Comments by Peter Gibbins (Top 2 by date)

Thanks for answering Original Griff

Yes - you're right. from what I read the JPEG compression uses the Discrete Cosine function which (if I understand the process) effectively compares pixels in successive 8x8 grids to produce a smooth "averaging" of the colours in that area to reduce the number of colours used which then allows the pixel count process to work. The "pixel count" is lossless, but the DCF process to reduce differing colours is lossy - the more so as you increase the compression.

I admit - trying to work the DCF without using pre-written libraries is way beyond my maths skills... So I want to attack the data store, avoid the complexity of the DCF function and just tinker with the colour values.

It may not work, but I'd like to try.

Hence, needing to understand the image data structure.

Peter