it has 1:1 the very same lag as the controller
on ps3 you have 100hz refresh rate on the controller, that is 10 ms
which means the xim on ps3 has 10 ms lag
on xbox360 you have 125hz refresh rate on the controller, which is 8ms
once again, that means your xim runs on 8ms lag
its always the very same as the controller, its not like the xim adds more lag to it
Well, I can't believe that is true unless the Xim Edge internal processing time is like really low (like less than a tenth of a milliseconds).
Let me explain this :
Time (in ms) ===> 0..........1..........2..........3..........4..........5..........6..........7..........8..........9..........10..........
ACTIONS :
Case 1 :
DS (no xim) ===> ..... ..... .... F (full forward with the view stick at T+A, it will "look above")
PS3 POLL ===> X (nothing to read) ..... ..... .... ..... ..... ..... ..... ..... ..... ..... ..... ..... ..... ..... ..... F (received at the next polling)
Xim Edge ===> ..... ..... ..... (unspecified internal processing time P) ..... F (translated from mouse)
PS3 POLL ===> X (nothing to read) ..... ..... .... ..... ..... ..... ..... ..... ..... ..... ..... ..... ..... ..... ..... F (received at the next polling)
So here, as long as the P + A is below the polling rate interval I (10 ms on PS3), it won't change anything.
But obviously, as soon as we got this situation :
Case 2 :
DS (no xim) ===> ..... ..... .... .... .... .... .... .... .... .... .... .... .... .... .... .... .... .... .... .... .... .... F .... ....
PS3 POLL ===> X (nothing to read) ..... ..... .... ..... ..... ..... ..... ..... ..... ..... ..... ..... ..... ..... ..... F (received at the next polling)
Xim Edge ===> ..... ..... .... .... .... .... .... .... .... .... .... .... .... .... .... .... .... .... .... .... .... .... (..... ..... .....) F
PS3 POLL ===> X (nothing to read) ..... ..... .... ..... ..... ..... ..... ..... ..... ..... ..... ..... ..... ..... ..... X (nothing to read) .... .... (will receive the "move" action 10 ms later, next polling)
So basically, if those internal processing time are negligible (like way below 1ms), the Xim Edge processing time can be considered as "null" and thus you can tell "this is the same reactivity as the naked controller's one" BUT, as soon as the Xim Edge internal processing time becomes higher than 1ms, it will increase the average processing time by 1 full polling rate interval at a frequency depending on that P processing time.
Well, I'm not really good at math but with P = 0, we have a "latency range" (processing time interval) from 0 to I (0ms to 10ms on PS3, more likely I, so let's use this value, the "0ms" case must be really rare)
With a P = 5 ms (for example)
our range will be 0 to I + P so more likely 1xI to 2xI (10 to 20ms) with a half chance (50% odd) of it to happen, so 15ms average.
With a P = 2 ms
our range will be 0 to I+P so 1xI to 2xI (10 to 20 ms) with a 2ms / 10 ms (20% odd) of it to happen, so 12 ms average...
I'm not quite sure about the formula but that must be something like this... anyway there IS a correlation between the Xim Edge internal processing time and the impact on the next polling state (previous move already available to be transmitted or not)
So, I would be glad to have more information about this.
Thanks !
