I agree.In my oppinion extracting the files into subfolders with the correct names and file extensions
And we should start using something like this as our map where each face AND tile is tagged with the same name as the jpg and path:
I agree.In my oppinion extracting the files into subfolders with the correct names and file extensions
Thats the same i was trying to do in my reference image, is just i reduced the identifyers to the minimal (only 3 digits), and im rotating it 180º to have the poles well oriented, compare with yours, is the sameI agree.
And we should start using something like this as our map where each face AND tile is tagged with the same name as the jpg and path:
View attachment 23047
Yes, but with the texture rotated 180ºI am fairly sure this is what they mean or do you not think so? Row then column.
![]()
idk, for me its much easier flip the single image before you start cutting it. All image editing should be done before then.But is much easyer to work with the images with his correct orientation, then flip them automatically with the .bat i posted the other day
And for the reference image i was doing it was a lot easyer to extract all files, them flip them with the .bat ... and simply forget about the fact that the .qrc stores them rotated
Thats the biggest advantage of what i suggesting, there are only 2 points of the workflow where you need to remember that the images needs to be rotated 180º... either inmediatly before extraction... and inmediatly before injection
Actually, if you automatize that 2 actions with a tool you can forget about this problems of the rotations forever
Whats the diference ?, if you do the rotation with the .bat (or any other automated tool) it doesnt matters if the number of rotated images is just 1 or 1000No, its much easier flip the single image before you start cutting it. All image editing should be done before then.
IDK, I just think all image preparation should be done first, and why flip 24 images when its exactly the same to flip one. Also the grid references make sense and it matches every other cubemap i have seen so far.Whats the diference ?, if you do the rotation with the .bat (or any other automated tool) it doesnt matters if the number of rotated images is just 1 or 1000
The .bat solved that problem, now it doesnt matters how many needs to be flippedIDK, I just think all image preparation should be done first, and why flip 24 images when its exactly the same to flip one
Even considering that theory is correct... are we going to live with something that looks like a problem (working with the images rotated 180) just to respect that coordinate system ?, i think is not worthyAlso the grid references make sense and it matches every other cubemap i have seen so far.
For the matter of rotations i dont see any problem, the image from google is going to have the north pole at north...i convert it like that, then i split it... and i open the splitted parts in photoshop to paint on top of them (for that i really need north pole at top)Try apply some planet textures with 24 parts and see what you think. I will test your qrcs if you want.
I think for now we'll focus on making something like this and if it someday needs a noob friendly version it'll be easier to remove functions than to add them. For now I don't see many people using this that don't know what they are doing.Editable by the user or not, such tool should unpacking all files, and rebuild qrc from all. Just in case that in future, someone will push research further and would be a shame if he discover that will be limited by packer/unpacker because the author decided to not unpack all.Jut my three cents.
Fingers crossed for repack tests success. ^^"