Expand description
This library contains a safe FFI for NVFBC from NVIDIA.
§Supported GPUs
As this uses a proprietary NVIDIA API, the supported devices are limited to NVIDIA GPUs. Officially the NVFBC API is only supported on GRID, Tesla, or Quadro X2000+ GPUs. Unofficial support is provided for GeForce GPUs by setting magic private data, similar to https://github.com/keylase/nvidia-patch/blob/master/win/nvfbcwrp/nvfbcwrp_main.cpp.
§Supported capture types
Currently only CUDA and system (RAM) capture types are supported.
§Example: Saving an image.
use nvfbc::{SystemCapturer, BufferFormat};
use nvfbc::system::CaptureMethod;
fn main() -> Result<(), Box<dyn std::error::Error>> {
let mut capturer = SystemCapturer::new()?;
let status = capturer.status()?;
println!("{:#?}", capturer.status()?);
if !status.can_create_now {
panic!("Can't create a system capture session.");
}
capturer.start(BufferFormat::Rgb, 30)?;
let frame_info = capturer.next_frame(CaptureMethod::Blocking)?;
println!("{:#?}", frame_info);
let image = image::ImageBuffer::<image::Rgb<u8>, &[u8]>::from_raw(
frame_info.width,
frame_info.height,
frame_info.buffer,
).unwrap();
image.save("frame.png")?;
println!("Saved frame to 'frame.png'.");
capturer.stop()?;
Ok(())
}
§Future work
Support for configuration is currently limited, to keep the code simple and concise. Future releases will add more configuration options.
Re-exports§
pub use cuda::CudaCapturer;
pub use system::SystemCapturer;
Modules§
Structs§
- Box
- Box used to describe an area of the tracked region to capture.
- Error
- Output
- Describes an RandR output.
- Size
- Size used to describe the size of a frame.
- Status