[LAU] Track bouncing

[nescivi]

On Monday 30 March 2009 06:47:32 Arda Eden wrote:
> Sorry if this is discussed before but,
>
> Some DAW software can bounce mixdown to an output file directly (like
> cubase or reason).
> But many audio people claim that this kind of bouncing is not good at all.
> They say that bouncing real-time
> (like with protools or by routing all the tracks to a new stereo track's
> input) is better resulting for audio quality.
>
> Now,
> My consideration is that, there should be no difference between the two
> because theoretically the software
> should be writing the same data in both ways.
>
> Am I missing something ?

Depending on the software, it could even be higher quality when bouncing to 
disk directly. However this is only the case if you are using control signals 
that have to be sample accurate and your software would otherwise render them 
based on audio blocks. This is what some people use CSound and SuperCollider 
in NRT mode for. In that case, it may take longer than a realtime bounce 
though.

In regular DAW software, I could also imagine that on certain parts where 
there is a lot of DSP going on (lot of effects for example), directly 
bouncing to disk may also be better, as the calculation does not *have* to be 
realtime, so where maybe in realtime you'd get hickups, rendering to disk may 
be safer.


It could be that some ProTools stuff needs to drag it through their hardware 
DSP's that will only run in realtime mode. In that case I could imagine there 
is a quality difference between letting the host PC do it, or the ProTools 
hardware DSP's. And this is probably where the myth comes from.


sincerely,
Marije