When i do the following thing manually: go to image, scale image, change the x resolution 600 dpi and the y resolution to 600 dpi. The width and height change accordingly to the proper size in mm. If I then export the image to png with compression to zero: the png is saved in the correct manner. If I reopen the file. the resolution and the size in MM are both still correct.

If I do the same through code. The image file has a 72 dpi and a width and height of 996 mm by 758 mm. I understand that technically this is the same but how can I make the code act in the same way as the manual actions.

desired_dpi = 600.0

width, height = pdb.gimp_image_width(image), pdb.gimp_image_height(image)

new_width = width / (pdb.gimp_image_get_resolution(image)[0] / desired_dpi)
new_height = height / (pdb.gimp_image_get_resolution(image)[1] / desired_dpi)

pdb.gimp_image_scale(image, new_width, new_height) 
pdb.gimp_image_set_resolution(image, desired_dpi, desired_dpi)


merged_layer = pdb.gimp_image_merge_visible_layers(image, CLIP_TO_IMAGE)

pdb.file_png_save(
        image,
        merged_layer,
        output_path,
        output_path,
        0,  # interlace
        0,  # compression (0-9, now set to 0)
        0,  # bKGD
        0,  # gAMA
        0,  # oFFs
        0,  # pHYs
        1   # TIFF warning
    )

 pdb.gimp_image_delete(image)

When i do the following thing manually: go to image, scale image, change the x resolution 600 dpi and the y resolution to 600 dpi. The width and height change accordingly to the proper size in mm. If I then export the image to png with compression to zero: the png is saved in the correct manner. If I reopen the file. the resolution and the size in MM are both still correct.

If I do the same through code. The image file has a 72 dpi and a width and height of 996 mm by 758 mm. I understand that technically this is the same but how can I make the code act in the same way as the manual actions.

desired_dpi = 600.0

width, height = pdb.gimp_image_width(image), pdb.gimp_image_height(image)

new_width = width / (pdb.gimp_image_get_resolution(image)[0] / desired_dpi)
new_height = height / (pdb.gimp_image_get_resolution(image)[1] / desired_dpi)

pdb.gimp_image_scale(image, new_width, new_height) 
pdb.gimp_image_set_resolution(image, desired_dpi, desired_dpi)


merged_layer = pdb.gimp_image_merge_visible_layers(image, CLIP_TO_IMAGE)

pdb.file_png_save(
        image,
        merged_layer,
        output_path,
        output_path,
        0,  # interlace
        0,  # compression (0-9, now set to 0)
        0,  # bKGD
        0,  # gAMA
        0,  # oFFs
        0,  # pHYs
        1   # TIFF warning
    )

 pdb.gimp_image_delete(image)

• resize
• dpi
• gimp
• gimpfu

1 Answer 1

If you want a given size in millimeters with a given resolution, the current resolution is completely irrelevant. Just do:

new_width=(required_width_in_millimeters/25.4)*desired_dpi

Your current computation assumes that the image already has the right size in pixels for the required print size and current definition.

Also, compression is irrelevant here, it just sets how hard the PNG algorithm tries to compress, the final image will always be pixel-perfect. With higher compression settings, you just waste CPU, but using 1-2 usually gives vastly smaller files than 0.