Bigger drives.. good enough reason to jump to 96k?

[sweetjesus]

Hello,

I've ordered my first 500GB hard drive. My current collection consits of a couple of 250's, some 320's a few 80's and some 160's and some various 2.5" drives about too..

Now that 500Gig drives cost less than registering Nebulae in the Gay Porn Workers Union .. is there a significant difference in jumping up to a higher sample rate?

I've come to identify these questions as the ones I feel most valuable in gauging whether or not it would be a good idea to go up a notch.

I'm hoping for some official answers too from the Abe's if they could be kind enough.


1. I've heard that unless you have an amazing soundcard, soundcards perform better at lower samplerates. Do things like 'jitter' and all other fandangled terminology differ at different samplerates?

1.1 Would something like jitter affect the sound noticably

2. I know it's not quite the same as headroom (bit depth) but do Live's devices perform better/differently with higher samplerates?

3. Does it make a difference in mixing or only for recorded audio material?

4. Does anyone here do stuff at 96k for music production that goes to CD?

[Tarekith]

I don't think it's really worth it for most people to be honest. I've done some songs at 96 just to see if there was a huge difference, and to be honest I didn't hear it. I think it's likely to be more of an issue if you're actually recording a lot of external instruments via mics and expensive pre's, but for the averge sample and loop based musician, not so sure. Then you also need to decide if you want to work at 96k or 88.1k, there's some debate there as to which is better (and why) for people downsampling to 44.1.

I still export my mixdowns as 96k files for mastering, but to be honest, that's mostly so I have a higher quality available should I ever need it for DVD or HD audio. Doubtful too.

[Pasha]

If I well remember I catched Mr. Henke telling that Operator at 96Khz was better...so I switched to that rate. My Audio Card is not 'top notch' but I noticed suddenly that latency was reduced by 50%. So far so good. The overall sound seemed richer, more 'body' more overtones...maybe only a psicoaudio thing but that was it. I then tried to record something and I noticed that files went soooo big and the same 16 tracks project I had before at 48Khz 24 bit, began to jitter a little. It seemed that bigger file means big work for the HDD even when only streaming, and the RAM option helped a little but with overgrown files you'll end up sucking your memory very fast. Moreover those HDD you're talking are external USB/FW discs. I know that some people might disagree on that but even with FW I noticed that little tendency to jitter and an External FW is slower than an internal SATA II. I think the way to go is eSATA that gives up to 1.5 GB/sec, amazingly. So I'd say that yes we have bigger disks available but the interface we got today (USB 2.0 and FW 400) are probably not enough for 96Khz. When eSATA will become mass market (Lacie has the d2 Quadra series, nice) probably we can afford it. We have Gigabit HDD available but capable only of 125 MB/sec. I really would like to know if someone used them already.
I came back to 44Khz 24bit and I'm happy. My USB 2.0 Drive sometimes jitters with complex projects but I'm using it only for backup and I'm on a Mac which are known to have lesser performance on the USB bus than XP has.

My 2 cents.

- Best
- Pasha

[forge]

i think you also need to get A/D converters, speakers and Amp that will take full advantage of it, given most of the nyquist frequencies added will be beyond human hearing, and probably get a word clock etc if you're going to get that anal about jitter etc

I'd personally save the HD space unless you have a few grand to spare on top notch hardware

bitrate on the other hand makes a big difference, 24bit whenever poss!

[Khazul]

Im my experience 96K is just a whole load of hassle generally due to the performance hit on your computer. The file size aspect is relatively minor next to that.

The one exception is I sometimes sample something at 96K/24bit when I know I will be doing some extreme processing on it.

[forge]

I just had a look in a Uni textbook I have at the moment that talks about it

it says "many musicians can notice a subtle but perceptible difference in audio quality"

it doesnt say how much cash you'd have to drop on it though!

most of the benefit is in preventing rounding errors in the frequencies above the sampling rate

for example, the 8th harmonic of a Piano note at 4.2kHz would be 33.6kHz (which of course needs at least 67.2kHz sampling rate to accurately catch it) and as 44.1 kHz can not handle it properly there can be undersampling and errors which can be detectable

but I'm really not sure how much of this you will pick-up on without some high end hardware

so the answer is - give it a try and see if you notice enough of a difference to make it worth the hassle

[Khazul]

Yeh - you can notice an audiable difference when sample at 96K vs 44.1 *if* the ADC and/or DAC has unbeleivably bad over-frequency exclusion filtering that results in reflected frequencies falling back into the audiable range - it maks the sound a tiny bit harsh.

Its alot easier to get the filtering and noise shaping right for 20Hz-20Khz at 96K than at 44.1K.

[Michael-SW]

Many effects and softsynths sound better at higer sampling frequencies. A quite comming technique is oversampling, which means that the synth operate internally at eg 96kHz even though your project is still 44.1 kHz. For instruments/fx that doesn't support oversampling (Operator?) then there might be an audible gain to increasing your project's frequency.

[Timur]

Well, usually you don't want to sample at 88/96k because of perceiveable quality of the recording (actually many converters/audio cards will show better signal/noise ratio at 48k), but for less aliasing artefact when processing the samples with lots of effects. So when you compare a 44/48k sample to a 88/96k sample you will most likely not hear much of a difference, given your converters work properly. Only when you run several effects there may be a difference (depending on the effect engine's quality too).

So best way to try if it makes a difference for you is to record the same source at lower and higher sample-rate, then run some effects on it and then sample down the higher sampled one to the lower sample rate (i.e. sample at 44k and 88k, run effect, then sample down 88k to 44k) and compare those. If you hear a difference it's worth the effort, if not then not!

When thinking about the output of soft-synths it's not so easy. You already quote that Operator is said to be better at higher sample-rates. I cannot comment that, but I know that Arturia Moog software used to sound alot grittier with better highs at high sample-rates when compared to low-sample rates. Saying that I am not sure if one could rate this more as a bug than as a feature, because actually any synth should be able to output the same high-range within the audible frequency-range of upto 20kHz even at 44k.

[dancerchris]

Im my experience 96K is just a whole load of hassle generally due to the performance hit on your computer. The file size aspect is relatively minor next to that.

My suggestion as well. Not only will you need big drives but RAID striping as well. This will increase your drives write speeds. Also you'll probably want some really hefty processing power like maxed out quad cores or a board that will run 2 quad cores. Vista running with > 4 gigs as well. My setup sputtered too much at 96k and the sonic differences were not noticable. (BTW I was loading big samples with DFHS and the samples quickly maxed out my memory limits.)

My $0.02

[nebulae]

I'd join the Gay Porn Workers Union before going to 96khz. All the reasons stated here are spot on:
- You'll only hear it if you're recording highly dynamic instruments like orchestras and string instruments, and even then with high quality mics, preamps, and AD converters. If not, then by going to 96khz, you'll get the following negatives without any real benefit:
- higher CPU usage
- higher disk usage
- higher ram usage

I can easily see many reasons to use 24bit 44.1khz, but any higher, and that Gay Porn Union looks a lot better for the money. They have an excellent dental plan.

[sweetjesus]

I'd join the Gay Porn Workers Union before going to 96khz. All the reasons stated here are spot on:
- You'll only hear it if you're recording highly dynamic instruments like orchestras and string instruments, and even then with high quality mics, preamps, and AD converters. If not, then by going to 96khz, you'll get the following negatives without any real benefit:
- higher CPU usage
- higher disk usage
- higher ram usage

I can easily see many reasons to use 24bit 44.1khz, but any higher, and that Gay Porn Union looks a lot better for the money. They have an excellent dental plan.

ok will stick with 32/44.1.

and .. why do u remind me of Lionel Huts?

[nebulae]

ok will stick with 32/44.1.

and .. why do u remind me of Lionel Huts?

Perhaps because I have a yellowish hue and a very large overbite

[sweetjesus]

very large overbite

you should look into that dental plan.

[Timur]

There is a bug in the Kontakt 2 engine concerning reverb when using 88/96k. I posted about that on the NI forum:

http://www.native-instruments.com/forum ... hp?t=54495