Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Automated batch xcf layer combining and png exporting
#5
Wow, can't thank you guys enough.  Here are some answers, results, and new questions. 

Rich:
> Are all the images the same size ?
The top layer is smaller than the middle which is smaller than the bottom.
> or do you need to place top / middle layers in specific places ?
They should just need to be stacked in order.  I pre-aligned each layer to a common sized boundary.  I wanted to simplify the automation effort to compositing only, and avoid any scripted image manipulation, at least for now.  Some alignment woes described below.
> Images are to be stacked before writing to png, so these layers are partially transparent ?
Yes.

Ofnuts:
Awesome shell script and sample output!  Can't tell you how exciting it is to have my output directory fill up with these combinations.

Both:
Yes ImageMagick 7.1.0-6 chokes on v2.10 XCFs. So I tried saving my XCF layers as pre-2.10. The only obvious method I found is to uncheck the last option when Saving, which leads to the Gimp complaint "Layer mode 'Normal' was added in GIMP 2.10".

When I saved the layers like this, ImageMagick could combine them, but sometimes layer alignment was lost in the resulting PNGs. I spent a bunch of time going in and re-aligning each image layer in Gimp and saving as pre-2.10 XCFs. But somehow the results were inconsistent.  So I ended up just going with PNG layers, which all aligned correctly when combined with the script.

Questions:
Is there a better way to enforce layer alignment than pre-aligning each layer to a common sized boundary?  (I never paid much attention to this, it always just worked right when doing things manually in Gimp.)
What's the preferred way to save v2.8 XCFs?
Gimp doc says PNG is lossless here, but when exporting to PNG, Gimp offers a compression slider, defaulting to 9. Is 9 lossless? Or maybe 0?
Reply


Messages In This Thread
RE: Automated batch xcf layer combining and png exporting - by gimpler - 09-06-2021, 06:05 AM

Forum Jump: