ffmpeg - Manually generate "empty" h264 p-frame -

let's call p-frame frame empty if doesn't change pixels in decoded video (i.e. no motion vectors, nothing).

what need able manually insert empty p-frame video stream on request (need send frames streaming client constant framerate, frame source on streaming server can supply frames different/lower one).

so need able correctly form byte sequence represents empty p-frame current resolution (i wonder other parameters needed?)

ideally, prefer have encoder-independent solution, since use 2 different ones: nvenc via nvidia video sdk , x264 via ffmpeg.

where should to?

i think there h264 nal unit dedicated data padding (nal_unit_type: 12 : filler_data_rbsp( )). might useful you.