16 bit with compression VS 24 bit

[django6]

I'm looking to learn more about the best practices for tracking a live band to pro tools. If I have 8 tracks of live sound, it would be nice to maximize the input levels, but not have to worry about clipping. I could use compression, but it appears that plug-in compression can only be used on pre-recorded material, not be recorded to the track. I can't afford to buy 8 high quality external compressors. I read others recommending using lower levels and 24 bit. How are others handling this scenario?

[albee1952]

24 bits and lower recording levels are the way to go(with brick wall limiting just in case). By recording at 24 bit, you can set your input levels so you just tickle the yellow on the meters, Remember, the old days of slamming the input to the max was all about tape compression and getting above the noise(tape noise). You don't have either of those things to worry about now.

[django6]

What do you use for the brick wall limiting? Is it possible with internal plugins or must it be external?

My understanding is that with 16 bit you still want to get the highest input levels short of clipping to get the greatest bit depth resolution.

I never heard about just "tickling" the yellow with 24 bit recording, but I guess that makes sense since you'll still have a high resolution recording even at lower levels. Does anyone else have opinions on how high to record levels?

[albee1952]

Successful brickwall limiting would need to be done on the way in, so plugins aren't going to help if the input A>D converter is clipped(another reason to keep the recording level down to just into the yellow).

[daeron80]

Nope. Albee's right on. Why track at 16-bit when you have 24-bit capability? It gives you a theoretical 256 times greater amplitude resolution at the cost of only 50% more disk space. In practice the resolution is greatly improved, if not the full 256x.

Plus, when you push your converters hard (over about -6 dBfs), they don't sound as good. The analog stages start to run out of headroom and crimp the transients. At that point, you're giving up far more than you're gaining with that extra one bit of resolution. Remember 6 dB = 1 bit. It's actually like 6.02... but that's so close to 6 that it's seldom worth worrying about the remainder. So with 24-bit you can have peaks below -24 dBfs and still have greater resolution and wider dynamic range than if you had recorded at 16-bit pushing the -6 threshold. It's just crazy not to when you have the choice.

[AfterJohn]

Record 24 bit files with PLENTY of headroom. Your master fader should peak at the most around -6 dB. You'll have so much more to work with when it comes time to mix.

[barters81]

Originally Posted by daeron80

Nope. Albee's right on. Why track at 16-bit when you have 24-bit capability? It gives you a theoretical 256 times greater amplitude resolution at the cost of only 50% more disk space. In practice the resolution is greatly improved, if not the full 256x.

Plus, when you push your converters hard (over about -6 dBfs), they don't sound as good. The analog stages start to run out of headroom and crimp the transients. At that point, you're giving up far more than you're gaining with that extra one bit of resolution. Remember 6 dB = 1 bit. It's actually like 6.02... but that's so close to 6 that it's seldom worth worrying about the remainder. So with 24-bit you can have peaks below -24 dBfs and still have greater resolution and wider dynamic range than if you had recorded at 16-bit pushing the -6 threshold. It's just crazy not to when you have the choice.

Great explanation. Thanks for this it makes what I vaguely knew about crystal clear.

[Bubblegum]

24 bit is the way to go. Why not if you've got it? much lower noise floor too (lower than pretty much any analog device is capable of going).

You can get away without brickwall limiting if you track low enough. If it's a live recording with unpredictable levels, there's no harm in having your loudest peaks at -18dbfs, so get the drummer to kick his bass drum louder than he could possibly hope to do during an actual performance for a reference peak. You'll still be well above the noise floor of your converter.

The hotter you record signals into digital,the more you are actually raising the level of noise generated by your analog gear (mic preamps etc.) and it all adds up when you have several tracks, so have them running where they were designed to sound best and least noisy then once you finish the mix and (most likely) use any compression/limiting at the end, the less noise you'll have.