调整MFT问题：Windows Media Foundation中的视频合成

Question

I'm trying to do composition with two separate video sources in Media Foundation. I am attempting to encode a video with a video overlay. To do so I am attempting to use the Video Resizer on the smaller input.

I've seen several threads on this, but I thought I'd ask around in any case.

Basically the idea is to create two source readers and a sink writer. The source files are h264, so I use the reader to decode into YUY2. While processing samples, I send the appropriate sample to the Resize MFT, then down the line (I haven't made it this far) I combine the two images to create the overlay effect with MFCopyImage.

My question is: I am getting an E_INVALIDARG when I call ProcessInput on the Resize MFT.
To initialize the mft, I am giving it the appropriate type from the reader via SetInput Type. After that I am setting all the appropriate properties via the PropertyStore, and then updating the framesize for the output type of the MFT. I have read the documentation and modeled my implementation according to the MFT Processing Model .

None of these steps raise any red flags until I actually attempt to use ProcessInput.

Although I have limited experience in Windows Media Foundation, I have been able to use the Framerate DSP with success. I would appreciate any advice. Thank you!

Answer 1

对于陷入类似情况的其他人，我最终没有使用Resizer MFT，而是使用了工作量少得多的视频处理器MFT 。