In my current game I’m working on, I managed to get networking fixed (thanks psspeed!) but there is an another issue. When for whatever reason the game lags (<10 fps) the translations are much smaller than a high frame rate action. I would like to scale the translations/rotations to larger/smaller framerates so there is no real difference and hopefully lessen the divergence of translations on the local/remote machines. Not to mention, I would like to set my movement constants to larger numbers (assuming worse case 1fps) rather than floats in the .5-.001 range.
I have done this by multiplying all vectors (and inputlistening inputs) by tpf in the update loop, and I have managed to regain what was my previous speed before by multiplying the previous max speed, max accel by obscene amounts (many thousands greater). I have yet to test in in a low framerate environment but I fear the possibilities.
Any thoughts, would this help syncronization (assuming sending of relative transforms over server) between remote/local machine states?
Thanks
edit: I figured it out…it was so obvious I feel pretty stupid right now…
Edit2: Actually no, there is an issue. When artifically introducing larger levels of lag (via lots of objects on screen), the turn rate (right now at 150pi) is much more pronounced at lower frame rates. What is the best way to ensure a proper scaling.
I suppose having a MYCONSTANTvaluespeedtpf in the onAnalong Listeners is not the ideal method, because that is how it stands right now.
Multiply by tpf
Yes, but artificially inducing lag shows it is not that simple. I would imagine that on 30 fps that effect is about 5-10x more pronounced. What was a gentle curve of flight in high fps becomes a very tight hairpin curve at 30fps. I imagine one simply has to figure out the releative difference and make continued adjustments based on the size of tpf might work, but I don’t have time to focus on that aspect.
Are there any solutions anyone else has come up with?