Topic: CEP2 multitracking

[dklein]

imho, there is no reason to be off by 5ms/min. (this only applies to pure digital sources).  if you're off by the amount you state, you should go back further and sync it there; therefor, no need to time stretch.

it's ALWAYS easier (and more practical) to cut a wav than to time stretch + or -.

marc

marc - not quite sure what you're saying.  Perhaps I wasn't clear:

Without any correction, my 2 sources drift by 5 ms/min.  That's just how it is - different A>Ds.  Wherever I synch in post, one minute later, one of the sources will be out by 5 ms.  My ears don't seem to like the sound when I'm off by more than about 5 ms.  That means I would need to cut the wave file of one source every minute, which is a huge amount of work, requires smoothing between samples AND also means that I spend more time out of synch than in it.

On the other hand, if I do a proper, pitch preserving stretch, I can synch up at the start point and end up within a couple of milliseconds an hour later.

What makes the first approach easier and more practical?  

[bluevolvo]

If you have a situation where the drift is too much for the 'cutting' method (I have a 5 ms/minute drift) than I recommend trying a stretch in wavelab (high quality, preserve pitch).  I tried a few other stretch functions with poor results, but this one sounds very good to me.  I also have wavelab temp files set to use 32 bit float.

You have to figure out the proper stretch

• synch up 2 points - one near the beginning and one near the end

figure out the amount of drift over that period of time

• Highlight the same area in wavelab and go to stretch - only use the ratio control - play with it using values to the 3rd decimal (xxx.xxx) until the time displayed is the duration you want to match up with on the other synch point (the slightly longer or shorter one)
• Once you find the best number, use that to stretch the whole file.
  
• Sometimes there's some trial and error required to zero in on the right value.  It's better to be a little longer and edit out the odd millisecond or two than it is to come up short.

I'd like to hear some people's opinon of the wavelab stretch - take a sample and stretch by a factor of 99.980 or something like that.  See if you can tell the difference.

I use AA2, but I concur...esp with the "99.980" statement.  32bit and high quality, preserve pitch.  I also lock and sync to whatever I consider my better source (rig/deck)

Nicely written.

[Charlie Miller]

Just curious for those who have tried ... if the initial 4 tracks are all done on a laptop and synced to the same clock, will they be aligned the entire time (assuming the two sources are aligned to begin with)?

If it's on the same clock then you're fine. BTW, I use one of these:
http://www.cascademedia.net/products.asp?catid=129&prodid=524

It helps me with my 32 track set up...

cm