I’m having some strange performance results that I hope someone here can clear up for me.
For reference, here is my setup:
Client: Apple TV 4K - Gigabit
Server: Jellyfin Host - 10GbE
NAS: TrueNAS - 10GbE
Infuse 8.0.8 > Jellyfin (Library Mode Off) > Media Content Fetched from NAS
Tests between the Jellyfin host and the file server confirm 10GbE speeds, both via synthetic benchmarks (CrystalDiskMark, iperf3, etc.) and through manual testing.
Now with the boring parts out the way, let’s get to the heart of the issue.
Before updating to 8.0.8, my speed test results when connected to my Jellyfin server would result in 900Mb/s+. Unfortunately, I do not remember what version of Infuse I was running at the time.
After updating to 8.0.8, my speed tests are around 250Mb/s average.This performance decrease is also noticeable when scrubbing through large files.
Here’s where it gets a little weird…
If I bypass Jellyfin altogether on my Infuse Client, and instead opt to connect directly to my NAS via SMB, I get the following results.
Again, I was getting 1Gb/s consistently in Infuse before updating to 8.0.8. Nothing else has changed on my end apart from that. But now I’m getting 250Mb/s when connecting to Jellyfin and 550Mb/s when connecting to my NAS. I’ve historically only used the Jellyfin route and it’s been rock solid at 1Gb/s until this latest update.
Are you using the speed test in Infuse when you see these numbers? If so, the Infuse speed test isn’t meant to give you max network speed, it’s simulating playing the video with the back and forth traffic needed to correctly stream the video so it will be less than the raw network speeds.
Yes via Infuse. But, again, that same test used to deliver 1Gb/s before this latest update - even with the largest, most high-bitrate content I have to throw at it (over 100GB 4K HDR files).
And it’s very noticeable when streaming content. Large files now buffer a lot more during seeks, whereas before they would load instantly. So the changes in speed are not purely synthetic but are actually measurable in daily use.
Based on the two tests (Jellyfin vs SMB), you can tell there’s something strange going on because the Jellyfin test results are all over the place and are vastly under-performing compared to the SMB results. I’ve historically always used the Jellyfin server as my endpoint, so this hasn’t been an issue. I can’t explain why the SMB is still roughly half as fast as it should be, but I never tested SMB speed in the past so I don’t know if this result is any different compared to pre-8.0.8
Does the Infuse app on macOS have a built-in ‘Speed Test’ feature like the tvOS version?
Streaming from my Macbook (wireless) results in a much faster, more responsive experience. I suspect the Infuse app on my Mac isn’t suffering from whatever happen to my TV post 8.0.8. Even though it’s most likely a 500-600Mb/s connection, seeks are instant and content loads instantly with little to no buffering in-between.
Something definitely changed after updating on the AppleTV, and re-installing the Infuse app hasn’t improved the results. Wish there was someway for me to revert back to the previous version. Is there a beta access available for me to test this out?
I dont understand the point here, even a full bitrate 4K disc wont need more than 100mbs, you playing videos, not copying them locally. If the videos play, they play.
I wasn’t able to replicate lower speeds like you are seeing, so it may be something related to the network or perhaps NAS/Jellyfin settings. Here’s a screenshot streaming from a vanilla installation of Jellyfin (10.10.3) on a remote server that is hosted behind Cloudflare with Infuse 8.0.9.
A standard Internet speed test run right after shows a speed of 801 Mbps, so getting ~86% of that from a remote server seems pretty reasonable.
My car has a 200mph top speed. I drive on the same road every day, and hit my top speed. Suddenly, I’m only able to hit 65mph. On every other road, I can get to 200mph - but just not on this one, which is odd cause I never had issues with it before.
Your response is akin to: “What’s the point? The speed limit is 35mph anyway.”
Brushing aside the legality of my analogy, the point is that it used to work - ran the update and now it doesn’t. Obviously something changed, and the changes are not just synthetic. There are very obvious differences in performance that I’m experiencing, which is what prompted me to look into the speed tests to begin with. That’s the point.
I don’t want to seem rude or abrasive, but commenting on someone’s help request with “I don’t see the point here”, is not helpful or insightful given the nature of the problem. I am well aware that videos still play - that was never the issue. That being said, I appreciate that you attempted to help, if that was your intention, but I’m afraid I’m still facing the same issue. Cheers!
Given that you’re running it behind a proxy and still getting those speeds shows that something is up on my end, cause I’m accessing it locally and don’t face that additional overhead. I’m going to try and setup a testing environment with a fresh install of Jellyfin and see if I can narrow down the issue.
