loading sdl surfaces from memory

Started by
6 comments, last by Exab 16 years, 9 months ago
For testing purposes I am trying to load a simple byte array into an SDL_Surface so I may display it on the screen. Libraries like opengl and direct3d provide concise calls for this purpose, but the only things I have seen in SDL are for loading surfaces from a file. I have tried allocating my own SDL_Surface and modifying its data, but I get a runtime error even when I change something as minor as the integral width of my data (and this is simply assigning the value, not rendering the surface). In short I have two pieces of data... int numBytes; unsigned char* buffer; and I want to be able to load that into an SDL_Surface.. any advice would be greatly appreciated.
Advertisement
Look into the RWops functions of SDL.

Here's a page with 3 tutorials, one using file pointers, one using zlib to load into memory and then create a surface...

http://www.kekkai.org/roger/sdl/rwops/rwops.html

And here's the documentation on the RWops functions: http://www.libsdl.org/cgi/docwiki.cgi/SDL_5fRWops

These functions are probably what you'll end up using:

SDL_RWops * SDLCALL SDL_RWFromMem(void *mem, int size);
SDL_RWops * SDLCALL SDL_RWFromConstMem(const void *mem, int size);

I hope this helps, it should be able to do what you want.

Edit: Once you have a SDL_RWops struct made, you can use the good old SDL_LoadBMP() with the _RW suffix added: SDL_LoadBMP_RW, the RW functions also exist in the SDL_image library so you can use IMG_Load_RW (if my memory is still good).
Thanks Exab, that is leading me in the right direction..

The thing is I tried the following test example which you think would be straightfoward, but I get a runtime error as soon as I try to access the width of my created surface.

	int numBytes = 20;	BYTE* buffer = new BYTE [numBytes];	for( int i = 0; i < numBytes; i++ )		buffer = i;	SDL_RWops* rwop = SDL_RWFromMem( buffer, numBytes );	SDL_Surface* textGFX = SDL_LoadBMP_RW( rwop, 0 );        //executable will crash on next line	int w = textGFX->w;


Note: I would prefer to not have to use files an an intermediate and do directly from new memory on the heap.
I haven't used those functions for a while, but I think you actually need the right headers and stuff if you use something like SDL_LoadBMP_RW.

A series of random bytes can't tell SDL the width of an image because that is usually stored in the headers of the BMP, it tells after how many bytes it should switch to a new pixels row. So unless you actually have the data you want to show, you'll want to use something else.

-----------

Here's what I should have thought of at first:

SDL_Surface *SDL_CreateRGBSurfaceFrom(void *pixels, int width, int height, int bitsPerPixel, int pitch, Uint32 Rmask, Uint32 Gmask, Uint32 Bmask, Uint32 Amask);

link to docs: http://www.libsdl.org/cgi/docwiki.cgi/SDL_5fCreateRGBSurfaceFrom

That should do it.
Ok that works thank you! I can see my surface being displayed as a 100x100 rect on the screen, but it only works for colors 0 and 255 (black and white)... I get something that looks like animated tv noise anywhere in the middle. This is understandable since I need to appropriately the set the color component masks and I can't find any examples, especially for doing a greyscale version.

	int numBytes = 10000;	BYTE* buffer = new BYTE [numBytes];	int depth = 8;	int width = 100;	int height = 100;	for( int i = 0; i < numBytes; i++ )		buffer = 100;	SDL_Surface* textGFX = SDL_CreateRGBSurfaceFrom( buffer, width, height, depth, width, 1, 1, 1, 1 ); 


It turns out that SDLCreateRGBSurface creates only a 1 bit palette if the depth is 8 bits per pixel. So seems like I just need to use SDL_SetPalette to create an 8 bit 256 color grayscale palette. Thanks a lot for the help Exab!
Quote:Original post by Khaos Dragon
It turns out that SDLCreateRGBSurface creates only a 1 bit palette if the depth is 8 bits per pixel. So seems like I just need to use SDL_SetPalette to create an 8 bit 256 color grayscale palette. Thanks a lot for the help Exab!


Are you sure of the values of Rmask, Gmask, Bmask & Amask? I'm not familiar with 8 bit surfaces with SDL, but 1 for each seems wrong.
Usually, from all the examples on the net, the masks are done like this:


-- For a 32 bpp depth --
Uint32 rmask, gmask, bmask, amask;

#if SDL_BYTEORDER == SDL_BIG_ENDIAN
rmask = 0xff000000;
gmask = 0x00ff0000;
bmask = 0x0000ff00;
amask = 0x000000ff;
#else
rmask = 0x000000ff;
gmask = 0x0000ff00;
bmask = 0x00ff0000;
amask = 0xff000000;
#endif

SDL_Surface* textGFX = SDL_CreateRGBSurfaceFrom( buffer, width, height, depth, width, rmask, gmask, bmask, amask );

----

I'm not sure how you would do a 8 bpp mask, but I guess this would work:

rmask = 0xff;
gmask = 0xff;
bmask = 0xff;
amask = 0;

This should make a grayscale image without alpha, I think.

This topic is closed to new replies.

Advertisement