I've been wondering how difficult it would be to create a peer-to-peer version of Jamulus. I imagine that it would result in lower latency.
From 10,000 feet, it seems to need to be something like this:
1. All participants run both client and server (modified as below)
2. Client sends the local signal to all other participants with a timestamp
3. Server adds variable delays to all inputs to match timestamps at mixer
4. Server mixer output is sent to local user only.
I'm assuming all participants use NTP to syncronize clocks within a few milliseconds, and there would need to be a way for everyone to identify each other's IP address, etc.
Without having looked at the code, it seems like this wouldn't be a huge project. Do you see any reason it wouldn't work?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I've been wondering how difficult it would be to create a peer-to-peer version of Jamulus. I imagine that it would result in lower latency.
From 10,000 feet, it seems to need to be something like this:
1. All participants run both client and server (modified as below)
2. Client sends the local signal to all other participants with a timestamp
3. Server adds variable delays to all inputs to match timestamps at mixer
4. Server mixer output is sent to local user only.
I'm assuming all participants use NTP to syncronize clocks within a few milliseconds, and there would need to be a way for everyone to identify each other's IP address, etc.
Without having looked at the code, it seems like this wouldn't be a huge project. Do you see any reason it wouldn't work?
See this thread: https://sourceforge.net/p/llcon/discussion/533517/thread/e6a9bb76e8
Here is more to the subject
https://sourceforge.net/p/llcon/discussion/server/thread/cc992c155c/#5611/2f75