Yeah you're right. I have no idea how I got those extensions mixed up. I meant the heavy framerate lag on the Joypad Object, though I "only" get knocked down to ~50 fps over here, using a Xbox 360 controller. Although if I run with DX8 Display Mode, with V-Sync checked and having that object in the frame, then I can only get 30 fps for some reason. Odd stuff!
My guess on that is that the joypad object is too frequently auto-polling for attachment of new devices, if you do that with joy2 you can also kill your FR.Originally Posted by chrilley
@Julian82 Hm right. Though is that only the "Poll for new devices" property? That it warns not to use on Vista, specifically?
Or the "Poll Joystick every loop"? That one is checked here and I cannot seem to budge the framerate by turning it on and off... Even on my poor surface pro, which is easily bothered by the new stuff I dump into the game, doesn't seem to be affected by the polling options?
Hey everybody. Because I'm afflicted with the Joystick 2 Object-makes-any-app-crash-immediately bug, I've been using the Joypad object instead. Since VACCiNE relies on Joystick 2, I think using Joypad is the way to go, for the time being (optionally alongside XBOX for vibration).
For those affected by framerate drops when using the Joyspad Object, I have a fix. Thanks to volPROFILER @Volnaiskra for helping me identify the problem.
The problem only seems to happen when the Joypad object exists in a frame and does not detect a connected USB controller. If I have the object not create itself at start, there's no lag - likewise, if I have the object destroy itself when it doesn't detect a controller, no lag. So the fix is pretty simple; Don't create on start. Every [interval], create Joypad object (I did two seconds). If Joypad object finds a controller, set a flag to 1. If Joypad does not find a controller, set that flag to 0 and destroy the object.
I didn't get any lag while actively using a controller - this method allows players to hotswap in and out of using a controller whenever they'd like, with (at most) two seconds of lag after plugging in their controller before getting a response. volPROFILER didn't show any meaningful performance impact from this method, either. I do also have all of my controller-specific code checking for that flag first, with some groups enabled/disabled based on the flag's state.
@JimJam if you still care, do you want to verify if this method fixes the performance plunge you saw? It might not be a universal solution, given how temperamental these extensions can be.
Scratch that. It's possible that the same solution actually just works with Ignore/Restore control, but after restarting my computer, Joystick 2 works and Joypad doesn't work. Disregard my advice.
Edit: Nevermind, new theory. Y'know how Joypad only supports two joysticks? Well, now I'm grappling with a bug where every controller I plug in is automatically assigning itself to Joystick 2 (despite Joystick 1 being free), and Joypad is lagging again. So a second ago, when I said the same solution might work with Ignore/Restore control? I think you just have to ignore control on a joystick slot if nothing is plugged into that slot. But I'm probably not gonna keep updating this, because the edit timer will end and I don't want to necro any more than I have.
Just wanted to pop my head up here and say hi, since I was mentioned a couple of times (though I only discovered the thread recently). Very interesting analysis, @JimJam , and rather disconcerting, too. It's a delight to see people take deep dives into stuff like this to really uncover the bare facts for the benefit of the rest of us. I'll probably take your app for a spin and do some measurements at some stage, though for now I'm content enough with my Joystick 2-based VACCiNE solution.
I haven't noticed any input lag myself, and my game is a high-skill speedrunner that relies on tight controls. Playtesters frequently compliment Spryke on its fluid and responsive controls, and that includes a bunch of speedrunning fanatics who took part in a speedrunning contest I ran at PAX a while ago. Many of them were better at my game than I am, and were shaving milliseconds off the levels in places that I didn't even know you could. They enjoyed themselves and seemed to be happy with the responsiveness of the controls. That said, I certainly believe you that there is real input lag, but perhaps my game's saving grace is that it runs at 120fps, which probably counteracts that somewhat. If the input lag is 2 frames at 30fps and just 1-2 frames at 60fps, then I guess it might be just 1 frame at 120fps. In this case, I suppose my 120fps game with 1-frame input lag would end up feeling more or less the same as a 60fps game with no input lag.
I've certainly never been a big fan of the Joystick 2 object, which is why I made VACCiNE in the first place, as I felt I couldn't even use it without first trying to fix some of its craziness. It's been a while since I did VACCiNE now, but if I remember correctly, I managed to fix some of the other issues you mention - eg. by introducing consistent and circular deadspots.
I'd certainly love to get even more responsive controls in Spryke though, so I'll definitely look into this stuff more in-depth closer to my release. But my own attitude at the moment is to take a wait-and-see approach. Yves has mentioned the possibility of making a new joystick object somewhere down the line, and who knows what else will happen on Steam's end in the next months and years. I already did plenty of wrestling with this object years ago to beat it into a semi-usable shape, so I'm not in a hurry to spend any more time on it now. Especially as it seems reasonably likely that new solutions and/or new problems may arise between now and Spryke's release. Though I feel for others like @Julian82 who have launches coming up soon and thus don't have the luxury of waiting and seeing, but desperately need an object that will work properly and play nice with Steam.
answered some where else
Posting Permissions
Reply With Quote