UI: Add analog speed limit mapping #15645

unknownbrackets · 2022-07-04T20:21:00Z

This makes it possible to map an analog (most especially, a trigger) to an alternate speed. The FPS is then limited based on how much the analog is pressed, linearly. It can slow down or speed up, depending on what the limit is configured to, but it can't do both.

One trick is that on my DirectInput controllers, the trigger ends up mapped from negative to positive whereas XInput just has a linear positive mapping. I'm a bit concerned there might be other weird results for other controllers, so I added better auto-detection of pad identifiers. For now, kept it to axis ID which might just work.

The auto detection could probably allow better auto-configuration logic for buttons too.

-[Unknown]

hrydgard · 2022-07-04T21:30:37Z

Quite a bit of logic, I'll review it tomorrow :) This is a really slick feature though.

iota97 · 2022-07-05T02:03:55Z

Core/Config.cpp

@@ -862,6 +862,8 @@ static ConfigSetting graphicsSettings[] = {
 	ReportedConfigSetting("AutoFrameSkip", &g_Config.bAutoFrameSkip, false, true, true),
 	ConfigSetting("FrameRate", &g_Config.iFpsLimit1, 0, true, true),
 	ConfigSetting("FrameRate2", &g_Config.iFpsLimit2, -1, true, true),
+	ConfigSetting("AnalogFrameRate", &g_Config.iAnalogFpsLimit, 240, true, true),


Should this belong to controls section? Related to #15406.

This isn't the button or anything, it's the limit. It's kinda more related to graphics.

-[Unknown]

I'm just wondering if we want to allow different control profile to have different limit or not. Not really interested in the semantic meaning here :)

Not a big deal anyway

iota97 · 2022-07-05T02:08:37Z

UI/ControlMappingScreen.cpp

@@ -337,6 +337,9 @@ bool KeyMappingNewKeyDialog::key(const KeyInput &key) {
 		if (key.keyCode == NKCODE_EXT_MOUSEBUTTON_1) {
 			return true;
 		}
+		// Only map analog values to this mapping.
+		if (pspBtn_ == VIRTKEY_SPEED_ANALOG)


Should we provide a message when user try to bind a key?

Well, it just doesn't react.

-[Unknown]

Mhhhh, it should react to at least "Back" button. If a controller have no analog axis this will brick the emulation basically.

Ah, I thought it did allow a way to cancel without touch already. Well, I let cancel go through. Thanks for pointing out.

-[Unknown]

iota97 · 2022-07-05T02:10:04Z

UI/EmuScreen.cpp

 			PSP_CoreParameter().fpsLimit = FPSLimit::CUSTOM2;
 			osm.Show(sc->T("SpeedCustom2", "Speed: alternate 2"), 1.0);
-		} else if (PSP_CoreParameter().fpsLimit != FPSLimit::NORMAL) {
+		} else if (PSP_CoreParameter().fpsLimit == FPSLimit::CUSTOM1 && PSP_CoreParameter().fpsLimit == FPSLimit::CUSTOM2) {


Need a || here

Oops, must've copy pasted something wrong there. Thanks.

-[Unknown]

Rather than just that it's a pad. This tries to get the identifier if possible.

Only shows up as a setting if mapped, to avoid cluttering settings.

iota97 · 2022-07-05T02:49:57Z

UI/GameSettingsScreen.cpp

@@ -335,6 +336,11 @@ void GameSettingsScreen::CreateViews() {
 	altSpeed2->SetZeroLabel(gr->T("Unlimited"));
 	altSpeed2->SetNegativeDisable(gr->T("Disabled"));

+	if (KeyMap::AxisFromPspButton(VIRTKEY_SPEED_ANALOG, nullptr, nullptr, nullptr)) {


This doesn't update when you get back from control mapping. If you exit to main menu and go back to setting screen it work tho'. Probably missing some RecreateView().

Yep, you're right. Made it recreate. I had tested with it already mapped by the time I changed this...

-[Unknown]

hrydgard · 2022-07-05T19:44:49Z

Code looks good, the MAPPED_TO_OPPOSITE stuff is a bit unintuitively named I feel, but okay I guess.

Will test it shortly before merging.

hrydgard · 2022-07-05T20:24:43Z

Hm, on my 360 controller, it's almost as if it's missing some events - by releasing the trigger quickly I could often get it in a state where it's stuck in a speed considerably faster than 100%. Weird...

iota97 · 2022-07-05T20:39:30Z

Core/ControlMapper.cpp

+	if (mode == AnalogFpsMode::MAPPED_DIRECTION) {
+		value = fabsf(value);
+		if (opposite)
+			return;


About the XBOX controller getting stuck, could you try set the speed back to normal rather than just returning here?

I got a really vague memory about some problem with controller going [-1, 0] because 0 is somewhat positive.

This does indeed help:

value = fabsf(value); if (opposite) { value = 0.0f; }

instead of returning.

Though, I would argue that the code that tries to figure out an "opposite" key for a throttle already is wrong, but that's a separate bug. The value that comes in in that case is indeed 0.0, and returning there means that we miss the return to 0.0.

Yeah basically if it's a zero value, the mapping finds itself as an opposite with value 0.0, and skips it with the existing code :)

Hm, I guess that's fine, I suppose any case where you don't want the negative (like mapping right analog X to speed) it'd make sense for it to clamp.

So direction is -1 for 0? I was expecting both to get called, because of this:

if (axis.value > 0) { processAxis(axis, 1); return true; } else if (axis.value < 0) { processAxis(axis, -1); return true; } else if (axis.value == 0) { // Both directions! Prevents sticking for digital input devices that are axises (like HAT) processAxis(axis, 1); processAxis(axis, -1); return true; }

So sure, one would get skipped, but the other wouldn't. I guess it's probably not exactly equaling 0 or something?

-[Unknown]

Code looks good, the MAPPED_TO_OPPOSITE stuff is a bit unintuitively named I feel, but okay I guess.

Well, changed to MAPPED_DIR_TO_OPPOSITE_DIR. I guess it could be THROUGH or UNTIL or something, but already kinda long. I guess it could be MAPPED_FULL_RANGE or MAPPED_EDGE_TO_EDGE or something, none of them are great.

-[Unknown]

Both do get called, so hm, I retract my statement about a bug, and now I'm slightly confused how this happens. The change does help though! I'm gonna debug some more because this doesn't quite make sense...

hrydgard · 2022-07-05T23:45:11Z

Argh, I retract my whole report about clamping. It seems my controller just doesn't always send the 0.0 message, and I was imagining the cause, maybe I moved it differently with the fix, or something. Effectively, we need a deadzone here. I propose from the measurements that we rescale values with a ramp so that 0.15 -> 0.0, and 1.0 -> 1.0.

I can add that post-merge if you don't feel like it after me complaining about something that wasn't your fault at all :) Sorry!

hrydgard · 2022-07-06T07:07:54Z

Thanks, works great now. It's a really cool feature!

hrydgard added this to the v1.13.0 milestone Jul 4, 2022

iota97 reviewed Jul 5, 2022

View reviewed changes

unknownbrackets added 3 commits July 4, 2022 19:14

Core: Track names of connected pad devices.

536c050

Rather than just that it's a pad. This tries to get the identifier if possible.

UI: Add analog speed limit mapping.

c15c4a6

Only shows up as a setting if mapped, to avoid cluttering settings.

UWP: Skip pad name lookup.

6e1457e

unknownbrackets force-pushed the analog-speed branch from cd67aca to 6e1457e Compare July 5, 2022 02:14

iota97 reviewed Jul 5, 2022

View reviewed changes

unknownbrackets added 2 commits July 4, 2022 20:20

UI: Allow cancel from analog speed mapping.

3cb4207

UI: Reset views when analog speed mapped.

f18c1a6

hrydgard added the User Interface PPSSPP's own user interface / UX label Jul 5, 2022

iota97 reviewed Jul 5, 2022

View reviewed changes

UI: Clamp analog speed to 0 on opposite value.

e159cb2

UI: Add small deadzone for analog speed.

0179da6

hrydgard merged commit e1ce0e2 into hrydgard:master Jul 6, 2022

unknownbrackets deleted the analog-speed branch July 6, 2022 13:50

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

UI: Add analog speed limit mapping #15645

UI: Add analog speed limit mapping #15645

unknownbrackets commented Jul 4, 2022 •

edited

Loading

hrydgard commented Jul 4, 2022

iota97 Jul 5, 2022

unknownbrackets Jul 5, 2022

iota97 Jul 5, 2022

iota97 Jul 5, 2022

unknownbrackets Jul 5, 2022

iota97 Jul 5, 2022

unknownbrackets Jul 5, 2022

iota97 Jul 5, 2022

unknownbrackets Jul 5, 2022

iota97 Jul 5, 2022

unknownbrackets Jul 5, 2022

hrydgard commented Jul 5, 2022

hrydgard commented Jul 5, 2022 •

edited

Loading

iota97 Jul 5, 2022

hrydgard Jul 5, 2022

hrydgard Jul 5, 2022 •

edited

Loading

unknownbrackets Jul 5, 2022

unknownbrackets Jul 5, 2022

hrydgard Jul 5, 2022

hrydgard commented Jul 5, 2022 •

edited

Loading

hrydgard commented Jul 6, 2022

UI: Add analog speed limit mapping #15645

UI: Add analog speed limit mapping #15645

Conversation

unknownbrackets commented Jul 4, 2022 • edited Loading

hrydgard commented Jul 4, 2022

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

hrydgard commented Jul 5, 2022

hrydgard commented Jul 5, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

hrydgard Jul 5, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

hrydgard commented Jul 5, 2022 • edited Loading

hrydgard commented Jul 6, 2022

unknownbrackets commented Jul 4, 2022 •

edited

Loading

hrydgard commented Jul 5, 2022 •

edited

Loading

hrydgard Jul 5, 2022 •

edited

Loading

hrydgard commented Jul 5, 2022 •

edited

Loading