I am opening test-tone .wav files from an external generator program and editing them in Wavelab Elements. (don’t think Elements includes a test-tone generator. Please let me know if I’m wrong.)
The files are generated at 88.2 kHz, 16 bits. After I reduce the level by 3 dB in WLE, are the audio data now held as 32 bit FP? The “File Properties” indicator at the bottom right of the Edit window still reads 88.2/16. One can change the bit depth there. Do I need to change it before reducing level to be sure computations are done in 32-bit FP? Or only before saving the file (I want to save it at full resolution)?
Wavelab operates at 32 bit float resolution internally, so any file you load and then process in it will be at 32bf until you save the file to whichever resolution you want. Ofourse at this point, if you save below 32 bits, you are reducing bits so dither would normally be applied now (I’m not sure how relevant that would be with test tones though…). The indicator at the bottom right shows the file resolution - literally, so it shows the properties of the source file.
Notes: (1) “should”, not “would”, because dither is not applied automatically - you have to specify it; (2) because you are assumed to be saving only the required bits, as I understand it WaveLab’s dither (I don’t know about the others) does not zero the bits that are about to be thrown away - thus they may appear on the bit meter; (3) certainly you should dither test tones - undithered pure tones are a fine way to hear the distortion that truncation without dither causes!
WL never shows more bits than the chosen bit depth at the dither slot, because they’re simply not there anymore. So if after processing you go back to 16 bits, that’s what the bitmeter will show. BTW, in WL this is unrelated to what you assumingly will save to - the bitmeter shows the realtime playback stream.
You’re right - even the internal dither now zeros the low bits (it didn’t used to, but I don’t know when it changed). But the bits are still there (though now zeros), and of course you can still store the audio as a 24-bit file after applying 16-bit dither (though it would be pointless).