I still have a question. Since the bottleneck of my patch seems to
be the video acquisition, I wonder if there is a way to capture
direct to texture, the same way we can capture direct to window with
the "window " message.
>> I still have a question. Since the bottleneck of my patch seems
>> to be the video acquisition, I wonder if there is a way to capture
>> direct to texture, the same way we can capture direct to window
>> with the "window " message.
> jit.qt.grab @colormode uyvy -> jit.gl.texture @colormode uyvy is
> the fastest way to accomplish this.
I tried, but this makes no difference than going straight to
jit.gl.slab in uyvy colormode as I was doing before.
I use the codec ‘uncompressed 4:2:2′ from the iSight. Would there be
a more efficient codec? I understood from previous threads that this
codecwas good because it did not ask Jitter to uncompress, expecially
since I use the uyvy colormode. Have I understood things right?
ps I manage to save 3% cpu by remplacing the second jit.gl.slab in
the feedback example by a jit.gl.texture. According to a post last
week, the texture is not doing reframing and that saves cpu. Is that
right or am I drunk again?