether I should save these in 96 or 44.1 kHz. The main purpose of the files should be loops for liveset. Does saving in 96 kHz make any difference in quality and is it worth saving the samples like this and having my soundcard set for 96kHz in the settings?
Thanks for your answers.

I am fretting over this a bit as well. My firebox can handle 24/96, and so can Live of course, and my new computer can surely handle it, and I have a terabyte of storage, but shouldn't this be determined by how things sound? I can hear a difference - but is it a meaningful difference? This is what I ask myself. A friend told me that if i was going to have my projects "mastered" professionally, then I should record at the highest my computer will allow. But if I self master with iZotope's stuff and go 44.1 16bit CD then i would be wasting my CPU and space doing the 24/96 thing.

Things were much easier with that damn Tascam 1/2" 8 track.

kaffein · Post by **kaffein** » Sat Jan 26, 2008 3:32 am

24bit/48kHz is ideal IMO. Keep in mind your music will be listened to @ 16/44.1.

24bit/96kHz is really good for audio that you will be doing a lot of pitch shifting with, and have HD mics to capture that range with. (Race car sounds in a video game for example)

sweetjesus · Post by **sweetjesus** » Sat Jan 26, 2008 3:35 am

kaffein wrote:24bit/48kHz is ideal IMO. Keep in mind your music will be listened to @ 16/44.1.

24bit/96kHz is really good for audio that you will be doing a lot of pitch shifting with, and have HD mics to capture that range with. (Race car sounds in a video game for example)

why 48kHz? i mean.. wouldnt the bitrate conversion + additional dithering end up degrading the sound when its out at 44.1?

kaffein · Post by **kaffein** » Sat Jan 26, 2008 3:40 am

sweetjesus wrote:
kaffein wrote:24bit/48kHz is ideal IMO. Keep in mind your music will be listened to @ 16/44.1.

24bit/96kHz is really good for audio that you will be doing a lot of pitch shifting with, and have HD mics to capture that range with. (Race car sounds in a video game for example)
why 48kHz? i mean.. wouldnt the bitrate conversion + additional dithering end up degrading the sound when its out at 44.1?

For future archival reasons... 24bit/48kHz is DVD quality audio.
If you have no plans to release it in a higher quality format, or get it mastered at a mastering house... Then by all means, save everything as 44.1/16bit...

dcease · Post by **dcease** » Sat Jan 26, 2008 4:00 am

i'm with kaffein here. 24/48...

dancerchris · Post by **dancerchris** » Sat Jan 26, 2008 4:41 am

It depends on your intended final product. DVD use 48khz and 24 bit. if your final is CD then use 44.1khz and 24 bit. 96 will suck you dry.

dcease · Post by **dcease** » Sat Jan 26, 2008 5:36 am

sweetjesus wrote:
kaffein wrote:24bit/48kHz is ideal IMO. Keep in mind your music will be listened to @ 16/44.1.

24bit/96kHz is really good for audio that you will be doing a lot of pitch shifting with, and have HD mics to capture that range with. (Race car sounds in a video game for example)
why 48kHz? i mean.. wouldnt the bitrate conversion + additional dithering end up degrading the sound when its out at 44.1?

fuck them. i make music for me! if a commercial project were released, i'm sure a professional mastering facility could downgrade the file properly... without too much degradation. some high end facilities use 192, that makes it to cd alright...

mr.ergonomics · Post by **mr.ergonomics** » Sat Jan 26, 2008 8:13 am

if you DAC suck then maybe more then 44.1 khz sound better (but only due bad designed aliasing filters in your DAC/ADC).

if you consider the audible frequency range (20hz-20khz) 96 khz doesn't give you any benefits, it's not more detailed in this frequency range, you only can reproduce higher frequencies (which we can't hear...).

for recording/playback it's really that simple and it's a mathematical fact.
for plugins 96 khz maybe an advantage due the aliasing issues etc.

leedsquietman · Post by **leedsquietman** » Sat Jan 26, 2008 9:24 am

24/44.1 best compromise between quality and file size and easier ro convert fir CD or mp3 audio. 48Khz only for audio for DV, DVD or if you're using ADAT.

Any Sample Rate Conversion is going to introduce artifacts, however more often than not they will not be audible, even though they might show on a detailed spectograph.

96 Khz and above is just a waste of filespace in most circumstances.

Timur · Post by **Timur** » Sat Jan 26, 2008 10:08 am

You need to know your final products, that's right. If you are going for CD use 44 kHz or 88 kHz, for anything else use 48 or 96 kHz. MP3/AAC/OGG can be encoded at 48 kHz and some players should very well be able to play that (IPod), on computers it could even be benefitial they're played via a Creative Audigy that forces everything to be played at 48 kHz and uses a bad SRC algorithm for conversion. Nevertheless it doesn't much matter for quality, because MP3/AAC/OGG all roll off frequencies somewhere above 16 to 17 kHz anyway. Many audio-interfaces have slightly better specs at 48 kHz (signal/noise ratio, dynamic range), but that's more a matter of measuring than hearing.

Good Sample-Rate-Conversion (SRC) can convert sample-rates up and down pretty much without audible artefacts. The internal SRC of the Creative X-Fi is superb for example and uses more than 50% of the X-Fi's processing power. Unfortunately it can only be used for playback and not to convert files. Good SRC will upsample your source file upto four times the original sample-rate then convert it at that higher precision before sampling down to the final sample-rate. Here is an image of how the X-Fi does it:

Wether you should be using lower (44/48 ) or higher (88/96) sample-rates depends mainly on the plugins you are using! Some plugin Instruments sound better at higher sample-rates (Arturia Moog) and some effect plugins will also sound better because of less Aliasing happening when using the higher sample-rates. You do not hear the higher frequency range, but you do hear the higher resolution: even low frequency signals sampled at double the rate have double as many informations/samples to work with.

But using high sample-rates for your whole project will considerably lower your overal performance, because at double the sample-rate your CPU has double as much data to calculate. You may also run into problems with HD bandwidth when recording audio at high sample-rates, because your drive has to move double as much data around. Some plugins (including those in Live 7) come with a "High Quality" or "High Res" mode, this effectively doubles the sample-rate internally for only the plugin and thus offers the best solution. That means that you can work at lower sample-rates (44/48 ) in your project/DAW while these plugins use higher sample-rates (88/96) for their own calculations in order to maintain higher precision and less aliasing. Some even come add anti-aliasing filters before down-sampling their output back to the lower sample-rate of your DAW.

So my advice is to stick to lower sample-rates to keep CPU load and HD bandwidth low and use high quality modes and anti-aliasing on those plugins that offer it if you can hear a difference. If you don't hear a difference then leave them off to keep CPU load low.

bEAN23 · Post by **bEAN23** » Sat Jan 26, 2008 4:32 pm

Well, the final product should also be CD or MP3, but the primary intension is to have the best possible sound when playing liveset. Live, on a sound system like this for example http://georgov.rajce.idnes.cz/18.1.2008 ... C03967.JPG

The thing I care about is whether the sound processors before the amplifiers also operate on 24/96 in most cases or not. So when they were all 44.1/24 or something, it would be useless...

Tone Deft · Post by **Tone Deft** » Sat Jan 26, 2008 4:53 pm

higher sample rate = lower latency, larger data files/throughput, more CPU load and a slight gain in audio quality that's completely debatable and in reality, negligible. - nuff said.

ilia · Post by **ilia** » Sat Jan 26, 2008 5:07 pm

44.1
record at 24 bit
submix/stems at 32 bit
final master 24 bit
you'll be good
for a while

Tarekith · Post by **Tarekith** » Sat Jan 26, 2008 5:20 pm

For playing live, I think anything over 44.1 is overkill. Even for prodution, I normally use 24/44.1 for everything, except:

- Rendering softsynths, these I render as 24/96 to help reduce aliasing. As Timur said, some softsynths sound much better at 96 than 44.1 renders, you'd be surprised at the difference in some cases. My thought is that rendering at 96 pushes any aliasing out beyond the range of human hearing even more. Note that after I render at 24/96, I do downsample the audio to 24.44.1 afterwards though.

- Rendering my mixdowns. These are rendered at 24/96 for when I master them.

96 or 44.1 kHz?

96 or 44.1 kHz?

Re: 96 or 44.1 kHz?