I wasn’t sure where to put this but I was curious if anyone knows why Plex Dash and the server dashboard show wildly different bandwidth usage. Like I have someone direct playing something right now and the dashboard shows usage that spikes between 14-29 Mbps with valleys of 0. Plex Dash shows spikes of 7-10 Mbps with valleys at 0. The file itself is a 1080p HEVC 4 Mbps video with 614 Kbps 5.1 AAC audio.
So a little update. I had my users try a bunch of different files last night and they all exhibit the same behavior. The Plex dashboard shows much higher bandwidth usage when compared to the mobile app. It also seems strange to me that they both show this super erratic spikey usage spiking up before dropping to 0 and then bursting up again. Looks almost like a heart monitor.
Actually the “spiky” behavior you see is probably proper operation. When Plex, or just about any streaming app, is streaming it buffers the stream until the buffer that is being used fills up and then it stops the stream from the source until the buffer drops below some threshold then it starts up again until the buffer is full again the the stream is paused again and so on wash rinse and repeat until done. It is done that way so the client device can deliver a steady stream to the display device without being reliant on the stream being “perfect.”
On most networks without the buffering behavior the playback would be quite uneven and jerky.
As far as the display inconsistency goes that would need to be addressed by Plex.
1 Like
I see, that makes sense. I should’ve thought about that more lol. It does still seem interesting that they are such short bursts followed by equally short valleys, like a second or two for each but equal in length, the logic still makes sense though. The stranger thing is the massive discrepancy between the dashboard and the app. I was kind of hoping someone else notices the same thing or unfortunately this thread will probably die without any of the Plex staff even seeing it.
A typical video file has several numbers you can cite as its “bandwidth”.
What you see in the properties of the file and in typical “media info” tools is the average bandwidth, which is calculated by simply dividing the file size by its play duration.
This however doesn’t mean anything, because there is no video file which has a perfectly even bandwith.
Videos are always “VBR” – variable bit rate.
Some scenes require more bandwidth, some less.
So to reliably determine the actually required bandwidth to stream the file without the dreaded “buffering” indicator appearing, Plex needs to determine the real bandwidth spikes.
Which it does during the in-depth analysis during server maintenance, btw.
The bandwidth number in the dashboard is the reserved bandwidth for playing the file – i.e. the “worst case”/highest bandwidth which can arise at one point during the whole length of the video.
What you see in the bandwidth graph of the network adapter is the currently used bandwidth (the above mentioned worst case might happen later in the file or has already passed). But of course this graph is also modulated by the “fill buffer – consume buffer - replenish buffer” cycle, which was explained above already.
If you are curious about those “worst case” bandwidth numbers, take a look at the Plex XML info of a video which has undergone the in-depth analysis. It should show the bandwidth numbers in the requiredBandwidths XML property.
Some additional details can be found here: What is Media Analysis & Do I Need It - #8 by OttoKerner
2 Likes
I appreciate all that great info! Though my question is still centered around why there is a massive discrepancy between the web based Plex dashboard and the Plex Dash app. Maybe I’m misunderstanding but it sounds like you thought I was asking about a difference between what Plex shows and my network adaptor on the server. This is not the case. The numbers I was looking at are Plex based numbers.
1 Like
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.
