16-32-16 conversion degrades Quality?

D

Deen

New member
Hi,

I'm using 1.2. Convert all files to 32 bit when loaded is checked.
All waves recorded in 16 bit.

I've been thinking. If Cep converts the waves back and forth
from 16 to 32 and back whenever I work on them in the edit view
then save after, wouldn't this constant conversion degrade the
quality of the recorded wave? Added noise perhaps?

I actually experimented listening to a 32 bit and 16 bit
file of the same wave. Processed with compression,
and I can't tell which sounds better. What's supposed to
be the difference. The manual mentions working in 32 bit
domain and convert to 16 in the final mix, but cep saves
waves in 16 bit. Everytime. So again does the quality
suffer this way?

Any form of enlightenment on this.
Thanks.
 
no i dont think that it degrades quality the whole point of converting a sample to 32 bit is for editing.. cool edit states that samples edit better while at 32 bit.. it says that if you do heavy sampling to a file while its a 16 bit you might lose alot of the effects that you add to the file while its trasfered back to 16 bit after the track(s) are mixed down

32 bit is said to offer more frequencies thats why it performs better with effect transformations
 
In the first place, you shouldn't be dithering down to 16-bit until the very last conversion - keep everything at 24-bit (or 32-bit floating point) until then.

In the second place, I think 1.2 has a bug, which *says* the file's in 16-bit when actually it's in 24-bit. I record in 24-bit in 1.2. The next time I load it, it says it's a 16-bit file, but if I load the same file into 2.1, it tells the truth - it's a 24-bit file.
 
I've been using 24bit native wave files for a while and just recently began using 32bit float native wave files. When I remaster something at home here I'll take a 16 bit wave file and convert it to 24bits (if using Sonar) or 32bitfloat (if using Cool Edit/Audition). It makes fade in/out, noise reduction, reverbs, etc. smoother and more detailed.

The internal Sonar busses operate at 32bit float anyway (don't know what Cool Edit uses) so as soon as you insert an effect that causes the other bits to come into play you might as well take advantage and save it IMO - later depending on what medium you're mastering for (CD, mp3, DVD, etc) you can convert and dither as you wish. There are various free VST bitscope plugins you can get to see if you're actually using the 32 bits. Ozone has a 24bit bitscope but that cost some $$$.

I recently have been upsampling to 88.2KHz or 96KHz for better freq response when remastering - same thing when inserting effects I tend to get a better sound. That takes a lot of cpu though - good thing Cool Edit allows track locking !
 
Thanks for the reply Chills and Dobro.

Hey Chills, thanks for clearing it up. SO the 32 bit conversion
doesn't degrade the audio, but rather preserves it when
effects are added before a final 16 bit mix down.

Dobro, I checked the convert all files to 32 bit when
loaded, but when I close CEp it's back to 16. So what I
did was save the Waves individually in 32 bit so I can
process it with effects. This keeps the waves in 32
even after closing cep. Thanks.

I'm curious Dobro, is it possible to record 24 bit using
a 16 bit card? I'm only using 16 bit onboard sound.
But I checked the Devices in Settings and it supports
24bit packed 32 bit unpacked. There's also
an unchecked option in Settings, System which reads Send
24 bit using 3byte packed data. What is this for.
I don't understand the Help file.
 
"So what I did was save the Waves individually in 32 bit so I can
process it with effects. This keeps the waves in 32 even after closing cep."

Okay, two things. First, go into Options>Settings and see what the default settings are like there for recording and pre-mixing and mixing down etc. Enable all the 24-bit options. (Do this when you get your hands on a 24-bit soundcard!)

Second,

"I'm curious Dobro, is it possible to record 24 bit using
a 16 bit card? I'm only using 16 bit onboard sound."

As far as I know - nope. If you've got a 16-bit card, all your tracks will be limited to a depth of 16 bits. Sure, you can enable the 24-bit option in Cool, but it will have no real effect.

"There's also an unchecked option in Settings, System which reads Send
24 bit using 3byte packed data. What is this for."

It's a setting for dealing with different types of soundcards. (All of which are 24-bit cards.)

There are two big advantages of using a 24-bit card. First, you don't have to track so hot, so close to 0 dB, so you tend not to clip so often, so you don't have to track with a compressor so often and with such vigorous compressor settings etc. With a 24-bit card, you can track with the peaks hitting at like -6 or even -9 dB. Nice. Secondly, come mixing time, you can enjoy the benefits of having all the math done in 32-bit floating point, which results in a cleaner sound overall when you're processing your tracks with EQ and compression and various effects. Nice too.
 
You must log in or register to reply here.
Back
Top