I’m using pixel streaming in an arcvhiz project where I’m working on 5 to 20 instances, but even using a few, we feel a lot of input lag. Does anyone know of a way to optimize pixel streaming, or a configuration that could help?
Note: We have 1GB of internet, with 500MB upload speed, and an i9 14700k computer, 64GB of DDR5 RAM, and a 4070 12GB.
Hi there!
Unfortunately, input lag is somewhat expected when running numerous streamers/instances from a single machine. This is something we’re fixing for 5.5, but is difficult to work around in the meantime.
I suggest looking into the SFU to better manage your stream resources. This should help ensure stream stability and quality with multiple instances.
I’ve created a simple video here going through rudimentary setup and explanation:
Additionally, there is some documentation on it here:
Please let me know if you had any further questions!
Hello @luhhiiz,
Selective Forwarding Unit is something you can configure and optimize for a better pixel streaming experience. Plus, performance bottlenecks on the server can be the reason for input lag during the stream as well.
Without SFU, because streaming will try to force your internet connection limits to increase the streaming quality, you may face some lags while using it. In addition to that, if you are far away from the server location, it might be another reason as well, physical distance to your server also matters.
If you don’t want to deal with SFU, or server orchestration, you can check Vagon Streams.
Our internally developed interactive streaming solution adapts itself according to the network conditions of your user by keeping 60FPS & 4K streaming quality, and we do the server/region orchestration process on behalf of you to serve your application from the closest server to your users.
If you need any assistance, I will be happy to help.
Best,
Serdar