Recording in 24bit

Any tips

Hey guys,

All the pieces of my new setup have now arrived … still in pieces on the floor at the moment, but soon to be a shiny new studio.

Consequently, I have made the move from 16 to 24 bit recording, and wondered if there is anything I need to watch out for in recording 24 rather than 16.

I assume that conversion from 24 to 16 for putting on CD is the very last thing you do before burning, but any other tips would be greatly appreciated.


Hi Dave, and welcome to the semi pro world of 24 bit. I think you’ll find that there is way more headroom than with 16 bit. I wouldn’t worry so much about going higher than 48kHz, though. Especially if you don’t have a real bad*** puter. The increased file sizes created don’t necessarily correlate to a better mix.

More headroom to me also means not needing to track everything at a louder volume to get a good sound. Your mixes should get much better (and clearer) now that you can hear more of the signal.

That’s a pretty lo-tech answer, but if you do a forum search for 24 bit, you’ll find lotsa detailed info, though.


Ok, what I’ve found is 1. Higher Headroom - you can really crank it 2. Lower noise floor - quiet parts sound more quiet! 3. Your archiving of mixdowns can be at a higher quality then what you are putting on CD

Just remember to dither to 16 bit when you are rendering/mixing down for a CD or coverting to MP3

I do not agree with t9c about not going higher then 48khz. If you can get to 96khz you will find your higher frequencies sounding much more crisp rather then kinda harsh. And my oh my does it sound good. Has to do with Nyquist… (this is not Nyquist but relates to it) it has more samples to capture those higher frequencies and not so harsh a filter to avoid aliasing.

I agree record the highest you can… google audio dither and noise shaping as well for when you do your mixdown. Ive not recorded at 96k Ive used 88.2 and 48. 88.2 mixes down to 44.1 in seconds instead of minutes like 48 does.

Ouch Squid, your grammar is hard to understand


I agree record the highest you can

This is one thing I almost halfway disagree with. I mean, it is good advice, so don’t take my disagreement that way.

When recording 24 bits it’s not nearly as necessary to record at levels that you would when recording 16 bits. There is so much more headroom, and more importantly noisefloorroom (130 something db down compared to 86 db or so 16 bits?), that it isn’t necessary to be up near 0 db, or even -6 db. It is when doing 16 bits.

When I record drums 24 bits I have the peaks in the -10db ballpark, and some tracks are even little lower (not that they need to be). By doing that there is plenty of headroom for those moments that the peaks just happen to be much higher for very short time. The attacks of drums can be WAY up there and unexpected. By keeping the “normal” attacks below -10db or so there is little chance of accidentally hitting that magic 0db digital crapout point. It’s quite easy to hit a rim shot that sends a peak way up there. It’s better to not get there. 24 bits has so much range that there’s plenty left over when the volume is a little lower than usual.

This is REALLY cool when you start adding up more and more tracks and don’t have to compensate so much to keep the overall volume below 0db.

6dB per bit, so 16 bit = 96dB, 24 bit = 144dB.

(Every added bit doubles the range; and doubling voltage-wise, is 6dB. (actually, nearer 6.0206 :cool:))

But that’s the theoretical best, and you’ll never get the analogue side, or the A/D’s anywhere near 144dB. (Or your ears :D).

And as you say Dave, dither down as the very final step, don’t even normalise after dithering down.

As for the 48kHz vs 96kHz or higher debate, there’s been a lot of discussion on here in the past about that.

Some people claim there’s no audible difference, some people claim the stereo imaging and transient “sharpness” is audibly better at higher sampling rates. Try both and see for yourself. :)


As for the 48kHz vs 96kHz or higher debate, there’s been a lot of discussion on here in the past about that.

I and others can hear the difference, however 96kHz puts a strain on your DAW. All studio work I do is in 96/24, however when I record my band live I do 44.1/24 and that seems to be quite acceptable.

My highly unscientific amateur hack take on this is that the source material decides how high you need to go.
Recording a mic’ed distorted electric guitar, I have yet to hear the sonic benefits of higher bit/sample rates (added headroom notwithstanding). If you happen to track a grand piano under pristine circumstances I suspect you should go as high as you’re able.

I did some A/B’ing with some Firebird (Gibson, not Gretsch) through Marshall recordings done @ 16/44.1 and 24/44.1 and on my setup the difference surprisingly was in the low(ish) mids, not in the high end as I expected.

Pro people tell me that cymbals @ 192 khz can’t be beat. Never heard it with my own lugholes though…


EDIT: There IS such a thing as too clean. Horse for courses and all that…

Thanks guys … some great looking advice in there as always.

Can’t wait to get playing around with it.


Quote (teryeah @ Nov. 12 2005,19:13)
Pro people tell me that cymbals @ 192 khz can't be beat.

192khz sounds very nice indeed...

Keep in mind that when sharing files the bottom line is sample rate. Different bit depth files can be imported into a single song, but all the wave files MUST be the same sample rate, or they have to be converted into the same sample rate. In other words, everyone working on files for a project needs to be able to record at that same sample rate, whatever that is.

The 24 bit argument I’m all about. 24 bits is a great way to go - headroom is fantastic.

There are two points I would make about 96K:

1. 96KHz sampling does put quite a strain on your DAW - especially when applying effects. Whether you have some DSP effects (i.e. Universal Audio UAD-1) or native in-the-box effects, you can get quite a few more plug-ins in your mix at 48K than you can at 96K. There’s 1/2 the data to deal with.

2. Jitter is an issue with all A/D converters…and typically, higher sampling rates = more jitter. This shows up in all sorts of audible ways - stereo imaging, etc. Obviously, the solution to that is a great clock - but until I own a pro studio, the clocks on my Delta 1010’s are all I’ve got. And I find that they sound better to me at 48K than 96K.

So, my overall take on it is - use what’s appropriate for the source material and what sounds best to your ears. If you find you like the sound of your equipment best at 44.1 KHz / 16 bit, by all means, go for it. If it’s pristine at 192KHz / 24 bit - and your DAW can handle it (n-Track can) - then rock on.

But don’t be swayed by the theoretical arguments. Trust your ears, and your budget.



The 24-bit Field Recording FAQ

The article also discuss CD 16 / 44.1 quality or lack of quality as compared with 24 / 96, and about consumer systems e.g. SACD etc.

If you don’t have an SACD player (or burner for that matter), but do have a DVD writer then I guess one could record in 24 / 96 and burn to a DVD that’s playable on a home DVD player. For listening quality the sound would therefore be better quality than a standard CD player.

Or - you could even just play back your 32 bit n-Track mixdowns from your computer through your hi-fi / stereo set up, and just appreciate from there.

The 24bit faq has a lot of useful info.
It’s also a bit outdated (nov '01) here and there.

There’s a poll going on over at the gearslutz forum (Music Computers).

23 of 32 responders still use 24bit/44.1k. ???

hansje - yes you’re right, a bit outdated, but the theory is still there. I thought I had a good grasp of the differences between 16 and 24 bit, but now this faq makes everything clear to me :)