Overview
Our customer wanted the frames captured from Blackmagic card to be transferred on GPU memory for further faster image processing.
Challenges
- Understanding the Interfaces provided by Blackmagic.
- Understanding how the frames captured can be transferred to GPU memory directly by using OpenGL texture.
- Understand the CUDA to verify whether it can be used with the gstreamer pipeline to transfer the data.
Solutions
- Using Blackmagic interface we can get the captured frames.
- We used CUDA to transfer the frames from CPU memory to GPU memory.
- Further, we sent the frames present on GPU memory to gstreamer pipeline, which was able to render the captured frames.
Outcome
- We were able to get the frames on GPU memory.
Technical know hows and understanding
- Understanding the interfaces and workflow provided by Blackmagic.
- Understanding the functionality of CUDA so that it can be used with gstreamer to get the data on GPU memory.