Don't forget about the - new for VP10 - ambient occlusion in the general and table specific settings (plus adjustment for the effect level under lighting options). It creates some decent shadowing of objects and the flippers and posts can look nicely enhanced with it. It is somewhat demanding on performance and I believe at this moment it only works with nVidia video cards. However, it can work at the same time as FXAA if you enable FXAA through the nVidia control panel instead of VP10s options. With the other
VP AA option, that looks even nicer than FXAA, AO can work as well but might push performance past a steady / vsyncable rate of 60 FPS.
On my development system (Windows 7 32-bit 4 GB RAM, with 5 year old Intel Core i5 760 and nVidia GTX 560 Ti OC) I get around 112 FPS for Monster Bash WIP15 in DT / Desktop mode at 1080p with control panel FXAA enabled, AO in
VP enabled and even using "Max prerendered frames" at 1 in
VP preferences, which is more demanding than at 2 but better for flipper interaction / timing. With the same setup but
VP's 4xAntiAliasing instead of nVidia control panel FXAA enabled (now off), I get around 53 FPS. In
FS / Full Screen mode with all the same configurations, I get around 84 for DT and 32 for
FS (reasonably shy of being able to vsync at a steady 60). With my mini-cabinet, which is more current and has an nVidia GTX 760 OC, I get considerably better but run at 1440p in an HV mode so comparisons are not as straight forward.
Setting vsync on should be the standard now for VP10 with the physics decoupled from the FPS since the PhysMOD extension of
VP. As long as the computer configuration is capable of getting at least slightly over 60 FPS consistently, then vsync should smooth everything out nicely and be set for every table. One important note though, is that with a locked in 60 FPS to a 60 Hz monitor you will get only one frame every 16.7 ms so it is important to turn down the
VP preference of "Max prerendered frames" to 1 instead of 2. It takes a little bit more performance but as long as vsync is used, you only need enough FPS to reach and stay at 60 (or 120 / 144 if you use some higher end computer monitors - only a handful of T.V.s can actually truly accept a 120hz signal at this point without frame skipping). If "Max prerendered frames" is left at 2 then you have 2 full frames / 33.3 ms before what has happened in the
VP program even begin to makes it to your eyes, plus have to add your display's input lag (which can be pretty high on T.V.s compared to computer monitors - the latter being used more for DT / mini-cabs) and then still have to add your controller's IO input lag on top of that (what you are reacting to will be older / already occurred and then your reaction / input has further delay depending on your keyboard / mouse / joystick handling). Flipper accuracy can be severely affected by all these aspects of lag / delay plus the more subtle nuances of flipper tricks can become completely unavailable more so from the controller input lag side but also from reacting to "old" video information as to the current state / location of the ball can and does have an affect too.
The link below is from a post where I wrote about USB input lag / polling issues in Windows and some measures to improve the lag and as a simultaneous result the number of sample / points in time where inputs can occur. The difference of 125 vs.1000 Hz for standard USB devices / drivers is time slices of 1 ms vs. 8 ms for a 125 Hz (creating a smoother curve and more detailed points in time in which a button press / flipper switch command will register and as a result yield more varied leading to more realistic results).
Edited by jimmyfingers, 21 February 2015 - 08:38 PM.
Top
Contributor
You deleted your unused resources? I thought that was an option for unused table resources, sounds - images . /cry
are all trademarks of VPFORUMS.