Audacity keeps converting my 24 bit samples to 32 bit float.

[Madalert]

audacity bit conversion problem.PNG (199.3 KiB) Viewed 523 times

When I drag and drop (import) a sound file of 24 bit into audacity, it converts it into 32 bit even though I have it preferred to be set as 24 bit. what is happening here and what can I do to fix this problem?

[DVDdoug]

I'm not sure where those default settings apply and where they don't...

But, there are very good reasons why audio editors generally use floating-point for internal-temporary storage.

Conversion from 24-bit integer to 32-bit floating-point (and back) is a lossless process so there's really no downside other than larger temporary files.

[steve]

I think it's a bug, though fortunately a harmless one.

Audacity works internally in 32-bit float format, so generally it is better to work in 32-bit float format so as to reduce the number of format conversions.

I have logged this on the bug tracking system, but I would recommend that you change the default sample format to 32-bit float.

[Gale Andrews]

I'm not sure where those default settings apply and where they don't...

Assuming we are talking about WAV files, this is what currently happens

Code: Select all

Quality setting    File Format       Imports as         

16                      16                   16
16                      24                   32 
16                      32                   32

24                      16                   24                   
24                      24                   32 
24                      32                   32

32                      16                   32
32                      24                   32
32                      32                   32

Conversion from 24-bit integer to 32-bit floating-point (and back) is a lossless process so there's really no downside other than larger temporary files.

I thought that there were some large 24-bit numbers that did not have exact 32-bit float equivalents: http://forum.audacityteam.org/viewtopic ... 47#p117447.


Gale

[steve]

I thought that there were some large 24-bit numbers that did not have exact 32-bit float equivalents: http://forum.audacityteam.org/viewtopic ... 47#p117447.

Strictly speaking, that is correct, but for signals below 0 dB the error is insignificant (we're talking about precision that is hundreds of times better than "CD quality" even in the worst case).