Input and Output

Inputting Video/Image Data

GPUPixel supports different input formats:

RGBA Data Input

SourceRawData accepts pixel data and runs the filter chain via ProcessData. The last parameter is the pixel format:

// For RGBA data (use GPUPIXEL_FRAME_TYPE_RGBA or GPUPIXEL_FRAME_TYPE_BGRA)
source_raw_input_->ProcessData(pixels, width, height, stride,
                              GPUPIXEL_FRAME_TYPE_RGBA);

Example with camera frame processing:

gpuPixelRawInput->ProcessData(pixels, width, height, stride,
                              GPUPIXEL_FRAME_TYPE_RGBA);

Currently SourceRawData only supports RGBA/BGRA; convert YUV to RGBA before passing to ProcessData.

Getting Output Data

After the pipeline has run (e.g. after calling ProcessData or SourceImage::Render()), read the result from SinkRawData:

// Get RGBA result
const uint8_t* rgba_data = target_raw_output_->GetRgbaBuffer();
int width = target_raw_output_->GetWidth();
int height = target_raw_output_->GetHeight();
size_t rgba_size = width * height * 4;
// Use rgba_data for display or further processing

// Get I420 result
const uint8_t* i420_data = target_raw_output_->GetI420Buffer();
size_t y_size = width * height;
const uint8_t* uData = i420_data + y_size;
const uint8_t* vData = i420_data + y_size + y_size / 4;