Compressing video using created pixel buffer in kCVPixelFormatType_420YpCbCr8BiPlanarVideoRange format

Question

I have a bi-planar raw video buffer data in YCbCr format, which I'm using as source to compress one new mp4/mov video in H.264 format on iPhone and iPad. Therefore, I'm using a CVPixelBufferRef to make a new pixel buffer and then use an AVAssetWriterInputPixelBufferAdaptor to append to the video writer.

However, when I'm calling appendPixelBuffer with the new pixel buffer with YCbCr data in it, appendPixelBuffer is only returning YES for the first frame appended. All the other frames are denied by the writer (by returning a NO). The problem is that if I use BGRA32 format raw video data, it's working just fine, so I wonder whether I have a wrongly created pixel buffer with YCbCr format.

I had two methods to create the pixel buffer:

1) Use CVPixelBufferCreateWithBytes:

cvErr = CVPixelBufferCreateWithBytes(kCFAllocatorDefault,
                                     videoWidth_,
                                     videoHeight_,
                                 kCVPixelFormatType_420YpCbCr8BiPlanarVideoRange,
                                     [videoFrame getBaseAddressOfPlane:0],
                                     [videoFrame getBytesPerRowOfPlane:0],
                                     NULL,
                                     NULL,
                                     NULL,
                                     &pixelBuffer);

2) Use CVPixelBufferCreateWithPlanarBytes:

cvErr = CVPixelBufferCreateWithPlanarBytes(kCFAllocatorDefault,
                                           videoWidth_,
                                           videoHeight_,
                                 kCVPixelFormatType_420YpCbCr8BiPlanarVideoRange,
                                           NULL,
                                           0,
                                           videoFrame.planeCount,
                                           planeBaseAddress,
                                           planeWidth,
                                           planeHeight,
                                           planeStride,
                                           NULL,
                                           NULL,
                                           NULL,
                                           &pixelBuffer);

planeBaseAddress, planeWidth, planeHeight, planeStride are 2D arrays containing the base address, width, height and stride of the Y plane and the CbCr plane.

So can you show me where I'm doing wrong, or if there is some sample code I can refer to, or if this is iPhone SDK's issue?

Compressing video using created pixel buffer in kCVPixelFormatType_420YpCbCr8BiPlanarVideoRange format

0 Answers0