24-bit recording?

[bohunk06]

Another question for you Cubase pros. I have always recorded in 16 bit because I know that is the format you need to burn on to a cd. Is it better to record in 24 bit and with a higher sample rate and then dither down (I think thats the right term)? If so, how do I go about getting it from 24-bit to 16-bit. Is this even something I should worry about? or just stick with my 16-bit recording?

Thanks guys!

[Waffleness]

My opinion is that unless you have the best microphones on the market, a console worth tens of thousands, lush DA converters, a properly analysed and treated room plus a top range monitoring system, 24-bit recordings are a waste.

All its going to do is eat your CPU power. You might be able to tell the difference if your an audiophile with the best headphones+amp around, but the phrase "You can't polish a turd" comes to mind.

[Dogbreath]

Even though you're going to 16 bit for CD, 24 bit tracking and mixing gives you way more room to get a cleaner, punchier, crisper CD.

It's not like the difference is 8. (24 - 16) It's exponential. I don't remember the numbers but it's boo-koo thousands of bits difference between 16 and 24 bit.

Like noise floor. Tracking "as hot as possible without clipping" kind of applies more to 16 bit than 24. You're looking for the best signal to noise ratio.
The problem is that you've already eaten up a bunch of the headroom that you're gonna need at mixing (EQ, processing) and mastering. Whether you do the mastering or someone else does...the headrooms gone. Or nearly so.

24 bit gives you exponentially more room. So you can track at lower levels, mix at lower levels and punch it up at the last stage. That gives you a cleaner, punchier and crisper CD.

Hope it helps.

[Farview]

Originally Posted by Waffleness

My opinion is that unless you have the best microphones on the market, a console worth tens of thousands, lush DA converters, a properly analysed and treated room plus a top range monitoring system, 24-bit recordings are a waste.

All its going to do is eat your CPU power. You might be able to tell the difference if your an audiophile with the best headphones+amp around, but the phrase "You can't polish a turd" comes to mind.

This is complete nonsense. The difference between 16 bit and 24 bit is night and day.

You should always record to 24 bit. If you need to mix it down to a 16 bit file, just choose 16 bit when you export the mix. If you are really worried about dithering, put a dithering plugin on the master buss.

Higher sample rates are not necessary.

[bohunk06]

That does help. Thanks for the replies. So if I were to record in 24 bit in Cubase Studio 4, what is the proceedure to get that back down to 16 bit to put on a cd? Is that done in Cubase or other programs?

Thanks!

[NYMorningstar]

After you record keep it in 24 bit while doing all your processing. Export your mix to a 24 bit stereo interleaved file for mastering. If you're doing your own mastering, then import the 24 bit stereo file, process in 24 bit and then export it to your 16 bit file for cd burning.

[Dogbreath]

I found this article over at Tweakz...

http://www.tweakheadz.com/16_vs_24_bit_audio.htm

about halfway down spells it out.

sums it up rather nicely, I think.

Peace mang.......Kel

[LemonTree]

Originally Posted by Waffleness

My opinion is that unless you have the best microphones on the market, a console worth tens of thousands, lush DA converters, a properly analysed and treated room plus a top range monitoring system, 24-bit recordings are a waste.

All its going to do is eat your CPU power. You might be able to tell the difference if your an audiophile with the best headphones+amp around, but the phrase "You can't polish a turd" comes to mind.

Absolute horse bollocks!

[suprstar]

Originally Posted by Dogbreath

Like noise floor. Tracking "as hot as possible without clipping" kind of applies more to 16 bit than 24.

IDK about that, whether you do that in 16 or 24 bit, you'll still have to pull all your faders back to -15 right off that bat, so you might as well track to -15 to begin with. If you track right up to zero, your noise floor gets amped up, then you'll pull the faders back, and the noise floor will be right where it woulda been if you tracked to -15 and left the faders at unity..

The location of that noise floor is the real difference between 16 and 24 bit. Always use 24 bits if possible.

[Dogbreath]

Good point man.

So maybe it applies more to analog than digital?

C'mon...throw me a bone here...

[bohunk06]

Thanks for all the replies. The article from Tweak was very helpful. I am going to try 24/44.1 and see if that makes a difference. I have a pretty good ear for music and as the article said, can hear the difference between mp3 and wave. So I am hoping that This will clean things up for me.

Thanks all!!

[NL5]

Originally Posted by suprstar

IDK about that, whether you do that in 16 or 24 bit, you'll still have to pull all your faders back to -15 right off that bat, so you might as well track to -15 to begin with. If you track right up to zero, your noise floor gets amped up, then you'll pull the faders back, and the noise floor will be right where it woulda been if you tracked to -15 and left the faders at unity..

The location of that noise floor is the real difference between 16 and 24 bit. Always use 24 bits if possible.

You should never track "right up to zero", and what you are describing is not the same thing. You SNR is gonna be determined by the levels at which you record, not where the fader is in the mix. Now, a lot of home studios are getting the bulk of their "noise" from the room itself, so this may render the gears SNR unimportant.

[Dogbreath]

and as side note bohunkdude...I track and mix at 24/44.1 all the time. You should be good to go.

[Waffleness]

Sorry if I got it wrong.

*shoots self*

[suprstar]

Originally Posted by NL5

You should never track "right up to zero", and what you are describing is not the same thing. You SNR is gonna be determined by the levels at which you record, not where the fader is in the mix. Now, a lot of home studios are getting the bulk of their "noise" from the room itself, so this may render the gears SNR unimportant.

Right, I know. I'm saying that if you track hotter, you're running your pre harder, it's gonna introduce more noise to the track anyway. SNR stays the same regardless of where your gain is set at, no? Your source sound is at a constant level, the track's noise is constant, I'd think SNR would be constant. You could get a better SNR by turning the pre gain down, and playing/singing louder, altho that would likely affect your tone..

[Tim Gillett]

Originally Posted by Farview

This is complete nonsense. The difference between 16 bit and 24 bit is night and day.

You should always record to 24 bit. If you need to mix it down to a 16 bit file, just choose 16 bit when you export the mix. If you are really worried about dithering, put a dithering plugin on the master buss.

Higher sample rates are not necessary.

Yes the difference is 48db for a total of 144db. But it's only fully realized when the entire signal chain has a noise floor also around 144db below its highest captured level.

But AKAIK there is no amplifier or DA AD conv which can achieve that s/n. And no human ear either.


Sure, record at 24 bit if you can but the recording, in terms of noise, is only as good as its weakest link.

That weakest link may be the mic/pre noise, room noise, converter noise, the dog barking down the street, or a combination of the above. As a minimum, your converter will be considerably noisier than 24 bit.

In that sense, no one is recording at 24 bit in real world terms.

Cheers Tim

[Farview]

Originally Posted by Tim Gillett

Yes the difference is 48db for a total of 144db. But it's only fully realized when the entire signal chain has a noise floor also around 144db below its highest captured level.

But AKAIK there is no amplifier or DA AD conv which can achieve that s/n. And no human ear either.


Sure, record at 24 bit if you can but the recording, in terms of noise, is only as good as its weakest link.

That weakest link may be the mic/pre noise, room noise, converter noise, the dog barking down the street, or a combination of the above. As a minimum, your converter will be considerably noisier than 24 bit.

In that sense, no one is recording at 24 bit in real world terms.

Cheers Tim

But the noise floor is being modulated by the audio and the resolution goes up in more steps with 24 bits. All the bits 'below' the noise floor are not useless.

[Farview]

Originally Posted by bohunk06

That does help. Thanks for the replies. So if I were to record in 24 bit in Cubase Studio 4, what is the proceedure to get that back down to 16 bit to put on a cd? Is that done in Cubase or other programs?

Thanks!

When you export the mix, just choose 16 bit. That is all you have to do.

[Tim Gillett]

Originally Posted by Farview

But the noise floor is being modulated by the audio and the resolution goes up in more steps with 24 bits. All the bits 'below' the noise floor are not useless.

Sure, I agree, but we only need a bit depth deep enough to not add any significant extra noise to the existing noise floor, typically of the converter, but in practice, whatever is the noise floor. Bit depth beyond that achieves no practical purpose.
In the same way, at the other end, "achieving" a headroom of 20db in a given track achieves nothing in audio terms. All it says is "I avoided clipping by 20db".
Adequate s/n of a recording system relates in practice to the dynamic range of the material we record and so it varies. It's not set in stone.

[Farview]

But unless the noise is DC, it still has to cross the 0 volt mark. That point uses the least significant bit. That is where the added 'resolution' is.

[Tim Gillett]

Again, you are assuming the dynamic range of what you are tracking is in excess of what 16 bit can handle. Again, it aint necessarily so.