I'm looking to cut up image data into regularly sized screen blocks. Currently the method I've been using is this:
def getScreenBlocksFastNew(bmpstr):
pixelData = array.array('c')
step = imgWidth * 4
pixelCoord = (blockY * blockSizeY * imgWidth +
blockSizeX * blockX)* 4
for y in range(blockSizeY):
pixelData.extend( bmpstr[pixelCoord : pixelCoord + blockSizeX * 4] )
pixelCoord += step
return pixelData
bmpstr is a string of the raw pixel data, stored as one byte per RGBA value. (I also have the option of using a tuple of ints. They seem to take about the same amount of time for each). This creates an array of a block of pixels, depicted by setting blockX, blockY and blockSizeX, blockSizeY. Currently blockSizeX = blockSizeY = 22, which is the optimal size screen block for what I am doing.
My problem is that this process takes .0045 seconds per 5 executions, and extrapolating that out to the 2000+ screen blocks to fill the picture resolution requires about 1.7 seconds per picture, which is far too slow.
I am looking to make this process faster, but I'm not sure what the proper algorithm will be. I am looking to have my pixelData array pre-created so I don't have to reinstantiate it every time. However this leaves me with a question: what is the fastest way to copy the pixel RGBA values from bmpstr to an array, without using extend or append? Do I need to set each value individually? That can't be the most efficient way.
For example, how can I copy values bmpstr[0:100] into pixelData[0:100] without using extend or setting each value individually?