MrSunshyne

I also hope this gets seen. No dice on forcing a PS4 or PS5 controller back into dinput mode unfortunately, looks like once they're in enhanced mode the only way to put them back is a reboot. At the very least SDL doesn't support it, and if I'm understanding the docs for the controller correctly, it's impossible, so it looks like the only real fix will be launchbox implementing support.

It's likely not an issue with what application is hooked up to the controller, but mode swapping behavior of the device itself due to the controller drivers in the emulators enabling enhanced controller features. I have more details in a thread here. Also, I know the open feature request has already been mentioned, but if you still want this resolved remember to go upvote it https://bitbucket.org/jasondavidcarr/launchbox/issues/7003/better-switch-pro-controller-support

Don't know if you ever found a fix for this, but it's probably related to your emulators using SDL to toggle the controller into enhanced mode (get full gyro and other features), which is not supported by Big Box. I have more details in a thread here. There is also an open feature request that would likely resolve the issue, so if you do still want this resolved please go upvote it https://bitbucket.org/jasondavidcarr/launchbox/issues/7003/better-switch-pro-controller-support

I know I'm a little late to the party here, don't know if you ever got further on resolving this, but I have a thread going about a similar issue https://github.com/libretro/RetroArch/blob/master/input/drivers_joypad/sdl_joypad.c#L302. To summarize, if you're running the Dualshock controllers via bluetooth it's probably an issue with your emulators switching them from dinput mode to what SDL calls "enhanced" mode (where you get gyro and other advanced features). Once they've been toggled they don't automatically switch back, so they'll keep working with programs that use SLD (or other drivers with full feature support) but will no longer input to programs expecting dinput/xinput. There's an open feature request https://bitbucket.org/jasondavidcarr/launchbox/issues/7003/better-switch-pro-controller-support that would likely resolve your issue, so please go upvote it if you're still having this issue.

Okay, I think I've troubleshot my own issue here, so I'm updating again in case anyone else is encountering it. The Dualshock 4 controller has two modes when connecting via bluetooth. A basic mode, which it uses by default, and is supported by Windows natively (pretty sure this mode is just dinput), and an enhanced mode that includes access to the gyroscope and other advanced features. The popular controller driver SLD has the capacity to access these enhanced mode features, and many programs that use SDL (or other drivers with full Dualshock 4 support) will switch the controller into enhanced mode. I can confirm Retroarch does this (in the code here), and believe, but have not confirmed, the same behavior in Dolphin, PCSX2, and Steam. There is an existing feature request https://bitbucket.org/jasondavidcarr/launchbox/issues/7003/better-switch-pro-controller-support to implement SDL2 as the controller driver which would fix this issue (as well as related issues with Switch Pro, Dualsense, and DS5 controllers). If you've encountered a similar issue to this, please go upvote the feature request. In the meantime I'll see if I can't script something using SDL to run when I exit emulators to grab my PS4 controllers and force them back to dinput mode so big box can use them again. I'll update here if I get it working.

On further investigation this appears to be a Windows issue, rather than Launch Box specific. After the bug arises no input is detected in the "setup usb game controllers" menu in Windows control panel. Other applications are still able to use the controller, but if Windows has lost track of it that suggest to me that there's a deeper issue at work here.

I'm encountering an issue in Windows 11 using a Dualshock 4 controller via bluetooth (have not tested wired). When opening Big Box, the controller works correctly to navigate initially. However, if the controller has been used by another application Big Box no longer appears to receive any input from it. I can no longer navigate with it, and cannot map the inputs. Disconnecting and reconnecting the controller resolves the issue, but only until the controller is used in another application. The order does not matter, i.e. if I connect a controller, launch a game in Dolphin, close Dolphin, and open Big Box, the controller does not work; if I connect a controller, open Big Box (controller works), launch a Steam game via Big Box, then close it and return to Big Box, the controller does not work. This issue does not impact any other applications, I can switch between RetroArch, PCSX2, Dolphin, and Steam and they all recognize and use the controller successfully. I haven't noticed any changes in the device description in the device list. Also, to address some basic things I've already tried: Using a different gamepad: I had access to an old 3rd party USB Xbox360 controller, and it did not experience this issue. However, it's not practical for me to by a full set of new gamepads and configure them for all my other applications (assuming they'd even work) Using another of the same gamepad: This has the exact same issues. Including the same behavior of working until another application uses it. So if the first gamepad stops working, I can connect a second one and use it to launch one game, at which point it will also stop working in Big Box. Toggling controller related settings inside Big Box: I've toggled all three settings (use controllers, use all controllers, problematic controllers) to all of their various options, both before and after the issue arises, with and without restarting Big Box, as well as having two controllers in the non-working state and switching between them as the selected controller (all controllers: off). None of this helps DS4Windows: unfortunately not a good option for me since it breaks controller input for all of my RetroArch cores. RetroArch has an open issue on GitHub about it (#14063), but it doesn't look like it's moving much, although maybe I'll take a swing at it. After writing all of this, I worry that this should have been a bug ticket rather than a troubleshooting request, but if anyone has ideas I'm happy to give them a shot.