I'd like to propose a change to Export behaviour, or possibly to the
Merge Visible Layers feature, depending on what other developers think
about the following:
1. Create new image with pink background
2. Add layer, draw picture of bird
3. Save as PNG => Result is an RGBA PNG, 121Kb
4. Flatten, Save again => Result is RGB PNG, 105Kb
The trouble is that "Merge Visible Layers", which is currently
recommended for any multi-layer image saved to a transparency capable
image file format, adds an alpha layer to resulting layer, even if
the lowest visible layer was a non-transparent background.
I don't think altering Merge Visible Layers is the right thing to do,
there would doubtless be a number of surprises waiting if I did it.
Instead, I propose to change Export to recommend "Flatten" when it
notices that the bottom layer has no alpha channel AND is visible,
when it would current recommend "Merge Visible".
This will improve compression ratios for PNG, TGA etc., as well as
increasing available colors for GIF and colormapped PNG.
Comments? Objections?
Nick.
