Slow speed test results (ATV / Jellyfin)

Latest version of Infuse running on Apple TV.
Jellyfin server running on Windows 11.
Infuse is connected to the share via Jellyfin.

Both devices are wired.
Both devices (according to speedtest) net an average of 980 down / 40 up
ATV via Infuse (according to Infuse’s speed test) net average of 15-20 down.

Playing media via a web browser on any device in the house = great speeds and results
Playing media via Infuse on the ATV = slow speeds and buffering.

Not entirely sure why this is happening. I tried adjusting the streaming cache, trying both Memory Only and Legacy, and while Memory Only may have helped a tiny bit with playback, the speeds are still unbearably slow.

What gives?

Thanks!

1 Like

Can you share a screen cap of the graph after running a full test on one of the problem files? Also a screen cap of the result speed numbers of that same test would help.

I didn’t bother with waiting for the whole file, since that would take way too long. Hopefully this will suffice!
Same device according to speedtest.net averages about 980 down. This sounds about right, since I have a 1Gb connection and all my wired devices average that speed.

Thanks for your help!

The speed tests run considerably faster than the actual video and will give a better graph of what’s going on. With less than 1% it could be several things, first thought would be that something on the server isn’t allowing the transfer to Infuse as fast as needed.

A test would be to try sharing from the server via say SMB instead of jellyfin and see if that changes how the speeds go.

Found a smaller file to test, so wouldn’t take too long.
It actually averaged better, but still nowhere close to where it ought to be.


Not sure if it’s a fluke or what, but if I stop a test and start it again, for a few seconds I see speeds upwards of 600-700, but it rapidly declines until it settles down low again. This will happen every time I stop/start a test. Again, could just be a fluke.

I’ll try via SMB only and will report back. Thanks!

As an SMB share, I get substantially better results. Not quite as high as I would expect, but I suppose this could also be a limitation of the storage media. In that case, this looks about right.

What gives? Is it common for speeds to be neutered when connected via JF? Or is there some setting somewhere I should be taking a look at? Thanks!


I don’t use any server like jellyfin or Plex so I can’t really say as to what’s normal but many users do and have no problems. I’ve run straight SMB for years and it’s been sufficient for everything I’ve watched so I’m happy. Others swear by FTP or other protocols but stick with what works.

The biggest thing you need to take away is that playback is the gauge to use for “adequate” speed, not “max” theoretical. As long as your videos play without issues then don’t worry about the speeds.

The JF functionality is great, not just for syncing watched progress between all my devices, but also concerning metadata (including artwork) for my library.

Playback is an issue with larger files. I have some 4K rips that damn near take forever to load and make for a terrible viewing experience. Over SMB, works just fine - but not when connected via JF.

I just signed up for Trakt and signed in on both Infuse and on my JF server. Will this, at least, sync my watch status? It might be a good compromise - albeit I would prefer to figure out what on earth is going on.