32 float conversion options?

I’ve managed to get some flawless recordings into WLE8. I’ve been using 32 float .WAV . Now I need to convert some of these recordings utilizing as little processing as possible.

My first question is… I need to get these 32 float .WAV recordings to 24 bit .WAV so players recognize the format. Can I simply “save as” 24 bit, IE do nothing else to move from 32 bit float to 24 bit?

you should be using a dedicated process. i’ve always used voxengo’s r8brain, for mastering anyway.

Ideally, you render your files through the Master Section and add Dithering. You choose PCM 24 bit as output.

I changed the thread title in order to not confuse.

So it’s absolutely necessary to dither from 32 float to 24 bit?

No, it’s not. It’s an option. It would be more important if you had to convert to 16 bit.

As your audio interface is very likely to be limited to 24 bit, if all you did was record and not add any processing in WL, then you effectively recorded 24 bit files anyway - only expressed as 32 bit float. So, no need to dither, and as PG said, dithering to 24 bit could be debatable anyway (in the sense that it is unlikely that a difference can be heard).

In fact, I would argue that recording to 32 bit float is unnecessary. Maybe when adding processing later, saving to 32f inbetween could be useful if further processing takes place later on.

The good news is, I knew that. 16bit is a lot more standardized than 24 bit. That does lead to a caveat… Is noise shaping necessary when going from 32 float to 16 bit? From what I can tell, it’s at least recommended. But I’m not sure if it is necessary?

That’s what I originally meant in my OP before I changed the title. “Option” is subjective. So, What should I look for when deciding what conversion direction to go in? Why dither or why not dither in the strict case of 32 flot --> 24 bit?

In this case, I process. I record my 12" records into WLE8 at line level. I apply gain/normalization in WL.

The idea here is to keep these record’s sound as pure as possible with as little processing as possible while making, lets call them base files to work off of, that can be converted into whatever file type and resolution someone else wants. I hope that makes sense.

bump. Any comments?

The idea here is to keep these record’s sound as pure as possible with as little processing as possible while making, lets call them base files to work off of, that can be converted into whatever file type and resolution someone else wants. I hope that makes sense.

Keep/process you file in 32 bit float and deliver to whatever format you need later.