dithering/bit truncation/mp3 conversion

generally when mastering from 24bit source files the plugin chain generates much higher wordlength (sometimes 40 or 50+ “apparent” bitdepth)
as i understand it WL8 internally processes these “apparent” bits in 32bit floating point.
so i save the processed files accordingly to 32bit float file format.
if i need to go to 24 or 16 bit i will have to dither it down to avoid bit truncation of the information contained in the higher bit rate.
the problem came up with MFiT (mastered for iTunes) which at this stage only accepts 24bit files (in the future higher…) for submitting it for their MFiT conversion process.
this is, as they claim, to get higher quality audio directly converted to 256AAC WITHOUT using dithering.
Q:
how can i generate 24bit master files from my 32+bit processing chain without using dithering and avoiding bit truncation? is that possible in WL8?

Q:
is it possible to convert within WL8 from 32+bit processing directly to mp3 without dithering and bit truncation?

Q:
can you explain why WL8 is referring to “apparent” bit depth.

thanks

how can i generate 24bit master files from my 32+bit processing chain without using dithering and avoiding bit truncation? is that possible in WL8?

If you don’t explicitly use dithering, no dithering is applied.
If you pass from 32 bit to 24 bit, obviously bit truncation is applied.
If MFiT accepts 32 bit, then just save to 32 bif float.

is it possible to convert within WL8 from 32+bit processing directly to mp3 without dithering and bit truncation?

Yes and no, as mp3’s encoding itself truncates to 16 bit.

can you explain why WL8 is referring to “apparent” bit depth.

This means WaveLab analyses the samples without taking care of what the file header say.

Yes and no, as mp3’s encoding itself truncates to 16 bit.

so best practice is to convert to mp3 from 16bit dithered files?