CopyWindow vs. DrawTexture timing

I am working on a PTB 3.0.8 script for a simple retinotopic mapping experiment in Matlab 7.0.0. I am presenting two, (20 degree) alternating checkerboard wedges flickering back and fourth at 15hz for 10 seconds. Reliable timing is of great importance in this experiment so I am trying to find the most resource efficient way of calling the pictures onto the display screen.
Currently, I am using "readImg" to read the .jpg file, "MakeTexture" to convert the file into a matrix form, and then using "DrawTexture" to display each wedge.
It was suggested to me that creating a separate offscreen window for each image and then calling each one onto the main screen using "CopyWindow" would be a more efficient and reliable way of doing the same thing. It makes sense to me that pre-drawing all windows would make calling them forward quicker but the experiment uses almost 50 images which means 50 windows open simultaneously. I have been playing around with it but I am still not sure which method is more reliable and I was wondering if anyone could offer any advice.
Thanks.
See our FAQ about textures and offscreen windows.

But the simple answer is, you're doing it right, with imread + maketexture to create the wedge images at startup time, and drawtexture inside the stimulation loop.

Also, 15 Hz is no challenge for any modern graphics card, so if you would have any timing problems, they would be caused by something else than the actual drawing.

Just familiarize yourself with the proper use of Screen('Flip') its timestamps etc., by looking at our demos and the intro pdf file to get best timing.

-mario

--- In psychtoolbox@yahoogroups.com, "First" <brianallencollege@...> wrote:
>
> I am working on a PTB 3.0.8 script for a simple retinotopic mapping experiment in Matlab 7.0.0. I am presenting two, (20 degree) alternating checkerboard wedges flickering back and fourth at 15hz for 10 seconds. Reliable timing is of great importance in this experiment so I am trying to find the most resource efficient way of calling the pictures onto the display screen.
> Currently, I am using "readImg" to read the .jpg file, "MakeTexture" to convert the file into a matrix form, and then using "DrawTexture" to display each wedge.
> It was suggested to me that creating a separate offscreen window for each image and then calling each one onto the main screen using "CopyWindow" would be a more efficient and reliable way of doing the same thing. It makes sense to me that pre-drawing all windows would make calling them forward quicker but the experiment uses almost 50 images which means 50 windows open simultaneously. I have been playing around with it but I am still not sure which method is more reliable and I was wondering if anyone could offer any advice.
> Thanks.
>