RGBA32F

RGBA32F is a texture or render target format used in computer graphics that stores color and alpha as four 32-bit floating point numbers. Each pixel consists of red, green, blue and alpha components, totaling 128 bits (16 bytes) per texel. This format provides full floating-point precision per channel, enabling a wide dynamic range and precise color and transparency values beyond the conventional 0.0 to 1.0 range.

Applications for RGBA32F include high dynamic range rendering, physically based rendering, post-processing, color grading, and scientific visualization. It supports linear color data and is suitable for preserving HDR data through various stages of a graphics pipeline, such as compositing and tone mapping. Its floating-point precision makes it a common choice for deferred shading and other workflows that require accurate color arithmetic.

Support and API mappings vary by graphics API. In OpenGL, the internal format is GL_RGBA32F. In DirectX, it corresponds to DXGI_FORMAT_R32G32B32A32_FLOAT. In Vulkan, VK_FORMAT_R32G32B32A32_SFLOAT is used, and in Metal, MTLPixelFormatRGBA32Float is the equivalent. When uploading data, matching formats and types (for example, GL_RGBA with GL_FLOAT) are required across APIs.

Considerations for using RGBA32F include higher memory usage and bandwidth compared to fixed-point formats, which can impact performance on bandwidth-constrained paths. It is generally used in modern GPUs that support floating-point render targets and is typically employed with linear color space before any gamma correction. While providing precision and HDR capabilities, careful management of tone mapping and color workflows is important to ensure proper display results.