Should I be recording in 24/96?

[curtiswyant]

I've got a Delta 1010 and I've never bothered to change the settings from 16/44. Just wondering if, since I can, should I be using the card to its full potential at 24/96?

[MadAudio]

You should at least up the bit depth to 24. With 24-bit, you don't need to record as hot as 16 bit, and IMO it sounds better. I also think 96k is overkill for most home recording, and besides it eats up drive space.

[reshp1]

What he said, 24bit gets you 256 times the resolution for 50% increase in HD space, CPU load. That means you don't have to really worry about getting a "hot,"near 0dB signal into the DAW, it'll sound just as good in 24 bit if you signal is say -6 dB.

the 96k debate has been done to death, but most people agree that they can't hear the difference, although if you do up sample rate, it's better to go to 88.2k, since the math is simpler once you inevitably have to come back down to 16/44.1 to put on CD.

[TimOBrien]

Definitely 24bit.

96k will severely reduce the number of tracks you can run and the number of plugins you can run. And eat up hard drive space.

And you still have to go down to 16/44.1 for CDs....

[XLR]

I can hear a big difference between 16 and 24 bits. But not so with the sample rate difference.

Tim

[boomtap]

I have always beleived that unless you have all of the other equipment to back it up (high end mics, clean power, and great room) then recording at 24/96 isn't really going to gain you anything. Even some of the Class A studios I go into with millions in gear feel that 24/96 isn't that special escpecially with the way mastering to high levels has ruined the overall fidelity of recordings. They always opt to eliminate all conversions. If you are going to a 16 bit final product then record in 16 bit. That is at least what I have bought into. Besides everthing gets mp3ed now adays, the only one who will apriciate the full size awsome unmastered recording will be you.

[XLR]

unless you have all of the other equipment to back it up (high end mics, clean power, and great room) then recording at 24/96 isn't really going to gain you anything

You're making a good point. With lower quality gear and rooms money is possibly better spent to improve the front end or the recording space. Also I think it's important to recognize that sound quality of converters is determined to a large extent by other things than wordlength and sample rate.

Tim

[Janesaid2me]

Def 24 bit its huge diff between 16....the sampling rate like others have said is hard to tell a diff...

[reshp1]

Originally Posted by boomtap

If you are going to a 16 bit final product then record in 16 bit.

I don't agree with this. In the final product you will have mixed it such that you are utilizing the full dynamic range. When recording at 16bit, you'd have to record everything as hot as possible to acheive this. What you are most likely doing is recording a track and then upping the gain later in the mix, either with the fader or compression, which further emphasizes the quantization error of low bit size. It's much better to record at 24 bits, because even at very low recording levels you have better resolution than you end format. You can they manipulate and add gain without fear of large quantization errors.

[Farview]

24, yes. 96, not so much.

[HangDawg]

Originally Posted by boomtap

If you are going to a 16 bit final product then record in 16 bit.

Please don't propagate garbage like this just because it's what you've bought into.

[NL5]

Originally Posted by reshp1

the 96k debate has been done to death, but most people agree that they can't hear the difference, although if you do up sample rate, it's better to go to 88.2k, since the math is simpler once you inevitably have to come back down to 16/44.1 to put on CD.

I could swear that i read a white paper on this, and that is completely wrong. There is actually a problem doing that, in that it causes artifacting that is eliminated by converting from 96 instead of 88.2 - does anybody remember that article - I'm pretty sure it was posted on here.

Anyway, I record at 24/48.

[HangDawg]

I think the MATH involved with sample rate conversion from 88.2 to 44.1 is more than just dividing by 2.

[reshp1]

Originally Posted by HangDawg

I think the MATH involved with sample rate conversion from 88.2 to 44.1 is more than just dividing by 2.

What more would you have to do besides drop every other sample? Maybe I'm just oversimplifying it, is there a "dither" for down sampling?

[apl]

Here's a useful article. The gist is take advantage of 24 bit when recording, and use the sampling frequency that your final product will be.

[amra]

I read an article the gist of which was that you needed approximately twice the amount sampling resoultion that you are trying to capture. I don't remember the details but I will try to find the link. Has anyone read or heard this before?


Amra

[Farview]

Originally Posted by amra

I read an article the gist of which was that you needed approximately twice the amount sampling resoultion that you are trying to capture. I don't remember the details but I will try to find the link. Has anyone read or heard this before?


Amra

That is that whole Nyquist thing that was discussed earlier.

[MadAudio]

Amra - you're talking about the Nyquist theory which states for lossless audio your sampling rate should be twice the range of human hearing. It has nothing to do with bit depth.

And I spent so long typing this that Farview beat me to it!

[bennychico11]

96k?? ha!! that's for pussy stuff
i do all my recording at 192!!!! I mean, can't everyone hear around 96,000Hz??? hmmm, maybe it's just me.

[giraffe]

yes 24 bit

i'm not 96 capable but i really like the idea of it because it is the point at which thare are 5 points on a 20k wave. i know not everyone can hear that hi, but the math looks good.

[boomtap]

Originally Posted by reshp1

I don't agree with this. In the final product you will have mixed it such that you are utilizing the full dynamic range. When recording at 16bit, you'd have to record everything as hot as possible to acheive this. What you are most likely doing is recording a track and then upping the gain later in the mix, either with the fader or compression, which further emphasizes the quantization error of low bit size. It's much better to record at 24 bits, because even at very low recording levels you have better resolution than you end format. You can they manipulate and add gain without fear of large quantization errors.

That is interesting and probably all true, but when I do this side by side in a class A studio, I just don't get this amazing recording when compared to 16bit. By the time it gets to mastering you just can't tell that the mix was 24 bit. I think that the very best equipment with the very best engineers, and the very best mastering house might be able to squeeze some extra out of 24 bit, but for the average project going to CD I just don't think you will be able to. I say do it if you want to, no big deal it won't hurt, I just don't think you will be gaining much.

If I polled 100 people and took 2 song versions 1 16 bit and 1 24bit(whole process), I bet that 100 of them could not pick out the diffrence. I mean people listen to music in thier cars, and on crappy headphones that don't reveal the clarity, and the rest of them can't tell FM from CD. Shoot AAC and MP3 rule the world now adays.

In short 24bit is better, but it is so minor that nobody cares. Until they can send sound direct to your brain with no sound pressure or cables, I think we are not going to see overwhelming advancement in sound technology.

[amra]

Amra - you're talking about the Nyquist theory which states for lossless audio your sampling rate should be twice the range of human hearing. It has nothing to do with bit depth.

Sorry, I meant sampling rate, not resolution. I wasn't talking about bit depths.
In any case I found the article I mentioned, and it is pretty interesting and explains the nyquist effect in terms us mere mortals can understand:

http://www.digitalprosound.com/Htm/S...ap2_Apogee.htm

[Bob's Mods]

I believe the current wisdom is to record in the sample rate of the final product.

Bob

[ds21]

Originally Posted by boomtap

That is interesting and probably all true, but when I do this side by side in a class A studio, I just don't get this amazing recording when compared to 16bit. By the time it gets to mastering you just can't tell that the mix was 24 bit. I think that the very best equipment with the very best engineers, and the very best mastering house might be able to squeeze some extra out of 24 bit, but for the average project going to CD I just don't think you will be able to. I say do it if you want to, no big deal it won't hurt, I just don't think you will be gaining much.

If I polled 100 people and took 2 song versions 1 16 bit and 1 24bit(whole process), I bet that 100 of them could not pick out the diffrence. I mean people listen to music in thier cars, and on crappy headphones that don't reveal the clarity, and the rest of them can't tell FM from CD. Shoot AAC and MP3 rule the world now adays.

In short 24bit is better, but it is so minor that nobody cares. Until they can send sound direct to your brain with no sound pressure or cables, I think we are not going to see overwhelming advancement in sound technology.

But what if you want to archive for the future or even DVD, do you really want to convert it up from 16 to 24, or at least start with 24 and be safer, that is unless you don't have the HD space or CPU to handle it, that's the only rational reason to go with lower resolution. Of course this is MHO.

[altiris]

what about 32 bit floating? Is it better to record a 32 bit, 32 bit floating , or 24 bit rate? Cubase default is 32 bit floating. Should I change it to 24 bit. If so, why?