quality loss when importing image maps

[helwin]

Hi!
I think it´s more convenient to start a new thread for this problem.

So I was toying araound with importing some plain image maps. Following your suggestions, I made user settings as

maps.features.offlinemaps.formats.gen_import.image_format=PNG
maps.features.offlinemaps.formats.gen_import.image_quality=95

and imported a 261 MB .png-file.

Alpinequest converted this into a 359 MB file (!) ("Type PNG@95") which took about 40 minutes, but the quality is significantly lower than the original:

tif_pngaqx.jpg

(154.33 KiB) Not downloaded yet

Upper is original desktop file, lower is Alpinequest.

Any ideas?

[Psyberia-Support]

Hi again Helwin,

First, PNG is lossless, so the "image_quality" value is not used for this format (see it as being 100%). I do not personally recommend this format, because if you import an image that is already compressed (like a JPG image), the PNG will use more space for the same quality than the original (should be the exact same quality). JPG with a 95% compression gives a very good quality, and you'll save some space (according that you import an JPG compressed at 95% or more quality).

Then, concerning the difference of quality, I think it's only due to the fact that the above image is smoothed (called "filtered" for images, a kind of intelligent blurring) to give a less pixelated result.
This was also done in AQ in the past but it has been removed after some user requests (complaining about the blurring effect).
In beta version 2.2.8c.r6921, I've added this back as a user settings (disabled by default), so if you have this version you can confirm (or not) if it's the only reason. Use this:

Code: Select all

map.canvas.tiles.filter_reprojected_tiles=1

The filtering will only be done when you zoom beyond the last zoom level.

[helwin]

Ok much better results with this:

tif_jpgaqx.jpg

(151.01 KiB) Not downloaded yet

Mind that the original image is a high quality .tif that has not been processed. I will try some more options and report.

Thank you again for your great support!