Yet Another Media Infrastructure. it is core part of media codec with hardware acceleration, it is yummy to your video experience on Linux like platform.
Apache License 2.0
147
stars
106
forks
source link
request for libyami SVC-t feature enhancement #750
This is libyami SVC-t feature enhancement from YunOS. The use case is webrtc, it may need to drop some frames due to bad network condition. They have basically accepted the SVC-t feature we provided in libyami for this requirement, but some enhancement is requested.
As we know, Libyami has already support SVC-t feature, and supported 4 layer maximally, but the every layer’s frame rate is fixed which introduces higher bitrate compared to AVC mode with same encode quality. This is because in AVC mode every P frame refers the nearest P frame before it, while in SVC-t mode the reference model is different. In current implementation, take 4 layers as example, 1/2 frames are not be referenced, 1/4 frames are referenced once, 1/8 frames are referenced twice, and rest 1/8 frames are referenced four times. The reference model is fixed currently. If we could supports different frames rates for different layers, it will be more flexible to control the encoded videos bitrate.
This is libyami SVC-t feature enhancement from YunOS. The use case is webrtc, it may need to drop some frames due to bad network condition. They have basically accepted the SVC-t feature we provided in libyami for this requirement, but some enhancement is requested.
As we know, Libyami has already support SVC-t feature, and supported 4 layer maximally, but the every layer’s frame rate is fixed which introduces higher bitrate compared to AVC mode with same encode quality. This is because in AVC mode every P frame refers the nearest P frame before it, while in SVC-t mode the reference model is different. In current implementation, take 4 layers as example, 1/2 frames are not be referenced, 1/4 frames are referenced once, 1/8 frames are referenced twice, and rest 1/8 frames are referenced four times. The reference model is fixed currently. If we could supports different frames rates for different layers, it will be more flexible to control the encoded videos bitrate.
Could you please support this? Thanks in advance