1 2 3 4
20 21
Emulator Coder, Skilled player (1112)
Joined: 5/1/2010
Posts: 1217
SoulCal wrote:
Well if the Vertical Interrupt occurs every drawn frame, and the game is at 30fps, could I essentially use this as the frame counter?
Yes, it can be used as a frame counter.
SoulCal wrote:
And why would the "frame" counter be off so much? Think it is just an error by the developers?
I don't know, possibly it assumes that there is 1 poll per frame and it actually measures controller polls, not frames (which will be messed up if game polls multiple times per frame).
Post subject: NEW Resident Evil 4 TAS WIP
Active player (335)
Joined: 1/19/2010
Posts: 383
Location: Texas
This is an improvement of about 16 frames up until the crow, and I also equip a grenade to run faster. This means I can't use the handgun to get lucky drops from the two guys at the bridge. Derek (the SDA RE4 record holder) told me this is for the best. Here is by dtm file if anyone wants to try and sync it themselves. http://www.2shared.com/file/kGfX7hN-/tillvillage.html See if you can record in 1080p. This video is in 1080p, but the only way I could get Dolphin to do this was to force widescreen mode, which just stretches the image. I'm not sure if RE4 even supports widescreen (since it puts the black bars at top and bottom anyway). How was the MegaMan 10 TAS recorded? RE4 for some reason dumps frames at about double the framerate, and I just use a video editor to correct it. Link to video
Experienced player (583)
Joined: 1/27/2011
Posts: 427
Location: Oregon
Ilari is correct about the poll counter. VI is what you should be paying attention to. What has been changed to "Frame" is now the current number of polled inputs. Since GC and WII poll multiple times per frame, the number advances quite bit with each frame step. The VI can be used for all intents and purposes as your frame counter.
Active player (335)
Joined: 1/19/2010
Posts: 383
Location: Texas
This concept about polled inputs has me a little flustered. For every VI advanced, the poll advances 2. That means that I should be allowed 2 inputs per frame, correct? If (ideally) I would be able to have 2 inputs per frame instead of 1, this would make TASing this game much tighter, or does this even matter?
Joined: 12/6/2008
Posts: 1193
damn dirty double posts...
Joined: 12/6/2008
Posts: 1193
SoulCal wrote:
See if you can record in 1080p. This video is in 1080p, but the only way I could get Dolphin to do this was to force widescreen mode, which just stretches the image. I'm not sure if RE4 even supports widescreen (since it puts the black bars at top and bottom anyway). How was the MegaMan 10 TAS recorded? RE4 for some reason dumps frames at about double the framerate, and I just use a video editor to correct it.
List: http://en.wikipedia.org/wiki/List_of_Nintendo_GameCube_games_with_480p_and_16:9_support As you can see RE4 doesn't support 16:9 you have to be contend with 480p. Forcing it through Dolphin is kinda iffy, imo, but oh well... Don't even know if it can be done differently on Dolphin. Let the encoders worry about getting it into 1080p in the end. Aktan even got the Majoras Mask TAS to that resolution and that is an N64 game.
SoulCal wrote:
This concept about polled inputs has me a little flustered. For every VI advanced, the poll advances 2. That means that I should be allowed 2 inputs per frame, correct? If (ideally) I would be able to have 2 inputs per frame instead of 1, this would make TASing this game much tighter, or does this even matter?
The gamecube polls the controller multiple times a frame. But if the game actually uses that input is an entirely different thing. You certainly can't move the character if the game doesn't progress. It might help with things like death hovering in TWW, however.
Active player (301)
Joined: 9/2/2006
Posts: 504
Aktan even got the Majoras Mask TAS to that resolution and that is an N64 game.
Mupen64 handles models and stuff well at higher resolutions. So it can be dumped in 1080p, no special encoding needed to make it look good(at least i think so).
Post subject: Resolutions
Active player (335)
Joined: 1/19/2010
Posts: 383
Location: Texas
Now that I think about it, what they probably did was encode it at 1440x1080 (the non-widescreen 4:3 resolution) and then added the black bars on the sides to fill it to 1080p (16:9 widescreen). I can encode in 1440x1080 with Dolphin but when I upload to youtube it puts it at 360p, just cause the resolution isn't exactly 1920x1080, although the resolution is far above 360p. EDIT: Also, I remember trying to TAS parts of Majora's Mask. If you run Mupen at certain resolutions, the .m64 files will not sync. So it should be encouraged by TASers to create their playback file with a resolution the original console actually supports (I'm sure this is common knowledge, but I'm new). DOUBLE EDIT: New WIP. I really hope someone would give me feedback on how this looks, and what improvements I could make. The only thing I could think of is somehow get an extra grenade (as if using 3 in 1-1 isn't enough) and blow up the trip wire AND the lock off the house at the end. This would save a weapon switch pause, I could run straight to the door without going around the tree, and I will keep the running speed of grenades for most of ch1. Whether a grenade has a big enough blast to blow up both is unknown to me. This is also about 38 seconds faster than the unassisted record, although about 8 seconds of that is because I don't collect the Pendant. Link to video
Active player (458)
Joined: 11/12/2010
Posts: 183
Location: Sweden
This looks nice, but is it possible to get something from the second crow next to the one you shoot? And would you be able to gain anything from throwing the grenade so you can loot the first one/two corpses you kill? And then arc the second grenade so you won't have to run as much? Also possible 3300/9900 drops when you gather money? Is the path shorter if you go through first house to the left at the last area instead of around it? Sorry to bombard you with so many questions, but those were the things I came to think of first. Apart from those things, it appears well performed. Good job!
"An artist who can’t take constructive critique on their work is only hurting themselves and their potential. Conversely, and artist that can’t communicate a critique in a constructive way isn’t helping anybody."
Active player (335)
Joined: 1/19/2010
Posts: 383
Location: Texas
It is possible to get money from the 2nd crow, but most crows only drop a few hundred ptas aside from 3 or 4; the 2nd crow at the beginning (before the 1st Ganado) and one in the castle drop a spinel. There is one on disc 2 that drops 5000 Gold. And of course the one with the grenade. There is also one that drops a flash grenade in the 1st gigante area. I am currently redoing the village area cause I can do it a few seconds faster. This requires one of the first two enemies in the village drops a grenade. It is hard to manipulate enemies to drop those. Once I have an extra grenade, you are right; I could arc it to the left side of the fire pit and kill the back group. With the 3rd grenade I can chunk it over the roof to the last set of enemies. I have seen several 3,300 drops since I began TASing this game, but only 1 9,900 drop. 19,800 drops are also apparently possible, but I have never seen one in the 6 years of playing this game. The only problem with the above village strategy is it leaves little chance to collect enemy pickups (aka, no money). I need at least 30,000 for a Rocket Launcher before the merchant trip on the way to the church. I can sell the first aid spray for 5,000, I want to collect at least 5,000 from enemies, and then I can collect the brass watch(end of ch1-2) and the headdress (by the merchant). Ideally, it would be best to get two 18,900 drops in the village and a grenade from the first enemy, but I don't see that happening unless the RNG is figured out. Having only 9,900 or 18,900 ptas drops basically allows me to have a grenade equipped for all of ch 1, since I won't have to equip pistol to collect the brass watch and headdress(this will save at most 15 seconds). I have figured out that grenade running is ~4.6% faster than with handgun, so the fewer weapon switches the better. This also assumes I get an extra grenade in the village to blow up the trip wire and bookshelf in the last house, so I won't have to equip handgun to blow up the wire (the lock can be cut with knife). And Tseralith, no, the left house is not faster; even if it was less distance there is a trip wire there. EDIT: I'm considering giving this TAS project up. The differences between how console and Dolphin run are too different (obvious point). Each load screen is about 1.5 seconds faster on emulator than console, in addition to inventory unpausing being another 1/2 second or so faster. I said I was about 30 seconds ahead of the SDA record for Ch1-1 but minus the pause and load times I'd say that is reduced to about 23-25 seconds. The cutscenes at the beginning of ch1-2 start about 2-3 seconds faster than on console. Shaving off 10 seconds per sub chapter by load times alone would result in a final game time 3 minutes less than the SDA record due to hardware differences.
Active player (458)
Joined: 11/12/2010
Posts: 183
Location: Sweden
Giving it up completely or giving it up for now? I don't know if Dolphin will "advance" to the stage where loading times will be the same as on the console, and thus prevent any possible TAS'ing for the goal to get a "correct" fastest time. So we might have to live with Dolphin being faster and let the viewers know it is so. I would still gladly like to see a TAS of the whole game, even if it's several minutes faster only because of shortened loading times.
"An artist who can’t take constructive critique on their work is only hurting themselves and their potential. Conversely, and artist that can’t communicate a critique in a constructive way isn’t helping anybody."
Active player (335)
Joined: 1/19/2010
Posts: 383
Location: Texas
I will try and do an official test with Dolphin to see if load times are incorporated into the game timer. If the game timer is independent of the load screens, I may continue. With a game like RE4 though, I bet more people would be interested in the final game timer more than a frame count. If memory serves me correct though, this is why no speedruns are submitted for the PS2 version; load times are too long (and also the lack of a few glitches). Load times in older Dolphin revisions caused desyncs because older revisions would load faster(yes, faster!) by about 15 frames for the first cutscene trigger. So I may be able to sync my run onto older revisions, but I bet with the frame numbers being off, the enemies will have different item drops.
Active player (458)
Joined: 11/12/2010
Posts: 183
Location: Sweden
SoulCal wrote:
With a game like RE4 though, I bet more people would be interested in the final game timer more than a frame count.
True, many in SDA (and many others too of course) are surely more interested in that than how to make a "perfect turn". Maybe special bugs or how well you can manipulate drops perhaps. But I understand what you're saying.
"An artist who can’t take constructive critique on their work is only hurting themselves and their potential. Conversely, and artist that can’t communicate a critique in a constructive way isn’t helping anybody."
Joined: 12/31/2009
Posts: 174
I'm sure load times will never be perfectly accurate. Loading from an ISO is faster than loading from a disc. I'm sure if you have Dolphin running from the disc, Dolphin itself would do momentary pauses while your system reads the data from the disc. Dolphin could easily stop advancing frames and whatever software that is recording the video/audio can't record those momentary pauses caused by I/O. A real GC/Wii on the other hand must be constantly sending video/audio and the recording device will be recording all of the momentary pauses to read data from a disc. I'm not sure if I'm correct (or even close) so don't quote me on that. The same thing happens to me with PCSX2. If I play KH2 from the disc, I get 1-3 second pauses every time I open a chest because PCSX2 has to wait for my system to get the data. If you were to record it, the pause is non-existent. I don't remember such lag from either of my PS2's but this shows PCSX2 and the recording software can wait for I/O to catch up while a PS2 and external recording hardware can't. I'm this has even been brought up about PSX runs at some point as well. Also, movies were never expected to sync from one emulator revision to another (especially when improving emulation accuracy). You should pick the latest revision (that can run the game) and follow through with the run unless you want to start over when a new revision comes out. If you are still concerned, maybe you should ask for a judge's opinion. I really want to see this run completed but I understand if you want to wait for Dolphin's development to settle down. I would be a little upset if DeSmuME is going through the same thing right now unless they add new features that saves me a few headaches later on to make starting over worth it.
Tub
Joined: 6/25/2005
Posts: 1377
SoulCal wrote:
With a game like RE4 though, I bet more people would be interested in the final game timer more than a frame count.
Actually, I'm not interested in either. Comparing TASes to speedruns is moot, anyway. I just want to see an awesome run. Precise headshots, insane lucky drops, clever strategies and possibly a glitch here and there, and I'm happy. So far, I like what I'm seeing. In other words, I'm hoping you'll continue regardless of loading time inaccuracies. If a comparison to console speedruns is important to you, research about the ingame timer should reveal methods to account for the differences. Does dolphin have a memory watch function? Figuring out the ram address of the ingame timer might be a good first move.
m00
Post subject: Console vs Dolphin Load Time Differences
Active player (335)
Joined: 1/19/2010
Posts: 383
Location: Texas
I was able to save at the end of ch1-1 and got a time of 3:32 with my current WIP, which is only 2 seconds slower than the SDA NewGame+ record. The NG+ also used Ditman to save about 10 seconds ahead of me in the first area alone. With that said... Watching the two segments side-by-side, my TAS was 3 seconds real time slower than the console (close enough, 3:30 compared to 3:32). Removing load times and comparing gameplay only, the console run is about 9 seconds ahead. So, my conclusion is that the game timer does indeed include load times. It seems each load screen instance saves about 1 second on Dolphin. Comparing my run to another New Game run in 3:58 for ch1-1, I am about 26 seconds faster in real time and game time. Console also uses a trick that, during loading cutscenes, you can pause the game with START and the game timer stops but still loads the cutscene. This saves the console about 5 seconds game time. Comparing both of these runs, it seems the loading time range is somewhere around 5-8 seconds for ch1-1. Here is that New Game run: http://www.youtube.com/watch?v=1SXMoIeTlTA So why am I so focused on the game timer when TASvideos uses frame counts? Well some TAS runs go for game time (several Metroid games do). Also, if each load screen saves 20-30 frames regardless, this will result in a run several minutes faster than it should be. Anyone who has played RE4 knows that there is a metric shit-ton of load screens. Zanoab is right in his assumption that an ISO would load faster than a disc. Sadly this is the obvious truth. For games like the other Wii submission MegaMan10, the whole game can probably fit into the RAM of the system, so load times don't exist.
Joined: 9/20/2005
Posts: 14
Location: Morocco
Wait a second. Don't load time change from one person to another (Wii model, DVD reader state and the state of the game disk itself) ? Hell, I bet even the same person may wait XXX frames for a cutscene to load, then if he replays the game we will have to wait YYY instead ...
Active player (335)
Joined: 1/19/2010
Posts: 383
Location: Texas
But with emulator, load times are always constants, or else playback files wouldn't sync. If you play Super Smash Brawl on Dolphin the first 15 second load screen doesn't exist like it does on the actual Wii. I don't care how long the load times are really varying from console to console, but rather Console to Dolphin. Besides if it were to vary on different consoles it wouldn't be that much.
Post subject: Re: Console vs Dolphin Load Time Differences
Experienced player (960)
Joined: 12/3/2008
Posts: 936
Location: Castle Keep
SoulCal wrote:
Well some TAS runs go for game time (several Metroid games do).
Its rare and often for good reasons (see sonic ect...), give us a good reason, theres none so far tbh, do we even get to see the game time screen at all, do we care ? Like you already mentioned this game have lots of loading, so you are suggesting to make the movie even longer just to win seconds on the end screen? imo this is very bad idea. We arent in sonic where every single levels (think rooms for re4) could be a whole seperate challenge and where pulling off any frames could warrant something more funny/entertaining, here its just toying with numbers. Ultimatly this isnt SDA so dont take game time objectives arbitrarly especially since this is something the site usually against (our movies are measured real time), if the runs arent comparable then it doesnt matter your time is higher than another run, given its still superplay, if your run doesnt do anything spectacular and is basicly just the same thing than a human would do, maybe you have a problem with the game choice.
Post subject: Re: Console vs Dolphin Load Time Differences
Active player (335)
Joined: 1/19/2010
Posts: 383
Location: Texas
arukAdo wrote:
Like you already mentioned this game have lots of loading, so you are suggesting to make the movie even longer just to win seconds on the end screen?
Increasing load times won't happen, so there's no reason to discuss that.
arukAdo wrote:
Ultimatly this isnt SDA so dont take game time objectives arbitrarly especially since this is something the site usually against (our movies are measured real time), if the runs arent comparable then it doesnt matter your time is higher than another run, given its still superplay, if your run doesnt do anything spectacular and is basicly just the same thing than a human would do, maybe you have a problem with the game choice.
Considering load times are not part of the actual gameplay or "doing anything spectacular" most people won't look for that. Since this is a TAS I'm creating I assure that the gameplay will not be "the same thing that a human would do." I consider this loading issue a flaw in the emulation rather than an improvement. Due to improved performance of the SNES emulator over the console, some tricks can't be pulled off for TAS, like the glitch in Crystal Snail's Stage from MegamanX2 caused by lag. I do see your valid points arukAdo, but I certainly won't waste time to add to the game clock (that's just silly!).
Experienced player (960)
Joined: 12/3/2008
Posts: 936
Location: Castle Keep
My point was just that you was trying to justify your tas been slower or equivalent with realtime-gametime explanations, i dont understand how you could be "slower" than the human speedrun, of course now well you talk about the first few rooms so yeah i hope for better later, and i guess its just theres no action in thoses. I wasnt making critic just for the hell of it and i wasnt trying to be offensive, but i probly confused some stuff you sayd because you was starting to compare frames with sda runs ect... Try to keep it simple, but yeah i was sarcastic i dont doubt this should be good tas candidate.
Active player (335)
Joined: 1/19/2010
Posts: 383
Location: Texas
Oh, that speedrun I am slower than is one where the player starts with all weapons and max health. He also has a gun that causes a glitch that increases running speed. The only reason I am comparing non-TAS to TAS is to see how much difference load times cause. Overall load times will save about 4 minutes for the entire run; about 8-10 seconds per sub-chapter. I am continuing with the run, but I'm redoing the village area. Hopefully 1-2 can be TASed in a week or so. EDIT: While re-TASing the village area, I can manipulate the first two enemies to drop 9900ptas, but I cannot get them to drop them simultaneously. If I hold Up+B only the woman drops 9900; if I let go of B for 2 frames then hold Up+B the first guy drops 9900ptas. I am not sure if the drops are connected (as in only 1 of them will drop 9900) but I bet they are. It seems like my delay of running switched the drop from one enemy to the other. Here are my dtm files if you want to see this for yourself. I am using the Japanese version. http://www.2shared.com/file/bvl-fgkH/villagelady9900ptas.html http://www.2shared.com/file/Pddo_hU_/villageguy9900ptas.html
Experienced player (960)
Joined: 12/3/2008
Posts: 936
Location: Castle Keep
SoulCal wrote:
He also has a gun that causes a glitch that increases running speed.
That sound a lot like a valid reason to use sram (maybe seperate branch), is there support for that yet in movies ?
Active player (335)
Joined: 1/19/2010
Posts: 383
Location: Texas
That function isn't yet in Dolphin. This game is killing me to TAS. Not just the drop manipulation, but that I have to watch my entire dtm file every time I want to add more to the TAS. When I try the read-only mode and load a state, it won't continue playing the game from where that load-state was. If I turn read-only mode off, the recording will continue from a load-state, but manipulating enemy drops like this is bothersome. When I create a dtm where the enemy drops a certain item, I play back the dtm and improve my movements but still get that item drop. Without the movie to continue playing in read-only mode, I have to save-state within 4 frames of the item being dropped (which is not efficient). Surprisingly I have had zero desyncs since starting this game, but the fact that I can't load-state and continue is bothersome. Also I am at a point where I could get hit by a Ganado in the village with no time wasted. I've heard that lower health means lower difficulty and greater chance of item drops. This seems to be the case for me, as getting hit causes 1 enemy I just killed 1 second prior to drop a grenade or 3300ptas. This could be a pure coincidence though... Also, a newer new WIP. This gets me 9000 more ptas and I have no item equipped for faster running speed. The most optimal situation would save about 6 seconds for the run; the first two enemies drop a grenade and 9900ptas, throw the grenade to the back group having the front enemy drop 9900(collected after cutscene), then immediately throw the last grenade to the last group and collect another 9900 and another grenade. I don't see this happening ever, so I'm continuing. Link to video
Active player (458)
Joined: 11/12/2010
Posts: 183
Location: Sweden
Impressive manipulation you managed to get now, and almost two seconds faster too! But try not to burn yourself out with all this loot manipulation now. I read what you mentioned in the video and wonder if it's possible to manipulate either the barrel to the right as you exit the house or the two barrels behind the merchant to be worthwhile? It will take a few seconds to reach those barrels though, but you might earn on it in the long run.
"An artist who can’t take constructive critique on their work is only hurting themselves and their potential. Conversely, and artist that can’t communicate a critique in a constructive way isn’t helping anybody."
1 2 3 4
20 21