Sample rate vs bit depth

Posted: Mon Apr 16, 2007 7:17 am

Hi everyone, I am setting up a 16 channel studio and have a few questions hopefully more clever people can advise me on.
I hooked up two firepods but the sample rate dropped to 44.1/48 khz. (a) Is this bad, as I'm going to CD eventually. What is more important - sample rate or bit depth? (b) Should I set my cubase to 32 bit floating point ? (my mics are going directly to firepod and all mixing and effects are done in cubase, and I'm using my cheapie desk to create four headphone sends). (c) Should I look for a 16 channel 96khz setup? I still like the presonus and would get a firestudio and digimax fs, I think the preamps are better on these units as well. Any comments and advice greatly appreciated

Posted: Mon Apr 16, 2007 9:04 am

Everything ends up at 44.1khz/16bit for CD's
It is possible to get slightly better sonic results at higher sampling or bit depth, although you're only as good as your weakest link, so in some cases it could end up being overkill.
Higher rates will absolutely require more drive space and archiving time.
Plan your life acordingly

Posted: Mon Apr 16, 2007 10:03 am

There have been several threads on this recently, so a search will give you lots of info (and a lot of divergent opinions). I'm for recording at 24/44.1. I want the bit depth to give me headroom, but I really can't hear any difference with the higher sample rates when you dither down to 44.1 to put on a CD.

[Edited for terrible grammar. It's a bad day in Blacksburg.]

Posted: Mon Apr 16, 2007 10:49 am

Steve,

In your circumstances, I would use the following:

24-bit 44.1 KHz sampling in the Firepods

32-bit (or 64-bit) floating point in Cubase

Dither to 16-bit on your final write of the stereo mix to CD, otherwise keep the intermediate mixes in F.P. or 24-bit PCM.

Note that we are all suggesting you stick with 44.1KHz sampling rate throughout.

Posted: Mon Apr 16, 2007 11:10 pm

Bah Humbug!! I'm still recording at 16-bit 44.1kHz without any problems. I always look at what the downstream product will be and that's 16-bit 44.1kHz for CDs, MP3 and Windows media not to mention i-Pod stuff. The 24 and 32-bit depth recordings give your software the ability to do all sorts of inappropriate, improperly adjusted, ridiculous level mismatche processing when adding clueless inexperienced effects and processing. Since I know what I'm doing, I know what I'm doing and I know I don't need anything more than 16 bit 44.1kHz as long as we stay with PCM based systems. When we finally move on to DSD (Direct Stream Digital) since it is becoming more affordable, everything will start to sound better and more like we used to get in analog but still with the overload problems just not the bandwidth limitations, brick wall filtering and phase irregularities.

I'm looking to re-lax, since I'm feeling a bit stuffed up
Ms. Remy Ann David who is generally full of it

Posted: Tue Apr 17, 2007 6:34 am

RemyRAD wrote:

....I'm still recording at 16-bit 44.1kHz without any problems.

.....

....but still with the overload problems...

I'm looking to re-lax, since I'm feeling a bit stuffed up
Ms. Remy Ann David who is generally full of it

Just picking up on a bit of self contradiction here.... :<)

Even when the target media is 16/44.1 or even less does not mean that advantage cannot be taken of higher resolution. The first of which is hinted at in this contradiction. With 16 bit recording I am sure you take great care with your levels but to prevent often horrible digital overs you must be making some compromise on dynamic range. Depending on what and where you are recording this may or may not be a concern.

Using 24 bit, although the full implied dynamic range is not available, there is still normally a useful improvement (usually 10-15dB) over 16 bit converters and this can be very effectively used to improve headroom whilst not compromising the dynamic range. There are other reasons but they require a more exhaustive mathematical analysis not suitable to hereabouts.

As just about all DAW's now process with 32 floats at a minimum, the impact on the mixing/effects process are more debatable but there are very compelling reasons (E.g. accumulation of rounding and truncation errors) to use higher resolution processing. That is probably why Pro tools use 48 bit fixed point and some other DAW's have 64 bit floats available don't you think?

Having said that the impact of all this can be totally lost to the human ear depending on the final material you are producing. If your programme material averages 15 dB below peak, I doubt anyone is going to hear an effect tail distortion or cut off, so who cares.

To me the limitations of 16 bit can be (and have been) demonstrated effectively elsewhere but that does not mean that excellent end results cannot be had from 16 bit recordings. It really depends to an extent on your final programme material. Personally I find that using 24 bit gives me a very useful amount of leeway when recording "live" musicians who always seem to find a little bit more gusto when performing a take than when giving me a level!

Posted: Tue Apr 17, 2007 12:36 pm

Yes, not to contradict myself but since I'm a professional who began many, many, many years ago and learned how to work within a 60 DB dynamic range and less, 16 bits is like a luxury vacation in dynamic range and 24 bits is like going to the moon.

Of course at 24 bits, it allows you to record at very conservative levels just to be ready when that 30 DB transient comes along. How many people do you know that are happy to make recordings where their "meters" are barely moving at all? Right! Hardly anybody. Now that's all very fine if you have also adjusted your microphone preamps/console accordingly with the same kind of conservative gain staging, which I rarely see. So with all this added confidence of 24-bit recording, comes more people who are learning how to create worse overload problems within their front end, preamp/mixer from poor gain staging, since most of the equipment being used has more of a limited dynamic range than the software/CPU combination that's capable of all this phony high-definition catchphrase these days. And that's what keeps people wondering why their recordings don't sound like the professionals recordings.

Impractically practical
Ms. Remy Ann David

Posted: Wed Apr 18, 2007 4:57 am

RemyRAD wrote:

Yes, not to contradict myself but since I'm a professional who began many, many, many years ago and learned how to work within a 60 DB dynamic range and less, 16 bits is like a luxury vacation in dynamic range and 24 bits is like going to the moon.

Yes, I very much appreciate that. I am also a professional of too many years and designed my first two amps with valves when those funny germanium things with three legs started appearing. Like I said though, it very much depends on the material and recording location. How many hiss free classical/romantic music recordings can you remember from that era? The point is we can now do better, as you say, even with 16 bit. The thing is we can do even better with 24 bit so why not - it's not exactly taxing the systems nowadays?

RemyRAD wrote:

Of course at 24 bits, it allows you to record at very conservative levels just to be ready when that 30 DB transient comes along. How many people do you know that are happy to make recordings where their "meters" are barely moving at all? Right! Hardly anybody. Now that's all very fine if you have also adjusted your microphone preamps/console accordingly with the same kind of conservative gain staging, which I rarely see.

Well I only suggested a 10-15 dB improvement with 24bit and only to use some of this for improved headroom. I would normally suggest (unlike some others) that you would only back off your gain staging by around 5-6 dB. That should give you plenty of extra confidence with overs.

RemyRAD wrote:

So with all this added confidence of 24-bit recording, comes more people who are learning how to create worse overload problems within their front end, preamp/mixer from poor gain staging, since most of the equipment being used has more of a limited dynamic range than the software/CPU combination that's capable of all this phony high-definition catchphrase these days. And that's what keeps people wondering why their recordings don't sound like the professionals recordings.

Impractically practical
Ms. Remy Ann David

Well there is never any excuse for not properly engineering things, whether analogue, 16 bit, 24 bit or DSD. I have to say that there are quite definite advantages to processing at higher bit depth than the recordings though - that's simple maths.

Whilst I quite agree with you that, in the right hands, excellent results can be had from 16 bit systems I do not see that as a valid argument for not adopting advances in technology. It simply means that we have the wherewithal to make even better recordings and final products - even IF the final medium is 16/44.1.

I think that the argument that some will be induced to make awful end products is a none starter. If they can't use their ears and do not have the expertise they would still produce rubbish with ANY recording system!

Posted: Wed Apr 25, 2007 1:45 am

Unfortunately, you are correct and so, yeah, rubbish. I too actually run my levels more like you described. I do think 24-bit is better than 16-bit, especially for those with contracts! Our simple home computers are more powerful than ever before but the 24-bit does substantially increase file size and reduce track count for those who do that have the most powerful machines. It's sort of like having a high-performance crotch rocket motorcycle. You really can't drive the way it was designed to be driven, unless you want to lose your license. Or, like playing a difficult musical passage or, learning how to run, you first have to take little steps or go slower before you are ready to blast off. After all, how will you know what overload is until you do?

Typically overloaded
Ms. Remy Ann David

This topic sponsored by: