roommini / realtime-video-quality-monitoring

Real-time Video Quality Monitoring

Architecture

1. Video sources broadcasts video streams
2. Dedicated containers with ffprobe generate statistics about the video stream
3. The statistics are piped to custom prometheus exporters
4. Prometheus probes these exporters and collects metrics about the video streams
5. The metrics are visable in real-time in Grafana

Deployment:

1. setup Grafana and prometheus by deploying the docker stack:

docker stack deploy rtvqm -c docker-stack.yml

2. build the exporter via:

go build main.go collector.go

3. stream webcam to ffprobe and pipe the results to the exporter via:

ffprobe -f lavfi movie=/dev/video0,signalstats="stat=tout+vrep+brng",deflicker=bypass=1  -show_frames -print_format csv | ./main -port 2112

Development

• build docker image:

docker image build -f Dockerfile . -t rtvqm:latest 

snippets:

capture stream from webcam:

ffmpeg -f v4l2 -framerate 25 -video_size 640x480 -i /dev/video0 -f mpegts udp://127.0.0.1:9999

probe udp stream:

ffprobe -f lavfi movie="udp\\\://127.0.0.1\\\:9999",signalstats="stat=tout+vrep+brng",deflicker=bypass=1 -show_frames

build and probe:

go build main.go collector.go && ffprobe -f lavfi movie="udp\\\://127.0.0.1\\\:9999",signalstats="stat=tout+vrep+brng",deflicker=bypass=1,bitplanenoise -show_frames -print_format csv | ./main -port 2112

play captured stream:

ffplay udp://127.0.1:9999

QCtool probing command:

ffprobe -f lavfi -i "movie=EXAMPLE.mov:s=v+a[in0][in1],[in0]signalstats=stat=tout+vrep+brng,cropdetect=reset=1:round=1,idet=half_life=1,deflicker=bypass=1,split[a][b];[a]field=top[a1];[b]field=bottom,split[b1][b2];[a1][b1]psnr[c1];[c1][b2]ssim[out0];[in1]ebur128=metadata=1,astats=metadata=1:reset=1:length=0.4[out1]" -show_frames -show_versions -of xml=x=1:q=1 -noprivate | gzip > EXAMPLE.mov.qctools.xml.gz