| Author
|
24 bit vs. 16 bit ... yet another question.
|
groovemachines
Started Topics :
5
Posts :
57
Posted : Jul 1, 2007 12:50
|
Hi guys..
I know this matter is pretty consistent.
But I am new to 24bit so I got a few questions:
1 - Most of my audio samples are 16bit wavs.
When working on a 24 bit project I have to convert it to 24bit as soon as importing? I mean.. will I loose quality when converting upwards?
2 - All of my were done in cubase in 16/44. Can I just simply open the projects, set them to 24 bits. convert the audio samples. and then render in 24 bits? will I benefit from it?
resuming...
3 - How to proceed when you want to work on a 24bit project but audio samples are 16/44 ?
thanks in advance!!! |
|
|
Tomos
IsraTrance Full Member
Started Topics :
84
Posts :
981
Posted : Jul 1, 2007 12:57
|
You can't get more quality than what you originally had, ever. You can't lose quality upsampling.
If you want to work in 24bit, you can still convert all your 16bit samples. They'll be no benefit unless you add some 24bit ones too, although some effects like reverb might sound better. Remember to dither correctly when you finish the track down to Cd quality or the whole exercise will have been pointless.
|
|
|
|
groovemachines
Started Topics :
5
Posts :
57
Posted : Jul 1, 2007 13:00
|
Quote:
|
On 2007-07-01 12:57, Tomos wrote:
You can't get more quality than what you originally had, ever. You can't lose quality upsampling.
If you want to work in 24bit, you can still convert all your 16bit samples. They'll be no benefit unless you add some 24bit ones too.. remember to dither correctly when you finish the track down to Cd quality or the whole exercise will have been pointless.
|
|
ok .. got it!
please tell me more about dithering correctly when downsampling to 16/44 again. |
|
|
|
dtd
IsraTrance Junior Member
Started Topics :
17
Posts :
490
Posted : Jul 1, 2007 13:25
|
hi groovemachines,
as tomos said, when converting from 16bit to 24bit, you won't lose quality. think of the numbers which encode each sample. adding precision won't alter a number. but going from 24bit to 16bit introduces quantising noise (as 2^24 distinct values need to be represented using 2^16 distinct numbers, rounding occurs. --> use dithering for this step.)
i would like to add that working with a 24bit project (converting your 16bit samples to 24bit) directly affects quality (ok, it still is a question whether one hears it or not.) but even a gain change (your fader in your sequencer) introduces rounding errors! so if you're working at 24bit, you already benefit here. same for all other signal processing stages. --> work in 24bit if you can.
hope i was able to help you a bit,
cheers!
 $ exp(j*pi) + 1 = 0. $ |
|
|
|
groovemachines
Started Topics :
5
Posts :
57
Posted : Jul 1, 2007 13:48
|
ok.. I think I am starting to understand a little more.
let me ask one last question about my pre-mastering chain. .. on my chain the last plug is a series of 3x waves L2 .
the L2 for example has its own bit and dithering settings.
lets say.. even if u are working on a 16bit project it works even with 24bit selected on the quantize setting of the L2.
what I mean is.. on L2 there is an option for dithering... if I work on a 24bit project. then decide to render it to 16bit. should I disable the dither section on the L2 and use only the dither plugin from cubase. or it doesnt matter?
and one more. is it better to render the whole wav 24/96 from cubase and then convert the wave itself to 16/44 or its better to covert from the 24bit project straight away?
too many questioons! |
|
|
|
Tomos
IsraTrance Full Member
Started Topics :
84
Posts :
981
Posted : Jul 1, 2007 14:06
|
The more dithering plugins you use, the more audible noise is introduced in an attempt to hit sample points more accurately. One good dithering plug is enough for the whole chain.
Somebody please please correct me if I'm wrong here (as this is the way I do it) you select the value you are quantizing to. So if your project is 24bit, and you're going to 16bit - select 16bit in L2.
I just convert to Cd quality inside cubase. The more programs you use, the most chance of problems is the way I think. Cubase has the option to do it - use it.
Or try a different app and see which sounds better, but you'd need some pretty supreme ears to tell the difference at that stage I think.
Did you say you use 3x waves L2's ?? |
|
|
|
groovemachines
Started Topics :
5
Posts :
57
Posted : Jul 1, 2007 14:19
|
yeah man. it is a little trick I have learned back ago!
it really works.
first insert one L2. then work on your threshold level. after you figure out what is the ideal value that gives you HIGH volume and no distortion. then you get this number. lets say it was -10db.
write that number down.
now insert 2 more L2 right below eachother.
now is the trick.
divide your main threshold (-10db) by 6.
10/6 = 1.66
now you set the threshold for the L2s using this calculations:
1st L2 Threshold = 1.66 x 3 = 4.98
2nd L2 Threshold = 1.66 x 2 = 3.32
3rd L2 Threshold = 1.66 x 1 = 1.66
got it?
try it and let me know... it is perfect!!
|
|
|
|
PoM
IsraTrance Full Member
Started Topics :
162
Posts :
8087
Posted : Jul 1, 2007 14:36
|
|
what sequencer do you use ?why 24 bit projects and not 32 bit? |
|
|
|
groovemachines
Started Topics :
5
Posts :
57
Posted : Jul 1, 2007 14:41
|
cubase sx3.
|
|
|
|
groovemachines
Started Topics :
5
Posts :
57
Posted : Jul 1, 2007 14:46
|
as far as I know my audiocard supports only 16 and 24 bits. being 24/192 the highest quality.
we are talking about another kind of bits.
the 32 bit you are talking is 32-bit float which is actually 24-bit digital audio waveform with additional precision, used internally by the s/w.... outside of the s/w, it's a 24-bit file.
i'm afraid there is no such thing as a 32bit/96 file for example. |
|
|
|
PoM
IsraTrance Full Member
Started Topics :
162
Posts :
8087
Posted : Jul 1, 2007 14:55
|
|
in a 24 bit cubase project when you bounce a softsynth at 24bit is there automatic dithering from 32 to 24 bit? if yes i think it s better to use 32 bit project with 32bit samples only so no dither is used in the projects , it should give less cpu load (if cubase have to dither everything back to 24 bit at the mixer output) and you can use a better algorythme than the one of cubase in the end of the mastering. |
|
|
|
Tomos
IsraTrance Full Member
Started Topics :
84
Posts :
981
Posted : Jul 1, 2007 15:53
|
|
Thanks for the L3x3 tip. I'll try it out and report the results. I guess there is some logic in there as I've heard more light compression is always better than one heavy. |
|
|
|
