SDL2 纹理的像素格式和表面颜色遮罩

SDL2 texture's pixel format and surface color mask

本文关键字:表面 颜色 格式 纹理 像素 SDL2      更新时间:2023-10-16

我试图创建一个用于像素操作的SDL_Surface,但是当设置表面的颜色蒙版时,因为填充颜色缓冲区时颜色不正确(请参阅评论,我试图将u8vec4解释为RGBA颜色):

const int win_width = 640;
const int win_height = 480;
std::vector<glm::u8vec4> colors{ win_width * win_height, glm::u8vec4{ 0, 0, 0, 0 } };
SDL_Surface* render_surface = SDL_CreateRGBSurfaceFrom(&colors[0], win_width, win_height, 32,
    sizeof(glm::u8vec4) * win_width, 0xFF000000, 0x00FF0000, 0x0000FF00, 0x000000FF);
SDL_Texture* render_texture = SDL_CreateTexture(renderer, SDL_PIXELFORMAT_RGBA8888, 
    SDL_TEXTUREACCESS_STREAMING, win_width, win_height);
// { 255, 0, 0, 255 } = Red
// { 255, 0, 0, 100 } = Darker red, alpha seems to be working
// { 0, 255, 0, 255 } = Purple!? Should be green?
// { 0, 0, 255, 255 } = Yellow, expecting blue
std::fill(colors.begin(), colors.end(), glm::u8vec4{ 0, 0, 0, 255 }); // Red
SDL_UpdateTexture(render_texture, nullptr, render_surface->pixels, render_surface->pitch);
SDL_RenderClear(renderer);
SDL_RenderCopy(renderer, render_texture, nullptr, nullptr);
SDL_RenderPresent(renderer);

您应该尝试根据端位数设置位标记。我是这样做的:

#if SDL_BYTEORDER == SDL_BIG_ENDIAN
    const uint32_t R_MASK = 0xff000000;
    const uint32_t G_MASK = 0x00ff0000;
    const uint32_t B_MASK = 0x0000ff00;
    const uint32_t A_MASK = 0x000000ff;
#else
    const uint32_t R_MASK = 0x000000ff;
    const uint32_t G_MASK = 0x0000ff00;
    const uint32_t B_MASK = 0x00ff0000;
    const uint32_t A_MASK = 0xff000000;
#endif

在你的例子中,你得到了颜色蒙版的另一种方式。


我认为Jongware是正确的,格式是ABGR。

// { 255, 0, 0, 100 } = Darker red, alpha seems to be working

我猜这里的alpha看起来是正确的,因为背景是黑色的,通过减少红色的数量到100,它会自然地与背景融合。

我还认为你没有设置SDL_BlendMode