RTSP sink output Jittery

Please provide complete information as applicable to your setup.

• Hardware Platform : RTX 4090
• DeepStream Version: 6.2 container
• NVIDIA GPU Driver Version: 525.147.05
• Issue Type: bug

I am using deepstream-app to generate RTSP video output and playing it on VLC. I managed to view the streams on VLC at http://localhost:8556/ds-test and http://localhost:8557/ds-test on the host system. However the stream is jittery when I play it and freezes often. Here is my config file:

[application]
enable-perf-measurement=1
mux-pool-size=8
perf-measurement-interval-sec=5

[tiled-display]
enable=2
rows=1
columns=1
width=1280
height=720
gpu-id=0
nvbuf-memory-type=0

[source1]
enable=1
type=3
uri=file://sample.mov
#uri=file:///opt/nvidia/deepstream/deepstream-6.2/videos/wah_train_0018.mov
num-sources=1
gpu-id=0
cudadec-memtype=0

[source2]
enable=0
type=3
uri=file://sample.mov
#uri=file:///opt/nvidia/deepstream/deepstream-6.2/videos/wah_train_0018.mov
num-sources=1
gpu-id=0
cudadec-memtype=0

[sink0]
enable=1
type=1
sync=4
gpu-id=0
nvbuf-memory-type=0

[sink1]
enable=1
type=4
gpu-id=0
codec=1
container=1
#output-file=sample_out.mp4
enc-type=0
sync=0
bitrate=4000000
profile=0
rtsp-port=8556
udp-port=5400
udp-buffer-size=2000000
nvbuf-memory-type=0
qos=0
latency=1000
rtsp-reconnect-interval-sec=30
link-to-demux=1

[sink2]
enable=0
#Type - 1=FakeSink 2=EglSink 3=File 4=RTSPStreaming
type=4
gpu-id=0
#1=h264 2=h265
codec=1
container=1
#output-file=sample_out.mp4
#encoder type 0=Hardware 1=Software
enc-type=0
sync=0
bitrate=4000000
#H264 Profile - 0=Baseline 2=Main 4=High
#H265 Profile - 0=Main 1=Main10
profile=0
# set below properties in case of RTSPStreaming
rtsp-port=8557
udp-port=5400
udp-buffer-size=100000
nvbuf-memory-type=0
link-to-demux=1


[osd]
enable=1
gpu-id=0
border-width=5
text-size=15
text-color=1;1;1;1;
text-bg-color=0.3;0.3;0.3;1
font=Serif
show-clock=0
clock-x-offset=800
clock-y-offset=820
clock-text-size=12
clock-color=1;0;0;0
nvbuf-memory-type=0

[streammux]
gpu-id=0
live-source=0
batch-size=1
batched-push-timeout=40000
width=2560
height=1440
enable-padding=0
nvbuf-memory-type=0

[primary-gie]
enable=1
gpu-id=0
gie-unique-id=1
nvbuf-memory-type=0
config-file=config_infer_primary_yoloV5_saipem.txt

[tracker]
enable=0
gpu-id=0
tracker-height=320
tracker-width=120
ll-lib-file=/opt/nvidia/deepstream/deepstream-6.2/lib/libnvds_nvmultiobjecttracker.so
ll-config-file=config_tracker_NvDCF_max_perf.yml


[tests]
file-loop=0

I am using YoloV5 inference and am trying to stream the bounding box stream via RTSP.

Could you share places where i have to troubleshoot to find why my stream is jittery?

Can you monitor the GPU and CPU loading while running the case?
You can use the command “nvidia-smi dmon” for GPU loading monitoring.

Here is the output when running nvidia-smi dmon:

gpu pwr gtemp mtemp sm mem enc dec mclk pclk

Idx W C C % % % % MHz MHz

0    314     76      -     98     25     27     35   9501   1905 
1     20     38      -      0      0      0      0    405      0 
0    314     77      -     98     27     28     38   9501   1890 
1     20     38      -      0      0      0      0    405      0 
0    312     77      -     98     25     26     34   9501   1905 
1     22     39      -      0      0      0      0    405    210 
0    316     77      -     99     26     29     34   9501   1905 
1     24     39      -      0      0      0      0    405      0 
0    341     78      -     99     38     45     82   9501   1875 
1     21     38      -      0      0      0      0    405      0 
0    351     79      -    100     41     49    100   9501   1800 
1     20     38      -      0      0      0      0    405      0 
0    344     78      -     98     29     30     46   9501   1905 
1     20     38      -      0      0      0      0    405      0 
0    322     78      -     99     26     29     44   9501   1890 
1     20     38      -      0      0      0      0    405      0 
0    322     78      -     98     27     30     34   9501   1905 

After a while the sm% drops like so:

gpu pwr gtemp mtemp sm mem enc dec mclk pclk

Idx W C C % % % % MHz MHz

0     79     60      -     46      1      0     22   5001   1365 
1     20     39      -      0      0      0      0    405    210 
0     78     59      -     14      2      0     54   5001   1365 
1     23     39      -      0      0      0      0    405      0 
0    125     62      -     29      1      0     23   9751   1950 
1     19     39      -      0      0      0      0    405      0 
0    143     62      -     13      1      0     24   5001   1380 
1     18     38      -      0      0      0      0    405      0 
0     78     59      -     55     10      0     41   9501   1590 
1     18     38      -      0      0      0      0    405      0 
0     99     60      -     23      1      0     38   9501   1590 
1     18     38      -      0      0      0      0    405      0 
0     95     59      -     57     10      0     30   9501   1575 
1     18     38      -      0      0      0      0    405      0 
0    105     60      -     25      1      0     20   9501   1575 
1     20     39      -      0      0      0      0    405    210 
0    123     61      -     24      1      0      1   9751   1950 
1     23     39      -      0      0      0      0    405      0

My stream looks like this with a lot of artifacts

Hi, is there anything abnormal from the nvidia-smi output?

“sm” is the GPU loading.
Is the picture you captured with the GPU loading 98%~100% or with the GPU loading is 29%~57%?

The picture is always like that regardless of the sm. Initially it starts of smooth but eventually where is a box drawn the box starts to “bleed”, leaving a trail of red everywhere.

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks

Where and how did you display the output RTSP video? On the same machine with the deepstream-app with “localhost”?

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.