Question about AVC sequence header

[levs42]

There's an open source RTMP server that I use for quick testing in the local network: NMS. It parses AVC sequence header when video data arrives here and extracts profile, level, width and height here from the AVC header. When I'm trying to publish using HK, that method fails to extract the data from the header. Looks like the muxer doesn't include much into the header. AVCDecoderConfigurationRecord defines multiple different fields to be included. Would it be beneficial to do so?

[shogo4405]

HaishinKit.swift/Sources/RTMP/RTMPMuxer.swift

Line 43 in 3e1604e

buffer.append(configurationBox)

I understand that the configurationBox value is the value of AVCDecoderConfigurationRecord and is being sent.

This value needs to be used as output by VideoToolbox,
and changing it is not recommended because it might prevent playback on the viewer side.

I think the reason why NMS can't parse it lies elsewhere.

[levs42]

Thank you, I'll look into it. So far it looks a bit strange, "This key is used by sample description bridges to hold sample description extension atoms that they do not recognize."