i’m a new Audacity user.
I intend to rip the audio stream(wave) from a full BD disc (mostly concert disc), and use Audacity to ‘cut’ them into their respective track,
(by selecting each ‘track’ and exporting them separately, with fade-in fade-out)
To test the repeatability of Audacity,
I load a 3minute wave file using Audacity v203,
select partial (1m0s to 2m0s) of the wave file and export selection to a new wave file.
Without doing any changes, i exported the same selection to a 2nd wave file.
But when i do a hex compare of these 2 files, the binary does not match.
Shouldn’t both file match bit for bit ?
Most people are not doing scientific calculations on audio files. Most people are applying effects and filters and those do not work well at 16-bit. They work at 32-floating and you need dither to come back down to 16-bit for export.
If you want to import uncompressed audio (such as WAV or AIFF format), apply basic Cut/Delete/Paste type edits only (no processing) and export in the same bit format as the original (probably 16 bit) and at the same sample rate (probably 44100 or 48000 Hz), then to get bit-accurate exports you need to set the “High Quality: dither” setting to “None” in “Edit > Preferences > Quality”.
In all other cases, “bit accuracy” is not possible because processing or changing the format will change the sample values. In these cases “High Quality: dither” should be enabled, (which is usually the case for most Audacity jobs, hence why it is enabled by default).
Yes, I set the Dither to None and the files match.
Another observation i have,
When i open a wave file, it will default to 32bit Floating (unless change in Pref),
If the original file was in 24bit, and i re-export the same file also in 24bit,
both files will be identical right ?
(from my testing, its telling me they are the same with dither disabled)
I just want to make sure i dont have to manually set the project bit rate to match the file bit rate for every file that i open (usu either 16b or 24b).
I must admit my usage is slightly different from most user.
And i like to ‘test’ things before using it.
Its no joke re-editing thousands of files later on.
Well, that’s the versatility of Audacity
Nice little program.
With dither off, then yes (provided that you’ve not done anything to change them of course). All 16 and 24 bit values can be represented exactly in 32 bit float format.
(A bit of trivia: positive peaks in 24 bit format can’t quite go up t0 0 dB. The maximum positive amplitude is about -0.00000104 dB. This is because in “signed integer” format there is one more negative number than there is positive number - the range for 24 bit is -8,388,608 to +8,388,607)
