[GEM-dev] strange behaviors with multipass rendering, setting gemhead order with float number, and image latency

[cyrille henry]

ah, ok, I think found the problem,
the set message only accept int, but float work for initial argument.
i'll fill a bug report

c

Le 19/12/2022 à 19:07, cyrille henry a écrit :
> Hello,
> 
> I experience strange behaviours I can't explain.
> the context is multipass rendering : the render order of each framebuffer is important otherwise you can experience 1 frame latency.
> I need to set explicitly the render order of gemhead using the set message, because I use the same abstractions multiple time and they need to be render at different time.
> I also need to use floating point number since I do have lot's of different gemhead. (gemhead use to accept only int, but if I remember correctly, this limitation is gone long any, is that right?)
> 
> I manage to reproduce my problem in a simple patch, see attachment.
> 
> When moving the teapot, you experience 1 frame latency for each pass of the abstraction. (for a total of 14 frames)
> But, if you remove the loadbang in the abstraction, everything works as expected.
> if you set the same value with the loadbang in the main patch : it works. If you don't set the order position  with loadbang, or with the set message, it also works (since the correct value is explicit in the gemhead object).
> 
> 
> And of course, on my big patch, changing the timing of the set message did not help...
> 
> When you experience latency, the rendering order is reversed.
> 
> 
> I would love to have an explanation.
> 
> Thanks for your help
> C
> 
> _______________________________________________
> GEM-dev mailing list
> GEM-dev at lists.iem.at
> https://lists.puredata.info/listinfo/gem-dev