Commit Graph

625 Commits

Author SHA1 Message Date
adelikat 698c6741ae Refactor IEmulator and remove the IController property, instead pass it as a paramter to the FrameAdvance() method 2017-05-01 20:09:11 -05:00
adelikat ded1c2d7b7 Break off BoardName from IEmulator into a separate IBoardInfo service 2017-04-30 21:01:54 -05:00
J.D. Purcell 083d9bec0e Throttle cleanup. 2017-04-30 20:38:10 -04:00
J.D. Purcell 0d853fd74a MainForm cleanups. 2017-04-29 15:18:08 -04:00
J.D. Purcell 3ebcaa8af0 Rewind cleanup / tweaks. 2017-04-29 14:04:19 -04:00
J.D. Purcell a72bdfe30f Allow rewind during movies. 2017-04-29 11:06:00 -04:00
J.D. Purcell 4848bb5cc3 MainForm: Remove unused line. 2017-04-24 10:34:13 -04:00
adelikat 13ca76f3f8 UI for changing SNES controller configurations 2017-04-15 20:34:32 -05:00
J.D. Purcell 16bf63ade8 Random cleanups. 2017-04-10 09:36:04 -04:00
shanehoman 58e1b52369 expose MainForm.SaveRam() method to lua 2017-04-02 23:12:38 -07:00
J.D. Purcell 07a8c2c027 Remove unnecessary rewind capture on ROM load. I reviewed the original commit where it was added, and it was only to initialize the delta rewind buffer. 2017-04-02 23:01:41 -04:00
J.D. Purcell 3f776dbf6f Rewind: Fix subtle off-by-one-frame bugs (only really noticeable when frame advancing). 2017-04-02 18:50:34 -04:00
J.D. Purcell c58d2929f8 SoundOutputProvider: Standalone mode to eliminate the need for an external buffer when used with SyncToAsyncProvider. 2017-03-25 16:26:46 -04:00
J.D. Purcell 231795c2cb Correction to previous AVI recording commit: buffer doesn't need to be so large. 2017-03-23 09:30:22 -04:00
J.D. Purcell 33d42152a8 Fix N64 audio weirdness in AVI recording. 2017-03-22 23:00:37 -04:00
J.D. Purcell 6cc5a914b7 Rewind cleanup. 2017-03-18 21:22:02 -04:00
zeromus edc3cea302 use gizmo to load roms from drag&drop and commandline (file>open already magically worked, due to explorer I guess) 2017-03-15 22:50:35 -05:00
feos 5c91c2f98c tastudio: allow exporting a named state 2017-02-24 21:58:41 +03:00
zeromus f458648fad improve rewind in tastudio by leaving editmode intact but taking care not to clobber frames. take special pains to ensure that (reverse) frame progress vs smooth rewind semantics stay as expected. 2017-02-16 00:29:28 -06:00
adelikat 2f3c0af4ea Open Rom dialog - mark intellivision as not experiemental and have it show up in release builds 2017-02-11 12:45:20 -06:00
zeromus ea2188dc46 make tastudio.setplayback correctly restore the recording flag after it finishes seeking from the greenzone to the target frame 2017-02-07 00:01:14 -06:00
zeromus 136c071ee8 12hz is too fast for default fps updates, i cant even read it. 2017-02-06 23:17:46 -06:00
zeromus a70e6e2fc4 fix #790 (Scroll wheel missing from input.getMouse()) 2017-02-04 20:52:32 -06:00
J.D. Purcell 52d604cb24 Fix turbo mode not disabling sound rendering (broken by 68a362c).
Apply FF&RW volume when unthrottled.
2016-12-20 22:41:59 -05:00
J.D. Purcell 970a9017d4 Fix typo in previous commit. 2016-12-18 13:03:58 -05:00
adelikat 40be0dd0b3 Refactor the way a Null ControllerDefinition is created 2016-12-16 23:34:18 -06:00
J.D. Purcell 7843664f68 Improve FPS smoothing to initialize with a real value instead of ramping up from zero. Re-initialize the FPS data after pausing or changing fast forward/rewind state to eliminate ramp-up and ramp-down in those cases. Only the initial behavior has changed; the FPS calculation/smoothing is otherwise mathematically identical. 2016-12-16 16:01:21 -05:00
J.D. Purcell 84480e4235 Fix sound bugs + cleanup. 2016-12-14 21:19:46 -05:00
J.D. Purcell afd55fdc6b Revert "Default to Sync sound instead of Async"
This reverts commit 6702f7795c.
2016-12-14 18:26:19 -05:00
alyosha-tas 6702f7795c Default to Sync sound instead of Async
It seems like 'CanProvideAsync' was always returning true so the else if at 1621 was defaulting to setting Async, which caused an exception later.

This seems similar to some TAStudio issues where different people are getting different results depending on Windows version, but I don't see any issue with not setting Async mode.
2016-12-14 15:07:04 -05:00
adelikat dd48864058 AsVideoProviderOrDefault() extension method 2016-12-12 13:47:11 -06:00
nattthebear dbdac5e3f2 Fix stuff or not 2016-12-11 22:37:28 -05:00
nattthebear a98316ba97 Soudn refactoring: Unbreak clock & vsync thrtottle 2016-12-11 18:35:42 -05:00
nattthebear ca60a33293 mess some stuff up 2016-12-11 15:01:42 -05:00
nattthebear eb72ff9053 Swag up NullSound 2016-12-11 14:07:12 -05:00
adelikat b73a500d6b Major sound provider refactor. Create a unified interface ISoundProvider, that is an IEmulator service and break it off of IEmulator. IAsyncSoundProvider was kept as a shim for now, for legacy cores that were inherintly async. 2016-12-11 11:15:00 -06:00
adelikat b697110098 Rename ISoundProvider to IAsyncSoundProvider to more accurately describe what the service does 2016-12-09 10:24:43 -06:00
adelikat 4d1629e271 Completely divorce IVideoProvider from IEmulator, use a NullVideo implementation in MainForm if a core does not provide one, Remove VideoProviderGlue. Note that NullEmulator does implement IVideoProvider still, since we like to have fun with that one 2016-12-07 13:21:18 -06:00
adelikat 9135aea773 Make a shim property in Mainform as a step towards reducing the use of Globals 2016-12-04 11:19:34 -06:00
adelikat 78bb41e3cf Mainform - make AV and Screenshot menus dependent on having a VideoProvider, and remove some usages of VideoProviderGlue 2016-12-04 10:50:28 -06:00
adelikat 5665de1797 Intellivision - add a "Fake analog" controller as the default controller for easier binding for casual play, but also leave a boolean option as well for more precise (TAS) control. Add a controller config dialog to select them. 2016-12-03 17:44:25 -06:00
feos 05d47a7318 tastudio: make analog hotkeys working, and their names fit in 2016-11-21 18:00:21 +03:00
feos 95dc2fd6b3 tastudio: move arrow keys logic to hotkeys
disallow sending Minus to MainForm
2016-11-20 23:26:13 +03:00
adelikat be93e78dd5 Intellivision - add .int files to the open file dialog in developer builds 2016-11-13 08:25:19 -06:00
zeromus 595a207c97 add new field to gamedb to specify which core should be used.
allow it to be 'neshawk' or 'quicknes'.

this should only be employed by end-users; we won't make those choices in the gamedb officially.

the quicknes blacklist is still implemented separately.

yeah, that's a lot of prioritization levels.
2016-11-12 19:23:54 -06:00
zeromus c7b639be28 apply ArcadePit idea: Let user change the string in the emulator title bar via a file
(but changed the format of the string to be more generally useful, and changed it to dll/custombuild.txt)
2016-11-12 18:49:03 -06:00
zeromus 0b4fc8b6a5 apply Arcade Pit work: "Extract NesCarts.xml for easier modification" 2016-11-12 18:32:18 -06:00
adelikat 60f84fa028 Merge pull request #706 from Meerkov/master
Fix Volume Controls
2016-11-08 10:42:27 -06:00
adelikat 8033a8c4fc Fix a null emulator check, fixes dump status icon not updating after a failed rom load 2016-11-07 14:26:02 -06:00
adelikat ef6c9131c6 Fix behavior when loading a new rom when a cheat is loaded, fixes #550 2016-10-06 13:57:39 -05:00
Meerkov 6a9510c9ae Re-enable Thread Sleep when not in Absolute Zero mode 2016-09-29 01:19:25 -07:00
Meerkov 129fdf2a68 Revert "Allow Fast Forward Volume to be Louder than Normal Volume"
This reverts commit 4bd6608940.
2016-09-29 01:10:23 -07:00
Meerkov 9eafac83d4 Improve Absolute Zero Mode by 15-30%
Short-circuits out of the Render function in the main program loop.

With an NES game running at 8500 FPS before, this brings it to 9700 FPS.

With no game loaded, running at 45k FPS, this brings it to 60k FPS.
2016-09-29 00:57:32 -07:00
Meerkov 151c88feb8 Removes busy-wait
This sleep doesn't appear to be needed. It's just busy waiting.

We save CPU cycles by not doing this.
2016-09-29 00:18:55 -07:00
Meerkov 1b8d32e82f Refactor the FPS Counter to it's own function
1) Takes the block of code that calculates the string to print and gives
it it's own function.

2) Stores a static display variable. Fixes incorrect FPS while fast
forwarding as a result.

3) Reduces updates per second to 12 from 16.
2016-09-29 00:04:39 -07:00
Meerkov 68a362c8ff Fix subtle bug in renderSound
Bug as follows:

renderSound was true when any of the following:
_currAviWriter!=null&&_currAviWriter.UsesAudio
!IsTurboing || _currAviWriter!=null

So, if _currAviWriter!=null, it didn't matter if it has the UsesAudio
flag on or not.
2016-09-28 23:14:18 -07:00
Meerkov 4bd6608940 Allow Fast Forward Volume to be Louder than Normal Volume
Allows FF Volume to override the normal volume, instead of multiplying
it (which could only decrease the volume).
2016-09-28 22:40:14 -07:00
Meerkov 4aa64b1f34 Refactor atten variable checks
Cleans up volume checks, makes it a bit easier to follow by removing
negations.
2016-09-28 22:37:26 -07:00
Meerkov d3547b15d4 Fix bug with Mute Frame Advance
It will actually mute when you hold the Frame Advance button.

Previously, the setting did nothing.
2016-09-28 22:02:38 -07:00
Meerkov 49d7f79ce5 Reduce redundant attenuation code in main loop
This simplifies the code for setting the attenuation by setting it to 0,
and moving all the checks that change it next to each other.
2016-09-28 21:48:35 -07:00
Meerkov cd46188efe Remove Unused Variable
UpdateFrame can never be true, so all of this code is pointless.
2016-09-28 21:24:09 -07:00
Meerkov 4e99ed17ae Replace spaces with tabs 2016-09-28 21:16:24 -07:00
Meerkov 91c5eda526 Update FPS Counter
Changes the FPS counter to fade.

Why: FPS counter has a bug where it doesn't normalize the frame count to
the time passed since last update. This can result in an error of 1 FPS.
When running at 30 FPS, that is ~3.33%. The lower the FPS, the worse
this error grows.

1) Adds a parameter for how often to update the FPS (e.g. 16 times per
second).

2) Normalizes to the number of seconds passed since last update (fixing
the bug).

3) Adds a smoothing parameter (e.g. weigh the most recent interval at
8x, and normalize with current displayed date). This results in less
jumpy FPS because it takes into account historical data.
2016-09-28 21:04:53 -07:00
adelikat 707c0944f0 clear the OSD before doing loadstate callbacks, fixes #635 2016-09-22 10:13:26 -04:00
adelikat 10a2ae1471 Detect when a joypad is added or removed while bizhawk is open fixes #698 2016-09-21 21:23:57 -04:00
feos ef445c90c7 tastudio: fix a couple of crashes 2016-09-09 19:15:00 +03:00
feos a0d718acc0 tastudio: don't pause on green arrow is seeking past it 2016-09-08 23:29:17 +03:00
feos 2b0dc82d16 tastudio: finally make < > buttons operate identically to advance/rewind hotkeys 2016-08-28 18:42:59 +03:00
feos 0246c952eb tastudio: complete overhaul of seek logic, replicating taseditor
todo: fix lastpositionframe dying if editing the same row twice with no autorestore, probably improve cursor color and rename old seek related vars
2016-08-28 01:08:43 +03:00
feos ef47d9fbca tracer: reboot on game close
more ^final^ tweaks
2016-08-22 21:57:18 +03:00
zeromus 94383ab143 fix deleting recent roms when missing 2016-08-17 20:24:14 -05:00
feos 80c90e7acc m64p: attempt to add execute callback
no memory callback seems to be working yet though
2016-08-11 21:58:09 +03:00
feos f94b352765 tastudio:
- fix escape handling
- frame advance doesn't unpause emu when seek frame is reached
2016-08-04 19:42:42 +03:00
feos bade8f9217 tastudio: properly handle global rewind function (step'n'run)
not attaching it to playback buttons, because they're obsolete and because they kinda simulate holding anyway.
2016-06-13 13:17:21 +03:00
Hathor86 7fb0eaf337 Add Save/Load state methods and events to ApiHawk
Now, ApiHawk can save and load name state. There are also events that
you can attached too. Furthermore, BeforeQuickSave/BeforeQuickLoad
events occurs just after user pressed button. We can react to this and
even override behavior.
Finally, I just allow .bat to take xml files (for ApiHawk doc)

PS: Yeah, I'm still alive
2016-05-29 16:50:16 +02:00
zeromus ef318aa0f8 fix noise in some cores during pause 2016-05-22 15:00:45 -05:00
zeromus 1a1a688b96 remove NeedsToPaint; client now repaints ~100fps while paused all the time. as it already did when input display was enabled. fixes #615.. probably 2016-04-20 12:17:41 -05:00
zeromus 1b8a51e447 rework sound config dialog and sound volume control to permit independent control of volume during FF & RW, and independent toggling of sound synthesis vs volume=0 suppression 2016-03-29 03:40:37 -05:00
zeromus 10272b4857 add alternate vsync method for d3d display method, in case youre the lucky owner of a system that sucks 2016-03-27 04:49:17 -05:00
zeromus 406b431280 add comments for later about triple buffering 2016-03-26 12:39:31 -05:00
zeromus c429ebdbcf fix error in "Display Vsync set to" OSD message 2016-03-26 12:39:31 -05:00
zeromus 204ff2365f I think the 1px padding hack (enabled by default) should have only applied to opengl (it was created when we only had opengl and it's meant to address a problem i think is opengl only). May possibly cause flickering when going fullscreen, if I'm wrong. 2016-03-24 14:47:18 -05:00
Hathor86 c0f7219b06 Contextualization of external tools + Basic of ClientApi
We can now specify a context for external tools. Global, specific for a
type of emulator / system or specific for a game (If not specified, a
tool is global).  It should avoid loading something unwanted.

Also add ClientApi, a static class that can interract with EmuHawk
window (SetPadding is the only thing implemented currently).
2016-03-16 23:06:53 +01:00
zeromus de576bb067 AVWriters now implement UsesAudio and UsesVideo so that the frontend can decide how much work it needs to tell the core to do to fulfill the AVWriter's needs (fixes #593) 2016-03-05 17:20:00 -06:00
adelikat f1f45afd83 Remove CoreComm link cable properties and turn it into an ILInkable service instead 2016-02-28 13:52:17 -05:00
adelikat 7a4a210ccc Remove some unused code and usings 2016-02-28 13:36:17 -05:00
adelikat c333403895 Prevent Advanced loader from pooping all over every other tool, this needs to be refactored 2016-02-28 12:15:23 -05:00
zeromus 3c96e9d434 ok so i messed up with the FileInfo, because it was already there and I just lazily used it. But lets try handling this with the sophistication it calls for 2016-02-27 16:19:41 -06:00
adelikat fa3ccc59e7 try/catch the File.Move when making a backup savestate and do nothing 2016-02-27 09:58:41 -05:00
nattthebear 35f2599c13 Remove savestate wait that caused nothing but trouble 2016-02-27 09:51:01 -05:00
zeromus 95bc69b448 some prepwork for opengl texture ID importing (skip rendertarget resolve on OGL display method) but there are still problems and it can't be enabled yet 2016-02-22 00:23:20 -06:00
zeromus 10b8a54e6c savestate backup had delete (it's asynchronous) vs move race condition 2016-02-21 22:01:36 -06:00
zeromus 0eb6e83384 glcontexts can specify versions and clean themselves up 2016-02-21 17:19:34 -06:00
adelikat c41851e877 Call the lua loadstate callback before redrawing the screen and other updates 2016-02-15 19:17:19 -05:00
adelikat b3728ff192 Make Screenshot Raw to Clipboard and Screenshot Client to Clipboard into configurable hotkeys instead of hardcoded shortcut keys in the menu items 2016-02-14 21:30:24 -05:00
adelikat 8168031bed Make TargetZoomFactor a per systemId setting 2016-01-31 21:05:08 -05:00
adelikat 894a9c2318 Rework ToolHelpers from a static object to ToolFormBase that inherits form and have a lot of tools inherit it, in hopes of moving a lot of copy pasta into a base class, move a few functions that show up a lot there 2016-01-30 20:24:53 -05:00
feos 099df32224 ability to save/load specific config files 2016-01-27 19:07:26 +03:00
adelikat 60692450e8 New Hex Editor - add the plumbing for the HexView widget, that will use the GDIRenderer as the basis for the drawing. 2015-12-24 11:15:29 -05:00
zeromus 22143f99fb work on stability of avi writing. Unfortunately, I did a half dozen things here, overlapping in complicated ways, and at least 2 of which were required to fix it: changed pinvoke wrapper of AVISaveOptions, removed ownership of pointers from codec token (simplified analysis and in principle [but not practice] determinism due to less GC involvement), changed life cycle management of codec token (irrelevant due to prior), changed functions used to allocate memory from AllocHGlobal to pinvoke HeapAlloc, made different assumptions of ownership/lifecycle of pointers accessed by AVISaveOptions. notably the pinvoke wrapper for AVISaveOptions was the last thing I fixed that sealed the deal, but I don't think it was that simple. 2015-12-19 05:13:31 -06:00