When dealing with PCM, you might as well stay in the 44.1kHz realm. There just isn't any real practical reason not to. Our automobiles would run better if we were using nitro methane but we don't. Neither should you. It's not practical. Why do people and other audio folks always want to be impractical? You are not impressing anybody with your lack of knowledge being impractical. Your releases will still be on CD, MP3, i-Pod, for years to come, which really don't care about higher resolution blah blah. You don't even need 24-bit unless you are a 2 bit engineer. 96 DB of dynamics processing power is more than most of us analog engineers ever had. And we still made recordings that to this day blow people away. So save your church and your congregation some money for more humanitarian purposes and go with the flow of 16-bit 44.1kHz like most of us professionals do that are practical people.

A terribly practical woman who is practically terrible
Ms. Remy Ann David

LOL- great thanks for the input!

It's very simple for me; 24/44 for most audio projects CDs, etc. No matter what I've read or heard, in my experience there is NO significant advantage to recording at 48k and then gearboxing it down to 44. (I don't pretend to hear the 2 db of difference at 22k in the stero soundfield that 48k supposedly yeilds, or whatever they claim it's better for...)

The potential loss and resultant corruption of sound during gearboxing does not, IMHO, give me any advantage over the supposed "Better" sound of 48k.

Conversely, I use 24/48 (and then dithered down to 16/48) for video projects: DVD soundtracks, etc. Before anything goes to the video software (Sony Vegas in my case), I let Sequoia/Samplitude do the dithering from 24bits to 16. I just like how it sounds this way.

But just to complicate things further, several radio and tv clients now have digital audio consoles and servers, which all run at 48k. What to do?

Since these projects will utlimately end up at 44k (CDs, etc.) I stick with the 44k rate and let them do the conversions. (They do the same thing for all of the 44k CDs they play on the air, so I'm not alone in this....)
Although a project may air once or twice on the radio (SR converted up to 48k), it will ultimately end up on someone's shelf in long-term life as an audio CD, so we stick to 44k.

What they do with it afterwards - MP3s, etc., I can't be overly concerned with, as long as I've done all I can on the front end.

Interesting. Thanks for the reply!
Does anyone use 96k regularly?

I have read many articles on the 96k versus 192k debate with hardware designers, but I guess I should have asked this basic question a long time ago.

Link removed

This article talks about the slowest sampling rate being around 44k for 22kHz frequenices (Nyquist). I read as well that NASA at one time believed that 60kHz would be enough, based on Nyquist ideas and real world conversion issues.

Anyway very interesting responses. Thanks very much!

The original 3M 32 track digital machine on 1 inch tape was 16 bit at 50kHz sampling. Said to have been the best sounding digital machine for that reason alone. And I think, we had to use a 12 bit and a 4 bit analog to digital and digital to analog converters because there were no 16-bit converters at that point in time.

Don't see those floating around much any more. Impossible to maintain.
Ms. Remy Ann David

I cant hear the difference between 44.1 and 48 but I can hear the difference between 44.1 and 192. can anyone say whats wrong with recording at the highest quality possible. I may be naive but I would like to see new high quality consumer audio products start poping up the way it is with video.

Go ahead and cut me down if you must, I wan't to be set straight.

Grizzly,

I think the most damning evidence against recording at the highest rate possible is the HD space. Even at 48k it doesn't take very long for a large HD to fill up. If we (my studio) recorded at 192 we'd be screwed in a matter of days.

88.2 if possible. Downsamples better to 44.1!

Anything special considerations for down sampling?

If i'm recording concerts I always record at 48k 24bit, as I know it will most likely end up with a video track on a DVD. Otherwise I stay at 44.1k 24bit.

HD space is getting really cheap, and I've got all the processing power to work at even 192k 32bit 100+ tracks, so maybe I'll be trying that out for some sessions shortly.

∞khz

Since you'r running 24 tracks you should choose between 44.1KHz/24bit and 44.1KHz/16bit. Both are accepted standards.

I use 44.1KHz/24bit for most projects. I'm a two bit engineer ;-)

24 bit gives you a higher dynamic range so it is easier to avoid overs. If you are confident it won't happen, you can use 16 bit.
44.1KHz is all you need for CD and any lossy format (wma, mp3 etc). To save a bit of harddisk space and processing power I wouldn't go any higher.

48KHz or multiples thereof are nice with 24fps video. If you don't sync with video there is no compelling reason to use 48KHz.

I'm on 44.1

Thanks Guys! I think I can hear a difference between 96kHz and 44.1kHz. But Maybe I am fooling myself.

OK,

interesting stuff so far, but if I may stray away from the original question for a sec:

Say you record tracks using 44.1/16 bit. You then want to do some detailed editing on some of the tracks, such as offline noise reduction, etc. I'm kind of referring to software like CoolEdit/Audition.

Surely this kind of editing is done in 32 bit - if you're audio is 16 then it's converted to 32 then back. Is this correct?

If so, is there any loss in the sound quality? Isn't better to start off with a higher resolution at the recording stage, before any such edits, and then 'moving down the ladder' to 16?

CoolEdit/Audition just like a lot of those editors work in 32 bit float. I forgot the exact explanation, but it means something like an intelligent comma placement to keep rounding off to a minimum which avoids artifacts. This is only for the editing itself. Your file will in this case stay in 16 bit.

Yeah,

never quite got the idea of the 32 bit float thing...

So it makes no difference if you record in 32 bit or 16 before editing within Audition?

Jaike wrote: never quite got the idea of the 32 bit float thing...

It's always good to read it again:

PowerStudio

No, I can definitely hear differences between 44.1khz, 96khz, and 192khz. It's just very subtle.

I can hear a difference. But I don't think it's because I can hear a 30 or 40khz tone, but because the roll-off filter is much steeper on 44.1khz than on 96khz, and there for the high sounds are much clearer, even when I've downmixed it to 44.1Khz

Interesting. I know I can't hear beyond 20k, but I also know in the real world Nyquist–Shannon sampling theorem, which requires a sampling rate of at least 2 times the frequency, as some issues.

“Exact reconstruction of a continuous-time baseband signal from its samples is possible if the signal is bandlimited and the sampling frequency is greater than twice the signal bandwidth”

So you need to sample at least at 40khz to get a 20kHz. However with interploation error and alaising the critical frequency (half the sampling rate) must be higher than the highest frequency required.
44.1k/2 = 22.05kHz

20k < 22.05k

However it does make sense the higher the sampling rate, the more acurrate the digital representation is going to be. But the trade off is the storage.

All you math geeks check out this for more:

http://en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampling_theorem

For Music projects, I record at as high a sample rate and bit rate as possible. I let the Mastering Facility playback the final mix at the higher res then reconvert to 44.1 16bit through their higher end converters. I also have them do them same process at 24bit and save that master for the future in case the standard ever changes to a better quality than 16 / 44.1

What it bit depth and sampling rate do you use?

24 bit, 44.1. Don't know when I'll need to change. 24 gives me plenty of headroom and 44.1 is where I end up anyway so why put the project through an extra step of mathematical conversion.

With my HD24 it's come down to the speed of the hard drive I'm using. If it's fast enough and has a big enough cache onboard it hopefully won't bog down at 24 tracks and 44.1. I would love to record at higher resolutions but just can't afford the $20K upgrade and with my tinnitus I don't think I would truly appreciate the difference anyway. Right now all my stuff gets cramed into cd's anyway.

Remy thanks for making me laugh everytime I drop in here!

Hey guys Im recording an album at the moment at its 44.1/24bit and my mate whos a producer also was like man you should record at 48 but i think while its great to have amazing production which im into ..if the songs are shit they will still be shit just at a higher bit rate.
I think production is an artform but ive seen guys make great albums with shit equpiment so i think its really done to the person and not as much the bit rate.
well thats my two cents

Cool Thanks!

I think its a good idea to keep one copy of the track at highest quality if its a big record because this means more sales when better technology comes along as people would like to play the music on that.

I record everything in 24/48. Theres a bit more dynamic range (so it seems) and with all the acoustic stuff that goes down I think it makes it easier to separate at mix.

That being said, if you're rendering to 16/44.1 and you work carefully with the levels it shouldnt make a difference at all.

I do want to record an album at 192 sometime and just see what it sounds like............

I have done stuff at 96kHz but nothing at 192kHz yet. My hedd only goes up 96kHz, but I would be intrested to try it.

Jaike wrote:
So it makes no difference if you record in 32 bit or 16 before editing within Audition?

Well, the editing in 32-bit float will do VERY little damage to a soundfile.

To my knowledge there is no converter that works with a bitdepth of 32, so I guess no one records in 32-bit.

Imagine: 20xlog(2^32) = 192,66 dB dynamic range. Outra(n)geous!

There was a discussion some months ago here about whether 32-bits was a viable bitdepth to work with. I lost track of that thread though.

192.66dB for us north american geeks.

You should record at the HIGHEST sample rate and bit depth possible. If your system can support 192, 24bit, DO IT! The reason is this: the Nyquist theorem says that the sample rate must be twice the highest frequency recorded. Now, even though we can only hear to around 20kHz it doesn't mean that higher frequencies don't exist. And what these higher frequencies do for us is give us a sense of space. As for the bit depth, why not use the highest one you can? All it's going to do is make the quiet parts seem quieter and the loud parts seem louder. That sounds like a good reason to me. There is one more idea that needs to be addressed. That is the use of dither. This is what is used to bring the digital recording down to a sample rate and bit depth used for digital media i.e. CDs. If you are recording something for a CD, use a sample rate that is a multiple of 44.1 (44.1, 88.2, etc.). This gives a lesser chance of artifacts in your recording. If you are recording for a DVD, use a sample rate that is a multiple of 48 (48, 96, etc.). I would use 88.2kHz 24 bit personally. I hope this post made sense and helps you out.

Welcome! Thanks for the comment!

My understanding is dithering only affects the bit depth not the sampling rate.

However what you say was talked about previously in this thread.

The real question is:

How much hard drive space and processor power are you willing to sacrifice for better resolution? And what is the point of diminishing return?

A lot of professionals here say anything above 44.1kHz is beyond that point.