Hey all,
Does anyone have any experiences with really long messages in PD. I mean in cases when using things like "pix_dump" that drop many floats.
Is there a maximum, what has the most anyone has used?
I'm currently trying a patch with pix_dump'ing a 400x300 RGBA image each 12s.
40x30 worked very well (where pix_dump outputs 4800 floats) and rock solid, but pretty low-res.
Thanks! B. Bogart
On Mon, 26 Nov 2007, B. Bogart wrote:
Does anyone have any experiences with really long messages in PD. I mean in cases when using things like "pix_dump" that drop many floats. Is there a maximum, what has the most anyone has used?
I used lists of 230400 elements for sure, but I might have also tried 921600 elements. This was with [#export_list]. With [pix_dump], I remember that it has problems with that. When I made [#export_pix] and [#import_pix], I was comparing with an implementation that involves [pix_data], and I had to fudge it because some of my lists were too long for [pix_data].
Try to find the size at which it starts failing. It might be 262144 (exactly 2 megs in atom format) or some similar number.
_ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard - tél:+1.514.383.3801, Montréal QC Canada
On Mon, 26 Nov 2007, Mathieu Bouchard wrote:
I used lists of 230400 elements for sure, but I might have also tried 921600 elements. This was with [#export_list]. With [pix_dump], I remember that it has problems with that. When I made [#export_pix] and [#import_pix], I was comparing with an implementation that involves [pix_data], and I had to fudge it because some of my lists were too long for [pix_data].
er, the problem is in either [pix_data] or [pix_dump] or both, I don't recall. I don't see the problem in the source code of either. The next step would be to check whether Gem has a global limit of number elements per list.
However, something really weird that I just found, is that [pix_data]'s grey outlet does not do any weighting. It should compute grey as 30% of the red plus 59% of the green plus 11% of the blue, but it just takes each as one third. You can see whose percentages are correct by drawing coloured squares surrounded by grey squares that are supposed to be greyscale versions of the colours. The right weighting will make the transition look much less discontinuous. (gamma issues may get messy here)
_ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard - tél:+1.514.383.3801, Montréal QC Canada
-
B. Bogart -
Mathieu Bouchard