Posts for Ilari


Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
antd wrote:
I used something called lsnes-sdl.exe.
BTW: That version is deprecated, use the wxwidgets one instead. Won't explain why it is slow, as that version has the game-video decoupling too...
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
Anavrin wrote:
It probably has VSync enabled, which synchronize the fps to the refresh rate of your monitor, which is most likely 60 hz.
lsnes can run core and display at different rates. So even if vsync is enabled (there's no explicit control for it), the core can run faster than monitor vrefresh.
Post subject: Re: lsnes seems slow on an i7
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
antd wrote:
Is there any way to make it run faster? Snes9x blazes along (900 fps). The fastest speed I've seen in lsnes is 60 fps. I've tried setting the speed to turbo and infinite with no significant change in speed.
lsnes is based on bsnes core, which is slow. - Is this some self-compiled version or some stock one (selfcompiled could be accuracy core, stock builds are compat core)? - Is this SGB game? Those are SLOW. - Also, does the game use enhancement chips, some of those are slow to emulate? Then there was an issue that caused lsnes to set speed to 60% on startup (but speed select still works).
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
From submission text: "This run aims to reach the fastest but not no getting hurts." Did you mean no death (I saw plenty of damage while watching the run on fast forward while dumping)? Would abusing death even save time?
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
GhostSonic wrote:
The only thing I had to sync it was program the bot to skip 1 frame ahead, which I guess All-Stars wasn't so picky about.
What does that mean exactly? Added an extra poll at beginning? Removed a poll from beginning? Something else?
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
GhostSonic wrote:
I've successfully verified KFCMARIO's Lost Levels run!
How much did you edit it (the polls sequence)?
Post subject: Re: VBA Rerecording Sound Synchronization for Gameboy
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
bortreb wrote:
My initial attempts were met with failure because the audio would gradually go out of sync with the video. Initially I thought that this was because the simulated gameboy did not use exactly 60 fps as its framerate, so I tried various other plausible framerates. Eventually I found to my surprise that there is no constant framerate which could work.
IIRC, The problem is caused by Gameboy allowing to turn the screen off. VBA then changes the framerate in unpredictable ways. Then there was the issue that one run caused the emulator to hang when dumping. My guess of cause is that the game first turns the screen off and then proceeds to lock up, never turning the screen back on.
bortreb wrote:
How are Gameboy TAS encodes rendered?
From internal VBA dumper. Interesting about the pops in sound. I guess VBA isn't just trying to duplicate frames to maintain A/V sync...
bortreb wrote:
This would of course do automatic frame dropping and duplication to make everything stay in sync.
Better to dump at correct framerate, duplicating if needed (dropping should not be needed).
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
jlun2 wrote:
Btw, haven't heard of any complaints about desyncs and the disk swap feature, so I'm kinda surprised that dolphin is more stable than another emulator which emulates an even older console.
Try anything using Wiimote. Plenty of desyncs, even in versions that try to fix the problems (the versions that don't... Forget about it). But stuff using GC controllers (including Wii games) seems pretty stable (if slow).
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
GhostSonic wrote:
I've considered this, but I'm not sure how to tell when the GBA is polling for input.
GBA has interrupts for input (but I think it can poll too).
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
CoolKirby wrote:
This movie needs at least one tag, for Genre, which should be Platform, right? Currently, there are none, which is never supposed to happen.
Oops, added Genre: Platform.
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
GhostSonic wrote:
Looks like it'll sync fine if you add in a few more frames at the game select screen.
Does the same way of modifying it work both on emulator and on the real thing?
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
Oh, some further details: - I can't get sensible listing of how PCSX2-rr injects input because of how the repository is structured (darn SVN). - I looked for the point input is transferred from the pad plugin to core in PCSX2 1.0.0. The data format looks quite nasty, so I presume rerecording would need to be added as special pad plugin or something.
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
Warepire wrote:
The reason for PS2 games not being allowed on the site yet is that PCSX2 is far from ready for TASing, the build used for TASes by the Japanese is quite bad, I guess they just have inhuman patience.
Furthermore, that version is based on quite old PCSX2, and porting it to new one isn't easy.
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
jlun2 wrote:
You mean the VI's in the image? Is it because the frame dump only dumps frames where the VI increases?
The stock Dolphin does exactly that.
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
GhostSonic wrote:
Is it possible to have the lost levels run sync on the normal Mario All-Stars rom? And would it still be a valid verification if done on the normal version of All-Stars rom?
Probably won't sync, since that run uses the version with SMW. But if it synced, it would be valid.
Post subject: Re: JPC-RR setup instructions out of date
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
BrotherMojo wrote:
In particular, when I try to use the FreeDos floppy .img file, it defaults to 1 Side, 63 Sectors, and "Unsatisfiable" Tracks.
2 sides, 18 sectors (giving 80 tracks) most likely. Can't properly test, because whatever version of Java install I have seems royally screwed. The .img file should be 1,474,560 bytes in size.
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
GhostSonic wrote:
I might try to take on the SNES as well by seeing if I can get an lsnes movie to playback on a real SNES. Has anyone actually tried that yet?
Note: the on_snoop (and on_snoop2, which just differs in port numbers it passes) Lua callbacks are explicitly meant for dumping data for console verification. There should be an example script earlier in this thread. As for runs on bsnes core: - 2055M (Actraiser 2): Might be a good candidate. - 2047M (Chrono Trigger): Requires resets, not verifiable. - 2073M (Speedy Gonzales): Rare version[*], hair trigger with timings. - 2000M (Speedy Gonzales): Rare version[*], hair trigger with timings. - 2296M (Lost Levels): Very good candidate to attempt verification on. - 2280M (Congo's Caper): ??? - 2250M (Timecop): ??? - 2277M (Metal Max Returns): ??? - 2337M (Super Mario World 2): ??? - 3517S (Yoshi's Cookie): Needs two controllers. - 3779S (Super Star Wars): Extreme hair trigger with timings. [*] The version that does NOT have the Acclaim logo. I haven't found any gameplay-relevant differences between the two versions (all glitches and tricks I know work in both versions). Still timings are different enough to cause movies to desync fast.
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
There is an archive containing all torrent files on the site. And also RSS feed linking current encodes. And just the set of current encodes is about 194.54GB as of now.
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
jlun2 wrote:
Is it possible/allowed to use a rewriteable cartridge for this? That'll make testing much cheaper.
I don't think it is allowed for final runs, but for testing yes. Note that many rewritable carts have extra stuff at startup that screws timings.
Post subject: Re: Lynx corrupted saves?
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
dwangoAC wrote:
I have put some considerable time into trying to get Lynx to behave but my results so far have been a complete failure. Can someone, anyone, please look at these movie files and tell me what the heck is going on? If someone could please take a look, I'd appreciate it. I'm running Mednafen Rerecording 1.1 Frontend 1.0 on Windows 7. Thanks for any thoughts you have
IIRC, newer mednafen-rr versions have totally broken movie recording for anything except PCE(CD), since MC2 won't support anything else. IIRC, the newest version that works is SVN r144 or something like that (no idea about how version numbers correspond to releases). Those versions use MCM, which can record Lynx.
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
SoulCal wrote:
My build does not put sound and video together. As for being picky about adding/dropping frames, it just seems like a lazy way of fixing the issue to me. It would make sense to me for each frame to be accounted for equally.
It is lazy way, but it is very easy for quick'n'dirty encodes. That's where the second mode comes to, where video is made to sync with audio by giving timestamp for each frame (variable frame rate).
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
SoulCal wrote:
RachelB gave me the build I am using, but I'm fairly certain it does not have the AVhack. My next question is what would be the advantage of the AVhack if it is adding and deleting frames all willy nilly? Wouldn't that cause inconsistency from what the AVI file should be?
Build being able to dump the sound into AVI file is pretty much a dead giveaway of AVhack. As for the reason to duplicate/drop frames: To keep the video synced with the audio.
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
SoulCal wrote:
I checked in VirtualDub how many frames both of my AVIs took for minimizing VIs vs minimizing frames. You were correct Ilari, that minimizing VIs is faster although it results in a higher frame count in Dolphin (but you established that as incorrect).
Are you using stock dolphin or AVhack one (the stock dolphin has really bad A/V desyncs)? - The stock dolphin emits one frame per what is shown as VI (that is, drawn frames). - The AVhack dolphin if dumping internal sound duplicates and drops frames as needed to keep video framerate constant. - The AVhack dolphin if dumping external sound emits timecodes file that shows when each frame is supposed to appear.
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
SoulCal wrote:
For almost all cases, there seems to be 2 frames per VI, at least for non-loading. The frame counter increments during load screens, but VIs stop. Counting the frames seems like the most reliable timing method, since it includes loading time. Logic tells me that lower frame #s = shorter AVI. I will do an encode later today comparing both the less VIs vs less frames. However, it may just be a labeling error from the developers. Notice how on the left, VIs is about 1/2 of the frames, but on the right VPS is double FPS. I just assumed the left was correct, since VIs pause during loading. Ilari saying that the VI counter should go up during loading sounds incorrect to me, but I could be wrong.
Damn hell... I just read the source... There are THREE different things involved. - True VIs. - Screens drawn. - Input frames. - VI is screens drawn. This explains why it doesn't increment while loading. - Frame is input frame (quite sensible). - FPS is number of screens drawn per wallclock second. - VPS is number of true VIs drawn per wallclock second. Well, this is just display stuff... But where it gets serious: The movie code seems to use screens drawn as number of VIs. Which would be VERY WRONG, because it leads to wrong movie times.
Emulator Coder, Experienced Forum User, Published Author, Skilled player (1142)
Joined: 5/1/2010
Posts: 1217
SoulCal wrote:
I would think the obvious answer would be to minimize the frame count, but TASvideos uses the number of VIs to determine the movie length. I submitted a dtm to the new WIP section, and it displays the VIs as the # of frames.
VIs occur at approximately constant rate, so those are used to calculate the time. Frames are at game control, and there can be multiple per VI or none at all. So minimizing VIs minimizes the time. Also, VI counter freezing during loading? One would think that frame counter would freeze and VI counter continue to count up...