3rd Strike Lag/Speed Tests

124

Comments

  • ryan.ryan. Joined: Posts: 1,918
    Holy shit well that might explain why I thought ps3 was shittier than Xbox. I've only tried ps3 at last years Evo where I played about 40-50 matches on a standard TE and it felt like complete dog shit. I blamed the poor ps3 never even thinking to blame the true turd, the TE.

    That also might explain why I've never felt comfortable on a TE on Xbox as well. About a half frame lag.

    Now I guess really it doesn't matter if you ONLY use your stick on console. But if you are borrowing a stick or primarily play arcade, this is pretty significant.

    Ps3 TE omg
  • TebboTebbo Play. Joined: Posts: 5,704
    edited July 2014
    neat info.
    16.66ish ms is about a frame (assuming around 1/60 of a second).
    looks like i should invest in an etokki omni lol
    Play more.
  • NoMoreFunlandNoMoreFunland Perfection equals obliteration! Joined: Posts: 1,229
    edited July 2014
    You sold me a bunk component Lance! Lol

    Real shame too. I wanted to get a PS3 sooner than later and try some of the PSN comp. Could buy a vx-sa.

    All these tests are great. So much theory flying around about lag early in the OE run. Amazing how people shut their brains down when they've decided a theory is the truth. Like people saying less than 1 frame of lag is imperceptible even though many people notice the difference.

    I wish MadCatz was paying attention to these things when they started making expensive joysticks.
    Post edited by NoMoreFunland on
  • Lance3rdLance3rd Alaska 3rd Strike Joined: Posts: 986
    It's hard for companies to pay attention when you have a good chunk of people in threads like that claiming 1-2 frames of lag doesn't matter or it's impossible to tell the difference. What a joke.

    Really discouraging how much dumb you have to weed through whenever you test stuff like this. Can speak from personal experience lol, it really saps your energy and interest. At least for us it was only one person, that guy is dealing with a large group of them. You set out to get to the bottom of something and learn more, and along the way you have to defend yourself from people who think the whole thing is imaginary and aren't interested in testing anything themselves but are 100% interested in voicing their opinion.
  • TebboTebbo Play. Joined: Posts: 5,704
    edited July 2014
    yeah there are some straight up dumb people is what i learned.
    or maybe just it's clear who is coming from a 3S perspective and who isn't. like in sf4 where the difference of a single frame matters a lot less because all the hit confirms are multiple normal sequences with static timing which give you a huge window to see if things are hitting.
    Play more.
  • petran79petran79 Here comes an old challenger Joined: Posts: 1,913
    sorry for not having knowledge in the matter, but one question:

    do better or improved input drivers compensate for input lag?

    Eg if you compare Shmupmame, released 18 months ago, to the latest MAME, a lot of changes were made to drivers, input included.
    If controllers can feel more responding and inputs are registered with greater accuracy, like in the arcade version, isnt this better than small input lag yet inaccurate inputs and timing?


    too slow!
  • DanderDander BANNED Joined: Posts: 7,227
    Yes there can be an increase in performance through the use of improved input drivers. I recall using a booster for my usb ports on my last rig and suddenly finding ggpo to be playable. That changed soon thereafter but the change was definitely noticeable.
    8tROOXi.png
  • isotopezisotopez Joined: Posts: 125
    If the improved drivers do in fact make the inputs more consistent, then it's a matter of whether or not you have more fun playing in persistent lag vs having your inputs occasionally lag. It's also possible that the variance in the lag data for the emulators is caused by inaccurate measurements from the monitor ghosting (we didn't have access to a CRT).
  • free_manfree_man Joined: Posts: 81
    If someone is in Japan and can play the Nesica version of 3S, please, can tell to us - after test it, of course - how is the speed of this system in comparison the others?
  • SpoRSpoR SpoR Joined: Posts: 350
    So by comparison, how does it run on GGPO, Supercade, and fightcade etc>
    Zerg!
  • cribbacribba Joined: Posts: 22
    Does anyone know if the default theme in Windows 8.1 causes lag like Aero in 7?

    If you play in fullscreen does the theme still cause lag?
  • exodusexodus anti-vid scrub advocate! Joined: Posts: 2,078 ✭✭✭✭✭ OG
    Windows 8 has aero permanently on.
    .
  • lamhazaartlamhazaart Joined: Posts: 91
    exodus wrote: »
    Windows 8 has aero permanently on.

    You can disable Aero in Windows 8, but it's a mildly involved process. You have to suspend winlogon.exe and then kill dwm.exe using procexp. You also have to rename some system dll files the first time you do it.
  • lamhazaartlamhazaart Joined: Posts: 91
    edited July 2015
    Regarding game speed, it turns out it's very easy to find out PC emulator framerate. Just need fraps+frametime viewer. Fraps' display already shows 59.5fps (only shows 3 decimal places) in speed-accurate emulators like Shmupmame and Groovymame. Looking at frame times, you see a frame interval correlating to that speed. The emulators themselves also indicate that they're running at 59.583. There's no need for adjustment for these emus.

    I found this out when testing black frame insertion in Groovymame, which can only run at exact monitor refresh rate (so, in my case 120hz). This caused Fraps to show 120fps, 8.33ms frame intervals (1/2 of 16.66), and an increase in speed when I timed single and multiple rounds equivalent to the difference in speed between 59.583fps and 60fps.

    TL;DR Shmupmame and Groovymame are already speed accurate without adjustment. However, my personal recommendation is to get them running at 60fps if you're on an LCD monitor - this means video will be smooth, frame delay will be consistent, and the current tournament standard(?) is 3SOE which runs at 60fps. If you're on a 120hz monitor, use black frame insertion to get near-crt motion smoothness.
    Post edited by lamhazaart on
  • IbokiIboki Joined: Posts: 206
    All things considered would it be better to practice offline on 360 OE or the ntsc-j PS2 port. I have ps2 pad/stick I'm comfortable with, and an xtokki converter for 360. I know ps2 speed is way off, but I find the xtokki converter to be pretty bad, even though it was hyped up by Markman and stuff. Once I can get my hands on a 360 stick it'll be OE for sure.
  • isotopezisotopez Joined: Posts: 125
    edited July 2015
    Ps2's speed isn't any different than oe's since both run at 60fps. Our tests showed ps2 being laggier than oe, but we aren't sure if that's because of controller lag or not. Just practice offline on whatever you have unless it's dreamcast.
  • IbokiIboki Joined: Posts: 206
    cool thanks!
  • shanoshano Joined: Posts: 284
    Doesn't seem like we can turn off Aero on Windows 10 from what i have gathered? Able to do a test on FBA and see the input delay results Windows 10 gives?
  • akuaku スピリチュアルやからね! Joined: Posts: 3,865
    Tried out Fightcade on Windows 10. Honestly I couldn't FEEL any additional lag/delay on it, but I'll leave it to someone else to explain if there actually is. :v
  • ChadChad Joined: Posts: 1,291
    Have you tried Fightcade on 8.1 aku? I played for a few hours and it was intolerable. Everything was underwater.

    Is 10 better?
  • isotopezisotopez Joined: Posts: 125
    shano wrote: »
    Doesn't seem like we can turn off Aero on Windows 10 from what i have gathered? Able to do a test on FBA and see the input delay results Windows 10 gives?

    We'll test it
  • akuaku スピリチュアルやからね! Joined: Posts: 3,865
    Chad wrote: »
    Have you tried Fightcade on 8.1 aku? I played for a few hours and it was intolerable. Everything was underwater.

    Is 10 better?

    Definitely felt better than 8.1, that thing felt like playing on a crappy Dell monitor. FBA was fine, Fightcade was Fightcade. I'm playing on a subpar monitor so there's always gonna be some delay on my part, but I never felt any additional delay in comparison.
  • JedahsMinistryJedahsMinistry BattousaiCV Joined: Posts: 542
    How about frameskipping/stuttering?
    Gettysburg, PA
    Netplay Poverty - http://www.twitch.tv/randomalt
    X-Men COTA and stuff - http://www.youtube.com/user/BattousaiCV
    Blurbs about games - http://twitter.com/BattousaiCV
  • vandarkholmevandarkholme Joined: Posts: 388
    I heard that different settings in 3soe bring out different results. Is that true?
    I'm planning to run a tournament with it and I'm not sure if I should worry about it
    Sutorito Faita : Sad strike
    vandark, you should hear a cool pherai story from me one day then. Add it to all your history. I'll probably remember it closer to EVO time and just retell it to the forum when I'm probably going to re-tell some more anyway. Its not the glory hole one though, that one's too legendary. You probably know more about that than me.
  • Smashbro29Smashbro29 Waiting for the new Framemeister... Joined: Posts: 7,199
    lamhazaart wrote: »
    exodus wrote: »
    Windows 8 has aero permanently on.

    You can disable Aero in Windows 8, but it's a mildly involved process. You have to suspend winlogon.exe and then kill dwm.exe using procexp. You also have to rename some system dll files the first time you do it.

    Tutorial please.
  • shanoshano Joined: Posts: 284
    isotopez wrote: »
    shano wrote: »
    Doesn't seem like we can turn off Aero on Windows 10 from what i have gathered? Able to do a test on FBA and see the input delay results Windows 10 gives?

    We'll test it

    Just curious if this is still happening?
  • isotopezisotopez Joined: Posts: 125
    shano wrote: »
    Just curious if this is still happening?
    We had camera issues. It's looking like we won't be able to test now
  • DontdodatDontdodat I know u want to Joined: Posts: 653
    Damn, I was going to run 3S meets on 360 (OE) CRT set ups. Probably still will because there are no cabs I can use.
  • DontdodatDontdodat I know u want to Joined: Posts: 653
    I don't think we have any Yuns good enough to make a diference with command grab. This is all a stepping stone to supergun and maybe a cabinet one day.
  • ilitiritilitirit Joined: Posts: 6,388
    edited March 2016
    aku wrote: »
    Tried out Fightcade on Windows 10. Honestly I couldn't FEEL any additional lag/delay on it, but I'll leave it to someone else to explain if there actually is. :v

    I'm trying it out right now and... I'm actually shocked that I can hit-confirm Ken's cr.mk into Shippu. I'm actually beginning to wonder if there's some lag-reducing technique that this build of FBA uses because AFAIK there's slightly more lag on the emulator especially on Windows 10 since you can't disable Desktop Composition and I'm playing in a window. I also don't think even ever been able to do the hit-confirm on anything else but PS2 + CRT and the actual cab. Only other thing that I can think of is that it might be playing a bit slower, but the FPS counter maintains a steady 60fps.

    Recorded at 30fps because my PC is ass:
    giphy.gif

    I'm using this to get Ryu to random block:


    Maybe that training mode module makes the game run a bit slower?
    Post edited by ilitirit on
  • GaijinblazeGaijinblaze fingerlicans Joined: Posts: 2,524 ✭✭✭✭✭ OG
    anyone can cherry pick a successful trial that had only six attempts. you would need to do a trial with at least 30 before anyone could begin to call it a hit confirm. also, the dummy sometimes begins the crouch animation before the move makes contact, which telegraphs what the dummy is going to do when the only two outcomes for it are stand hit and crouch block. if you can get the dummy to crouch random block, it would be more meaningful.
    This message will self-destruct.
  • ilitiritilitirit Joined: Posts: 6,388
    edited March 2016
    anyone can cherry pick a successful trial that had only six attempts. you would need to do a trial with at least 30 before anyone could begin to call it a hit confirm. also, the dummy sometimes begins the crouch animation before the move makes contact, which telegraphs what the dummy is going to do when the only two outcomes for it are stand hit and crouch block. if you can get the dummy to crouch random block, it would be more meaningful.

    Of course I repeated it more than that, and I did it on standing and crouching (this is admittedly harder than on PS2 - the super often doesn't come out). And on different characters. I just recorded that because it was the easiest to demo (that short sequence takes up over 1GB with DXtory - for some reason FRAPS stopped working a week ago).

    I even captured the frames required:
    FMkyB9V.png

    If you press later than 19f after you initially did the cr.mk the super won't come out (well, it seems to be random, like sometimes it works on frame 20 but I put it down to the input recorder dropping frames, unless it's recording the neg edge).

    No I don't have 100% accuracy on hit, but who does? That would be inhuman IMO, and TBH there's no way I'm doing this online anyway. The point is that this game isn't that laggy on FBA on Windows 10 (feels like there's less lag than on Win 8), unless there's something else very funky going on.

    Post edited by ilitirit on
  • free_manfree_man Joined: Posts: 81
    It would be a nice setup if I play 3S using Shmupmame in a Raspberry Pi(running Retropie, maybe?), a PS3/PS4 arcade stick and CRT TV or it doesn't have hardware enough to run it? If so, Shmup would have the same performance it running on a PC?
  • ViacheViache Joined: Posts: 781
    edited May 2016
    Fightcade on W8 was basically unplayable for me. Fightcade + W10 seems okay.
  • free_manfree_man Joined: Posts: 81
    Viache wrote: »
    Fightcade on W8 was basically unplayable for me. Fightcade + W10 seems okay.

    For me it was the opposite. Playing on Win 10 I feel that it's more laggy than playing on Win 8.1. I don't know why but it's it.
  • ParryAllParryAll Dangerous Posts Joined: Posts: 3,311
    Guys I just wanted to say a few things.

    I noticed with NKI's test of ST he made two mistakes.

    One was he measured Ryu from the animation point, when startup frames in that case begin while crouching. So he's off by 1 there.

    The next mistake he made was counting the first frame. From the moment the button press is registered, what's being displayed at that moment can't change, only the next frame can be different. Time moves forward that's why we see it as a motion, when it's really a series of stills. I'm sure you know this but the point is, the first frame isn't a frame of delay..

    I'm not saying you guys did anything wrong I'm just curious to know the details of the experiment.
    "Third Strike just has that flow which somehow alienates you from your surroundings and really plugs your brain into the match your playing and your able to comprehend so many different situations and mind games and mixups in such a limited amount of time and all this without even thinking" - Nubilous

    "[3rd Strike] makes you feel like the character is an extension of yourself when you get good enough." - AzureWolf

    "Street Fighter is now like The Simpsons. It's been bad longer than it's been good." -ParryAll
  • isotopezisotopez Joined: Posts: 125
    ParryAll wrote: »
    Guys I just wanted to say a few things.

    I noticed with NKI's test of ST he made two mistakes.

    One was he measured Ryu from the animation point, when startup frames in that case begin while crouching. So he's off by 1 there.

    The next mistake he made was counting the first frame. From the moment the button press is registered, what's being displayed at that moment can't change, only the next frame can be different. Time moves forward that's why we see it as a motion, when it's really a series of stills. I'm sure you know this but the point is, the first frame isn't a frame of delay..

    I'm not saying you guys did anything wrong I'm just curious to know the details of the experiment.

    From the original post: "I counted the first camera frame where the led is lit as frame 1..."

    Even though nothing new can happen during that frame, the player still perceives the elapsed time during that frame as part of the lag, which is why I included it.
  • ParryAllParryAll Dangerous Posts Joined: Posts: 3,311
    edited June 2016
    isotopez wrote: »
    ParryAll wrote: »
    Guys I just wanted to say a few things.

    I noticed with NKI's test of ST he made two mistakes.

    One was he measured Ryu from the animation point, when startup frames in that case begin while crouching. So he's off by 1 there.

    The next mistake he made was counting the first frame. From the moment the button press is registered, what's being displayed at that moment can't change, only the next frame can be different. Time moves forward that's why we see it as a motion, when it's really a series of stills. I'm sure you know this but the point is, the first frame isn't a frame of delay..

    I'm not saying you guys did anything wrong I'm just curious to know the details of the experiment.

    From the original post: "I counted the first camera frame where the led is lit as frame 1..."

    Even though nothing new can happen during that frame, the player still perceives the elapsed time during that frame as part of the lag, which is why I included it.

    That's fine but I would argue that's not lag.

    I have no idea if other testers account for that sort of thing and thank you for letting me know.

    Just so I'm clear about the second part: what I'm suggesting is. Take any move, and then grab the frame data online from startup , to active, to recovery ok. And then measure the animation from start, I.E. button press, to when the first active frame hits. Because often times startup frames aren't animation frames you see. Using the Ryu in ST cHK example, during the first startup frame he is still just crouching, the animation itself doesn't start until the second start up frame.

    Again I'm not saying you're wrong I'm just suggesting a mistake you may, or may not have made and to do it again if so.

    Also if you could please PM me the equipment used and where you got it. I really want to do my own tests on a bunch of different games.
    "Third Strike just has that flow which somehow alienates you from your surroundings and really plugs your brain into the match your playing and your able to comprehend so many different situations and mind games and mixups in such a limited amount of time and all this without even thinking" - Nubilous

    "[3rd Strike] makes you feel like the character is an extension of yourself when you get good enough." - AzureWolf

    "Street Fighter is now like The Simpsons. It's been bad longer than it's been good." -ParryAll
  • ParryAllParryAll Dangerous Posts Joined: Posts: 3,311
    edited June 2016
    So in other words. Just going back to my first point. If a video game had zero frames of input lag. Meaning from the microsecond the button is pressed, processed, the displayed frame cannot change only the following frame can show a change. That's just how TV/Monitor technology is you get that. If the next frame displayed by the monitor showed a change in animation or movement id say that means 0 lag, you'd call that 1 frame. Which is fine but I just want to make that distinction for people reading. Which is that by his definition all video games have to have 1 frame of input delay and 0 frame input delay can't exist. This is actually difference in philosophy, nothing more than that.

    And I think most people here know this but for those that don't. Television/monitors, whether you're watching a movie or playing a game, the technology is flashing images. Every second in a 24 fps film 24 unique images display in that one second. They used to call it a flicker rate, and the max flicker rate for old CRTs is 60hz, 60 frames per second because that's what the images are frames. Think of it like an old film camera reel that passes through a projector. TV/Monitor technology is just that. Our minds do not see the flickering we observe it as one continuous motion. It's a trick basically.

    So for video games running at 60hz like any Street Fighter is 60 images a second. You break that down and you get 16.67ms per frame. When you press the button, an image is already displayed. Only the next image displayed can show a change, not the one already being displayed. Get it? But our brains again process that as instantaneous, and as one smooth moving image. But it's actually not moving, just changing every frame.

    So if the hardware can register and process your input within 16.67 milliseconds and effect the next frame, I would call that 0 frames of input delay, isotopez defines that as 1 frame and perhaps other testers do (or do not) but again I'm simply making this a point for people reading to make up their own minds about. By isotopez's view all games will have at least 1 frame of lag, and there's nothing wrong with that its just worth pointing out.

    I would be really curious how other testers go about measuring. It's probably the same as you isotopez.

    I'm also curious if a supergun and artificial controller was used or an arcade cabinet and if so what kind of cab.

    There is a noticeable difference for me playing 3rd Strike locally at the arcade vs at home on my 360 and CRT. I don't think half a frame is noticeable which is why I'm questioning these results. Again I'm not saying you're wrong.

    Do you understand the second point I was asking about startup frames and using the entire animation (or at least half of it) to eliminate error?
    "Third Strike just has that flow which somehow alienates you from your surroundings and really plugs your brain into the match your playing and your able to comprehend so many different situations and mind games and mixups in such a limited amount of time and all this without even thinking" - Nubilous

    "[3rd Strike] makes you feel like the character is an extension of yourself when you get good enough." - AzureWolf

    "Street Fighter is now like The Simpsons. It's been bad longer than it's been good." -ParryAll
Sign In or Register to comment.