09-28-2021, 11:22 AM
IMHO, assuming all the image have the same size, a much better way, to summarize:
1) make you images "heads with transparent background" (more on that later)
2) Create the gradient image
3) Overlay each of your image on the gradient, or add the gradient behind each.
Step #1
Step #3
The only downside of this method is that if your files have random names the naming scheme is lost, but if they have a sequence number, ofn-tiles can re-generate it when saving. But in the whole it could be best for you to handle these images as "sprite sheets", and only export final versions as individual images.
1) make you images "heads with transparent background" (more on that later)
2) Create the gradient image
3) Overlay each of your image on the gradient, or add the gradient behind each.
Step #1
- Use my ofn-tiles script to load all your images as tiles on a single layer
- Make their background transparent using Colors ➤ Color to alpha:
- fuzzy-select the background
- Select ➤ Grow by a couple of pixels
- Colors ➤ Color to alpha and remove the background color
- Create an image that has the size of one head image (you can copy paste one of the heads from the image in step #1)
- Add a layer below it
- Fill that layer the gradient
- Edit ➤ Copy to copy the gradient layer to the clipboard
Step #3
- Back to the heads image
- Add a layer below the heads layer
- Start the bucket fill tool
- Set to "Pattern" fill
- Set the pattern to "Clipboard" (the first in the list) (this should give you the gradient layer from step #2)
- Bucket-fill the added layer. The pattern repetition should align with the heads
- Image ➤ Flatten image or Layer > Merge down on the head layer
- Use ofn-tiles again to split your image into its components.
The only downside of this method is that if your files have random names the naming scheme is lost, but if they have a sequence number, ofn-tiles can re-generate it when saving. But in the whole it could be best for you to handle these images as "sprite sheets", and only export final versions as individual images.