16 bits or 24 bits?

[The Green Hornet]

Can one of the crew tell me if it "really" makes any difference doing a CD on 24 bits?

For some reason, I may not have tried using the 24 bits. Once I may have remembered that the 24 bits "won't burn" a CD. [Using my AW1600]

I get pretty clean sounds from the 16 sounds. But, I always like to ask when I like the information and learn.

Hope the golf season starts moving in THIS directions soon.

Thanks,
Green Hornet

[mixsit]

I presume you mean tracking, but ultimately going to cd.
16 does work fine, the cool thing with 24 is you can come in nice and clean with 20db below full scale of headroom and still have conversion error noise well below you mic or room noise.

[MessianicDreams]

Originally Posted by The Green Hornet

Can one of the crew tell me if it "really" makes any difference doing a CD on 24 bits?

Yeh, it makes a difference.

At 24bit, you're CD won't play in most commercial CD players!

[The Green Hornet]

Yo "MIX" and "M. D."


THANKS! Since my CDs from the 1600 Yam won't "play" in most CDs, like my car or home Sony player, I guess I'll have to do 16 bits.

Some great music MD. So much talent. Appreciate your help.

Green Hornet

[steve.h]

If you're tracking, 24 bits will get you higher headroom, and if you ever feel like burning to a DVD-Audio disc, you'll be able to at higher resolution (although it's likely you won't be able to tell a difference if you belong on this Home Recording forum!).

However, as you've been told, the Red Book CD audio standard is at 16 bits, 14.4khz: anything else won't play correctly.

[mixsit]

Originally Posted by The Green Hornet

Yo "MIX" and "M. D."


THANKS! Since my CDs from the 1600 Yam won't "play" in most CDs, like my car or home Sony player, I guess I'll have to do 16 bits.

Some great music MD. So much talent. Appreciate your help.

Green Hornet

Just in case- You do know the normal route is track and process (mix) at the higher depth, then reduce to 16 for the cd?

[Dogbreath]

Originally Posted by steve.h

, 14.4khz:

Hi Steve.

[doulos24]

i like 8 bits but im an odd ball

[Massive Master]

16-bit: Around 65,000 points of resolution, dither noise on each individual track.

24-bit: Around 16.7 million possible points, 48dB more downward headroom (*DOWNWARD* headroom - "foot room" if you will), no dither noise, dynamic range far beyond almost any piece of gear imaginable.

Always work in 24-bit until the last possible stage.

[Ironklad Audio]

^listen to this guy...

track, mix, and master at 24...then chop the files down to 16-bit at the end of the whole deal

[RAMI]

Not to disagree with you guys, because I don't disagree with you. I also record in 24bit until the last possible stage.

But I think Ethan Winter (or someone else like him that knows a lot more about recording than I do) insists that there is absolutely no difference between 16bit and 24bit. I'm pretty sure he did some experiments and has numbers to back that up.

Just throwing that into the argument. Hopefully Ethan (or whoever it is) will come in and explain their theory.

[SouthSIDE Glen]

Originally Posted by RAMI

ut I think Ethan Winter (or someone else like him that knows a lot more about recording than I do) insists that there is absolutely no difference between 16bit and 24bit. I'm pretty sure he did some experiments and has numbers to back that up.

He has some technically flawed but publicly believable tests that demonstrate that stuff (e.g. dither, noise) happening down at the least significant bit or two of the 24 bits are quite often inaudible and therefore irrelevant, but that does noting to change the fact that 24 bits gives one a nice wide 140dB canvas to work with with absolutely no dynamic worries and plenty of footroom to be able to structure the gain on the analog side of the chain nice and quietly, keeping the potential *analog* noise floor down.

Bottom line: The only "downside" to working in 24 bit is a 50% increase in file size from somewhat less than 700KB per minute to approx 1MB per minute of audio track. That's a small cost to pay for the 50% increase in digital canvas space.

G.

[Massive Master]

Not that I'd want to argue with Ethan about anything either - But I think his point is with "file A" and "file B" -- One is 16 bit, one is 24, can you hear the difference?

I'm sure few could hear the difference.

This is about stacking (perhaps dozens of) tracks together - Put 20 tracks at 16-bit (adding all that dither noise together) and you'll notice it.

GRANTED: It's still going to be less than tape noise and the DAW is probably going to be throwing calculations in 32-bit FP no matter what the word length of the source files. Add self-noise of gear and what not into the equation and you're "eliminating excuses" by recording in 24-bit. But IMO, well worth the additional horsepower.

[Waffleness]

Originally Posted by SouthSIDE Glen

Bottom line: The only "downside" to working in 24 bit is a 50% increase in file size from somewhat less than 700KB per minute to approx 1MB per minute of audio track.

Plus the associated CPU power that is required to drive a 24 bits signal chain.

[SouthSIDE Glen]

Originally Posted by Waffleness

Plus the associated CPU power that is required to drive a 24 bits signal chain.

That is true and is technically a "cost", you're right. But in today's world of multi-GhZ CPUs, that cost is a tiny one. It may mean the difference between being able to handle only 50 simultaneous tracks instead of 75 before hitting the CPU very hard, or something on that kind of order; usually comfortably still enough for all but the most sophisticated of projects to run without noticing a difference.

G.

[steve.h]

Originally Posted by Dogbreath

Hi Steve.

Ah fuck hah, I went dyslexic and nostalgic at the same time! (Thinking about modems apparently).

44.1 it is!

[Ethan Winer]

Originally Posted by RAMI

I think Ethan Winter (or someone else like him that knows a lot more about recording than I do) insists that there is absolutely no difference between 16bit and 24bit.

No, there is a difference - 24 bits uses half again more disk space and CPU power.

Seriously, if I were recording a classical music concert I'd definitely record at 24 bits. No matter how loud an orchestra (or jazz band etc) plays during rehearsal, they always play louder when the red light is on. So using 24 bits lets you aim for peaks around -15 or even -20, then you can normalize later to full scale with no degradation.

All that said, I absolutely stand by my contention that for a delivery medium 16 bits is plenty. Nobody can ever tell the difference in a blind test. Same for high sample rates. This article from Mix magazine describes an extensive test where hi-res material was reduced to 44/16 and nobody could tell the difference:

http://mixonline.com/recording/mixin...ing/index.html

Again, this is for a delivery medium.

--Ethan

[Ethan Winer]

Originally Posted by Massive Master

Not that I'd want to argue with Ethan about anything either

Aw, go ahead John, really, I can take it.

This is about stacking (perhaps dozens of) tracks together - Put 20 tracks at 16-bit (adding all that dither noise together) and you'll notice it.

Just to clarify, in DAW software all the tracks are manipulated as 32-bit floats, so dither is applied only once to the final summation as the output file is written to disk or sent to the sound card during playback. So it's not like the old days with tape recorders, where the noise floor on each track was -70 if you were lucky, and adding 24 tracks brought it up to about -55 dB.

It's still going to be less than tape noise and the DAW is probably going to be throwing calculations in 32-bit FP no matter what the word length of the source files.

Exactly. And just to keep this in perspective, if you record an acoustic guitar in a home studio you're really lucky to achieve a total dynamic range of 70 dB. Often it's much less. Only a lamer would screw up capturing a 70 dB dynamic range in a studio (not live) setting using a medium having 96 dB available. I say 24 bits is for cowards!

--Ethan

[Massive Master]

Originally Posted by Ethan Winer

Aw, go ahead John, really, I can take it.

The part that I really didn't want to argue with was summed up in the delivery medium - I'm with you there - If you can't deliver and absolutely stunning recording in 16-bits, jumping to 24-bits wouldn't do diddly (just as if stunning recordings couldn't be made in 44.1kHz, then jumping up to 192kHz isn't going to make or break anything).

I just (also with you - but the original post didn't make it clear on your stance) prefer to track in 24-bit. I'm a whore for headroom.

[Todzilla]

I have found that sample bit depth (16 vs. 24) makes a much bigger difference in tracking and mixing clarity than does sample resolution (44.1 vs. 48 vs. 96 vs. 192).

I track in 48/24 all the time.