Adds 16bpp support for BMP File Format #67608
Conversation
|
I also added the test project I was using while testing the code here: https://github.com/ScorpionInc/godot-projects/tree/main/BMPTest if that would help anyone reviewing out. It's not the prettiest project but it uses alot of different BMP image formats and colors so you can see the edge cases better. |
|
The check: "Static Checks / Static Checks (clang-format, black format, file format, documentation checks)" Fails still however the suggested results look the same as my current values. Is there some invisible format character in there somewhere that I'm missing? Thanks for your time. |
You need to use tabs for indentation, not spaces. Make sure you don't have mixed tab/space indentation too (we don't use spaces for alignment). From a quick look at the CI log, you appear to have mixed indentation on the lines highlighted in red. |
ScorpionInc
force-pushed
the
Add_BMP_16bpp_Support
branch
3 times, most recently
from
7bc32aa
to
48c748d
Compare
|
Alrighty should be ready now hopefully thanks as always for everyone's time and help. o/ |
|
You can check out with my MR #67802 too. |
|
If I remember correctly, Reduz didn't like color palletizing (color reduction) or using formats typically LDR for HDR because it caused confusion and thus made Godot Engine less usable. This appears to be converting RGB565, and RGB555 to RGBA8. So I think it's fine. |
|
I am not seeing where the color scaling is done. This is probably just something I missed because I'm dumb. I've tried compiling the code from your fork and testing it with many different 16 bit formats of BMP and have not gotten resuts I expected. This may be because I wrote my code and project for an older version of Godot and havent forwarded my code yet. However your code seems to produce the same results as stable branch but with more warnings and a black box vs more errors and transparent sprites. I'll try applying my code changes to the current version of Godot (4.1.dev.custom_build[21d080e] at this time) and see if I get simular results, but wanted to share what I'm seeing so far. Project: I'll keep you posted on what I find. I like the look up for masks, based on what I read in your merge it sounds like the missing piece to get some of the alpha formats working. I'm not tring to say anything bad about your code, just wanted to show I'm not seeing it working on my machine yet. I'll keep you guys posted on if updating the base code breaks my version as well. Have a great day o/. |
|
So applying my code to the most recent Godot build seems to break mine as well see attached. I see that there was a change to the source but it just looks like a 2bpp implementation which shouldnt change any of the data we are working with. |
|
So I figured out the black boxes. It is due to, as usual me forgetting something. Because I didnt reimport the images and used the incorrect .import cached values. The project was displaying the black boxes I was getting when using your build instead of recalculating the textures with mine. After reimporting mine is back to what it was before. However reimporting didnt change the output of your build/fork from black boxes/textures for some reason when tested with my project. However the test formats you provided still work with both versions, I think this is due to only testing a few forms of 16BPP but I'll keep poking around if you'd like? |
ScorpionInc
force-pushed
the
Add_BMP_16bpp_Support
branch
from
48c748d
to
ef02571
Compare
|
Sorry accidentally closed this while fetching upstream due to a file conflict(This has been fixed). |
|
Thought I'd share a little update. So far I got almost all versions of 16 bit bmps working im only having issues with the X8R8G8B8 format. for some reason the red mask reads a 0x0000 instead of the expected: 0x00FF. Interestingly A8R8G8B8 also does this however it displays correctly where X8 does not. Still working on it but I have masks implemented mostly and thought I'd share. Edit: So the reason I'm going to give up on the X8R8G8B8 format is for one it doesnt fit in the theme of this branch as it's a 32bit format and two im stumped. It's still something I'd like to fix eventually but I'll try it in a seperate branch. Also I found that a combination of lacking color scaling and having moved the code into the switch is what caused the black textures I got when using Anutrix's code. Leaving that part alone and 16BPP works fine. I'll be updating on github soon(Hopefully I don't break it this time.) |
ScorpionInc
force-pushed
the
Add_BMP_16bpp_Support
branch
3 times, most recently
from
1e3de1a
to
d0b0a2f
Compare
ScorpionInc
changed the title
|
|
This commit adds some basic 16bpp support for BMP File Format. Added support for reading and using of 16 bit mask values from file. All values are scaled to ARGB255 format based on bit depth of source color channel. Removed warning, it's no longer required as 16 bit with alpha bit(s) are now supported. Adjusted spacing, added spacing and brackets to make clang static check happy.
ScorpionInc
force-pushed
the
Add_BMP_16bpp_Support
branch
from
56cede7
to
e74a0ef
Compare
|
Thanks! And congrats for your first merged Godot contribution 🎉 |
This commit adds some basic 16bpp support for BMP File Format.
Doesn't support R5G6B5 pixel format yet as I haven't figured out how to detect it's usage seperatly from the A1R5G5B5/X1R5G5B5 pixel formats from the available header information. Should be easy to add support for it though once that's nailed down.
Doesn't support Alpha bit yet. This is because while it seems to still have it's transparency correct when I open test images in Gimp, I have no such luck in Godot and I'm not sure why. I think some relevant information maybe being stored in the colorspace or the V5 BMP header that I'm either ignorant of or have missed.
Edit: Added spacing and line breaks for legability.