[Explained] Why should you use 24bit rather than 16bit depth while recording

[xeonblade]

Hey people, it's me! A guy with super MSPaint skillz! This is a continued post from [Explained] Why should you use higher Sample Rates while recording. But let's cut the crap and go directly on the subject.



SNR = signal to noise ratio
Noise floor = noise produced by electronic device
0dBu = signal ceiling (in decibel volts)

In this case Signal to Noise ratio is the same as Dynamic Range.
Note that:
SNR = ratio between an arbitrary signal level and noise
Dynamic range = difference between the largest and smallest signal a system can record or reproduce

For each additional 1bit you get 6dB of SNR.

Unlike 16bit, 24bit depth is able to cover whole human hearing range. (And here I don't mean 20Hz-20kHz, I mean 0 to 140+dB)
But that isn't the point. Again like the sample ratio, it's about sound quality.
16bit depth offers 65536 levels of audio signal while 24bit offers 16 million levels. This gives incredible amount of sound detail.

Look at it like colors: With 8bit colors you can see 256 shades of a color, with 16bit you have 65536 shades and with 24bit picture you can see 16 million shades. (Just change ur monitor to 16 bit and see the difference)
Something similar is with the sound! Just remember how toony 8bit remixes sound.

Unlike sample rate bit depth doesn't increase latency while recording and it's strongly advised that you use 24bit if possible, it gives your audio more room to breathe in the realm of digital world.


If anything is unclear, please tell me because English isn't my native language and it was kinda hard for me to explain this simple enough.
Please notify me if I made any mistakes that should be fixed.

P.S. BE AMAZED BY MY MSPAINT SKILLS!

[Purelojik]

makes sense also when playing with levels and such. it all give you more "headroom". or basically room to expand your sound. which is nice cause it wont seem like things are getting suffocated.

EDIT: sorry jumped the gun, saw that you had already said this at the end of your post. my bad

[sh4z]

Nice Post.

[OrsusMetal]

Nice to see it explained. Also, AWESOME SKILLZ.

[xeonblade]

Tnx both of you, there is [Explained] Why should you use higher Sample Rates while recording if you are interested.

[ROAR]

I've been learning about this in school and this is a
really great and simple explanation.

Great job man

[VBCheeseGrater]

The color analogy is a good one.

[Larcher]

Once again, dem ms paint skills <3

[Menigguh]

32 bit say what!

[Exit Existence]

*Record's in 24 bit, then render's in 196kbs mp3 to put online*

[ROAR]

^Yea you show em!

[S-O]

While I am down for 24 bit recording, 32 may just be a bit outrageous! 32 bit would have 192db SNR, which, let's consider the noise floor to be 0db, since anything below it would not be very discernible. Now, the threshold of pain is ~140, and ear drums tend to rupture ~160 or 180 air pressure blast, and lungs tend to rupture around 200.

So, 32 bit may just kill you!

[Rational Gaze]

I still think if you're going to be entertaining these discussions, you should include the harddrive space conversions. Just a thought.

[the unbearable]

Originally Posted by Exit Existence

*Record's in 24 bit, then render's in 196kbs mp3 to put online*

320k represent!

[ralphy1976]

So I guess you did not get laid since last thread?

[xeonblade]

Originally Posted by Rational Gaze

I still think if you're going to be entertaining these discussions, you should include the harddrive space conversions. Just a thought.

You live in time when 2TB HDD costs around 70-80$ and you cry like max HDD is 40GB?
(Not including this price jump because of floods in Taiwan!)

Originally Posted by ralphy1976

So I guess you did not get laid since last thread?

Nope, why do u think I wrote another thread?

Originally Posted by Menigguh

32 bit say what!

32bit is an overkill which gives results only in theory.

Originally Posted by Exit Existence

*Record's in 24 bit, then render's in 196kbs mp3 to put online*

Record in 192kHz 24bit, release in 24bit 96kHz 320kbit mp3! Or 192kHz 24bit FLAC

[KingAenarion]

Originally Posted by xeonblade

You live in time when 2TB HDD costs around 70-80$ and you cry like max HDD is 40GB?
(Not including this price jump because of floods in Taiwan!)

Except project files increase very quickly in size.

My band album is currently sitting at 60GB for 11 tracks... and I haven't even started tracking Bass or vocals.

Now imagine you're a studio and you're recording 6 days a week. If you're backing up your projects and keeping copies of projects that need to come back at a later date... You're going to fill up a terabyte quite quickly.

Add that to sample libraries and space fills up VERY quickly.

[xeonblade]

Originally Posted by KingAenarion

Except project files increase very quickly in size.

My band album is currently sitting at 60GB for 11 tracks... and I haven't even started tracking Bass or vocals.

Now imagine you're a studio and you're recording 6 days a week. If you're backing up your projects and keeping copies of projects that need to come back at a later date... You're going to fill up a terabyte quite quickly.

Add that to sample libraries and space fills up VERY quickly.

We're mostly talking about home recordings here for personal uses and I'm replying having that in mind. 96kHz is half size than 192kHz and quality is unnoticable. 24 bit vs 16 is +50% in size rather than +100%.

[Seamless]

Hard drive are also more expensive than they used to be now since the flooding in Thailand.

Now, I'm not a super pro so I might be completely wrong about this, but I still don't quite see the value in 24 bit, at least in heavy music production.

"heavy" in terms of genre, metal, techno stuff. That sort of thing.

Having a wider dynamic range, and even a larger distance between the noise floor and 0db doesn't seem to matter much when you're constantly pushing everything to be 0db.

I would totally use 24bit if I were recording something like an orchestra doing a classical performance where dynamic range is extremely important and also obvious.

But for things like drums, guitars, bass, and even vocals, the dynamic range of the kind of mix we shoot for is usually "on" or "off". Not something that's enhanced (from what I can tell) by 24 bit.

[xeonblade]

You don't have to use it. I just explained it. Please don't start shitting over my hard work for no reason.
If it's metal it doesn't have to be superquality? It ISN'T just Dynamic Range. It's about sound quality.

I've already mentioned "EXCEPT for this period where HDD are more expensive cuz of floods"

[Seamless]

I missed where you qualified the floods. My bad. I'm also not advocating NOT using 24 bit. Just my, singular, view on its usefulness.

Now, "superquality" in a recording is an interesting problem. Like, for example, could you tell the difference between a 16bit Superior Drummer sample or a 24 bit Superior Drummer sample? (I might actually be wrong about that being relevant at all)

Would you be able to tell a production done primarily in 16 bit from one done primarily in 24 bit?

And the all time favorite: Could you tell the difference between an MP3 and a less compressed file format?

The answer to those questions for just about everyone is no. Not that I'm saying you shouldn't make songs of a higher file quality if it makes it sound better for you just because no one else can tell.

Heh, I just realized, actually, that if I use the particular method that I do use to record what I record in FL Studio then it doesn't matter. This is because FL's default (and somehow unchangeable) setting is to record the inputs at 32 bit float. So, until I bounce the tracks to a final mix into 16 bit files, I'm running with an even larger distance from the noise floor.

And lastly, I suggest that you not take arguments personally. Since, at least from me, they are not meant to be.

[xeonblade]

Don't mind my bitching. I'm just sick of people jumping on my head and shitting on my tries to help anyone (which almost always happens, but doesn't have to be your case).

Edit: Most people maybe wouldn't be able to make a difference between 16 and 24bit, 44,1 vs 192kHz or low bit rate vs high bit rate, but if you take 16bit, 44.1kHz, low bitrate and compare to 24bit 192kHz high bitrate it makes a difference.
Mostly guys just ignore those small thing that combined give some effect and just start shitting how normal people don't even notice the difference.

[Rational Gaze]

People aren't. We get that you're trying to help out. But you said I was crying about something. No, I was not. That's a shitty attitude to have when someone is just trying to suggest something that is an important part of the conversation when you're working on a large project where resolutions are high, hard drive space may be running out, and CPU loads may not be optimal. Not everyone makes music on the latest hardware. But aside from that, you need to learn to relax.

[xeonblade]

You are right.

[Ryan-ZenGtr-]

I like what your doing here xeonblade and I've enjoyed your presentations.

Each of us will be limited by our hardware quality, so we all have our own opinions on how practical each format is for us.

Thanks for taking the time to put together your presentations. I'm sure it will help many people, and expose others to some of the practical limitations and possible workarounds discussed in the thread.