Hi,

I encountered, that when I load a 24bit TGA from file, that A8 returns as format 8888, while bytespp is = 3; this is, I guess, because of the conversion for being used with DirectX or so... right?

Well, because I am writing some functions for saving BMAP*'s to file, I came up with this solution:



So, whenever I get a format that possibly implicates an alpha channel, I also require the (original) bytespp field to be = 4.

Is this correct for all loaded images from file and all procedurally generated images?

DirectX loads all images in the format that is used by the 3D hardware. So there is no fixed rule which format you get. It depends on your hardware.

I don't think that there is a 3D card that would support the 888 format, so a 24 bit image is most likely loaded in 8888 format, even when the image contains no alpha channel.

Thanks for the answer... I found out, that DirectX stores BMP and PCX images as 888, though.