| mark4man |
| (Newbie) |
| 09/21/07 07:06 PM |
32-Bit Floating Point Mix Engine (II)...So What?
|
From the WaveLab Help Files:
Quote:
The audio path in WaveLab uses 32 Bit floating point processing. You can therefore "overload" it, i.e. use levels above 0 dB, without running the risk of causing digital signal clipping in the signal path. The only section of the audio path that can introduce clipping is the output of the Master Section.
OK...
Since we don't seem to be getting anywhere on this topic...maybe I should ask my question in this way:
What the hell does it matter if you can "overload it...i.e. use levels above 0 dBFS"...if that "overload" never makes it to the output ??? When WaveLab (or any audio application) converts that overloaded 32-Bit floating point signal to integer to send to the audio card...& everything above 0 dBFS is clipped...we're not hearing an "overload free" signal anyway, in our studio monitors...theoretically...are we?
So then what's the point ??? (no pun intended.)
But I think we somehow are; & I think there has to be some other factor at play here....something in the way WaveLab perhaps processes that overloaded signal prior to sending it to the output...because...If I take a multitrack montage that has a peak signal of +6 dB & listen to it...listen to the playback...it sounds much different that if I bounce it to a 24-Bit .wav. It sounds cleaner (& for some reason distortion free.) In the 24-Bit bounce, it has some scratch.
BUT THERE SHOULDN'T BE ANY DIFFERENCE BETWEEN PLAYBACK & RENDER...since playback essentially encompasses the same thing...the 32-Bit floating point mix is truncated to 24-Bit to be sent to the audio card. Exact same thing happens in a bounce.
So...there has to be some factor as to why playback sounds better....why what we're hearing sounds better.
Am I the only soul left on earth asking this question? This is important sh_t, is it not?
mark4man
BTW - While we're on the subject...are there any white papers from Steinberg hangin' around the web that detail exactly what WaveLab does send to the audio card; & how that signal is routed/processed?