[Ardour-Users] Automation from analog mix?

[Arnold Krille]

On Saturday 18 December 2010 10:35:48 David Kastrup wrote:
> Arnold Krille <arnold at arnoldarts.de> writes:
> > On Friday 17 December 2010 17:21:00 David Kastrup wrote:
> >> Jörn Nettingsmeier <nettings at stackingdwarves.net> writes:
> >> > On 12/17/2010 02:43 AM, David Kastrup wrote:
> >> >> I was wondering: if you are recording using a mixer (and using
> >> >> pre-fader recording outputs), wouldn't it be feasible to record the
> >> >> analog mix and use echo compensation techniques for figuring out the
> >> >> fader settings (if we are more ambitious, equalizer settings) and
> >> >> pull them into the automation?  That would allow one to use a good
> >> >> analog mixer (with all its nice controls) to do all the mixing
> >> >> hands-on, and still get output with the quality achievable by
> >> >> digital mixing.
> >> > you mean i would feed pink noise into each of my mixer channels,
> >> Nope, just the usual signals.
> >> > buy a 32ch a/d converter to get the tape outs,
> >> It's already there if you are recording the sources.
> > No, the 16 channel converter is already there to record the main
> > signal directly after the pre-amps. You need another 16 channels for
> > the signal after the fader to get all the information you want for
> > automation of eq, gate, compressor and volume.
> Nope.  Just the stereo mix at the end.  It has all the required info
> after decorrelation ("echo compensation techniques") if you take into
> account that changes are confined to small bursts of activity.

Yep, I had the same idea over night. But while this will be easier on the 
number of needed channels, this is much harder on the number of parameters you 
have to determine on every single sample. Think of the order of 100 parameters 
with a 100^10 possible values that need to be checked against the real result. 
Better use that cpu-cycles for recording... And if you run the algorithm 
overnight on the data, better spend that time and energy on doing the mix 
right in the first place...

You do not want to confine your algorithm to echo compensation, you want the 
full parameter-set of the mixer automated from that stereo mix and the 
individual signals. That is a constant "burst" of very high activity on your 
machine and still you will not get it right.

And still you have to acknowledge that mixing for live is something different 
then mixing for cd. So much different that manufactures advertise "total 
recall" for studio- or (as in XOR) for live-usage...

Have fun,

Arnold
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.ardour.org/pipermail/ardour-users-ardour.org/attachments/20101218/a6cee3a0/attachment-0002.pgp>