I'm currently 'porting' a patch from MAC to WINDOWS and have run into the jit.dx.grab interlacing problem mentioned previously here: 

https://cycling74.com/forums/index.php?t=msg&goto=62489

The solution to downsample then upsample the matrix using a hardware accelerated videoplane works great except I am trying to save some images out dynamically.

The upscaling of the image happens right at the end of the chain with a videoplane, so when I save the images out, they are at the reduced 720 288 size.

char 720 288]->[jit.matrix 4 char 720 576 @interp 1]

Is there a way to do hardware accelerated transformations in the middle of a chain of events? 

EG [jit.matrix] -> [hardware_scale] -> [jit.matrix]

The other option is to somehow capture the images after they have been upsampled ie glreadpixels

I'm currently 'porting' a patch from MAC to WINDOWS and have run into the jit.dx.grab interlacing problem mentioned previously here: https://cycling74.com/forums/index.php?t=msg&goto=62489

While I can use the software method:
[jit.dx.grab 720 576]->[jit.matrix 4 
char 720 288]->[jit.matrix 4 char 720 576 @interp 1]
This is a lot slower.

Is there a way to do hardware accelerated transformations in the middle of a chain of events? 
EG [jit.matrix] -> [hardware_scale] -> [jit.matrix]

jit.dx.grab - Hardware deinterlacing