New discovery: Xbox 360 version of sf4 series has different timing from arcade/ps3/pc

Classic Fighting Games Ultra Street Fighter IV
101

i cant speak about the PS3 version , i never tried it, but for the Xbox version to match the input lag i have to turn on Vsync on PC to get the same input lag feel and practice a bit before i go to play on it everytime, ther’s no way Vsync off on PC is the same as the Xbox, the first time i switched i couldnt even do simple combos, the lag was just ridiculous , it was to the point where i felt like an online laggy match on PC, but it was offline and on Xbox

as for the monitors, we play on a BenQ gaming one, also i can tell you that my friends who play only on the Xbox and never even tried the PC version have no trouble at all and dont feel any lag, and they are better players than me, they even laugh when i try to explain to them the input lag difference with Vsync

just ask someone for an Xbox you’ll see the difference instantly

102

http://i1181.photobucket.com/albums/x437/corruptstaffmember/frame%20by%20frame/1-1.jpg
http://i1181.photobucket.com/albums/x437/corruptstaffmember/frame%20by%20frame/2-1.jpg

http://i1181.photobucket.com/albums/x437/corruptstaffmember/frame%20by%20frame/1.jpg
http://i1181.photobucket.com/albums/x437/corruptstaffmember/frame%20by%20frame/2.jpg
http://i1181.photobucket.com/albums/x437/corruptstaffmember/frame%20by%20frame/3.jpg

What does it mean?

103

As far as I can tell from the video tests of my own (link in original post), it means that the Xbox version is running 2 frames faster than arcade, PC, ps3 due to lack of proper vsync.

104

Ya I guess AMD decided to use third world vsync and the third frame is a result of not having Toshiba on board to design the Xenon.

. . .

No sir, I’m not touching this topic again.

105

If you want proof of the botched Xbox v-sync look no further than this video https://www.youtube.com/watch?v=0kZv_eme1TM&list=UUgG9t9G0jRTz83bL1mwxmhw

Notice how the tearing % under the frame rate is jumping all over the place on the 360 yet remains locked at 0% on ps3 (proper v-sync). Keep in mind that street fighter x tekken runs on the same engine as sf4. Also notice that the ps3 frame rates matches or exceeds the 360 (basically locked at 60fps) with 4 characters on the screen at the same time, this also makes me very suspicious of claims that the ps3 drops the frame rate with 2 characters on screen on sf4.

I never said amd has third world v sync, just that the sf4 engine was developed on an nvidia platform and that is why v-sync seems to have problems on 360 (amd gpu) vs ps3 (nvidia gpu), if the game was developed on an amd gpu things could have just as easily gone the other way.

Anyways if you want to post pictures of a video that I’ve referenced and used as part of my own analysis since my first post than that shows that you really haven’t understood anything that I have tried to explain. The information is there for those interested.

106

Because your “information” you gathered is flawed and it’s pathetic that you’re trying to push this agenda forward.

107

I’ll try to get my hands on an Xbox for a final definitive comparison over the next few weeks (leaving on vacation soon). All other videos show it should match pc with v-sync off but I would still like to test it myself with my own high speed camera. Thanks for the info.

108

Needs legitimacy of the real hardware to prove it. It’s not enough to say that it is nearly the same, what matters is consistency.

Also, judging that your point of the thread is to compare what’s closest to arcade, there is an issue. Ask any good player that a new game came out from the SFIV generation that has these same issues (if necessary, pretend that it is SFIV but it is brand new), and he/she had the choice of system to choose from given these facts from underground tests (disregarding biasness via region/system and controller layouts assuming the standard for players is arcade sticks).

Lay out the terms that:

Xbox 360
*Less input delay than PS3 and PC (with v-sync), can be nearly the same with PC if the PC version has no v-sync.
*General computer specs benefits over PS3 and is usually consistent in the frame rate department, but still gets disadvantage over PC.
*Easy access for general casuals/tournaments.

PS3
*More input delay than Xbox 360 and PC (no v-sync), can be nearly the same with PC if the PC version has v-sync (read next bullet).
*Inconsistent frame rate that brings another issue to input delay and timing (ie: it will either lag 1~2 frames more depending of that certain frame).
*Easy access for general casuals/tournaments.

PC
*(With V-Sync) More input delay than Xbox 360 and can be nearly the same with PS3. (Without V-Sync) Less input delay than PS3 and can be nearly the same with Xbox 360 (or potentially, maybe less. More proof is needed.)
*General computer specs benefits over PS3 and could be consistent in the frame rate department, but has basic PC flaws.
*Not so easy access for general casuals/tournaments.

Arcade
*No real proof is given out for tests of the real hardware. It may play a similar role to PC but no one can prove that it is a guaranteed fact.
*Not so easy access for general casuals/tournaments (in most regions.)

Out of these four, arcades can be scratched off due to lack of arcades in the whole world sadly. Most people would rule out PC too due to the hassle of bringing it for about anywhere. That leaves PS3 and Xbox 360.

This is something I don’t see mentioned in this specific thread, but majority of the players here and in tournaments always advise that whatever system lags the least, is likely the most preferred in their books. Looking at the PS3 section, it clearly disassociates this preference players have regarding input delay. There are more input delay issues with the PS3 than the Xbox 360.

It also feels awkward how you sort the PS3 with the Arcade and PC in the title, considering the PS3 version has an inconsistent frame rate. Seriously, knowing that you have to guess if it either lags one frame late or one frame early that happens much more often than the other systems (outside arcade, no real tests of the actual hardware exist atm) is a problem.

109

I have been asking for some sort of proof (video or frame capture) of ps3 frame rate dropping since the begining of this thread. Digital foundry’s in depth comparison(these are experts) found both 360 and ps3 versions of sf4 to be LOCKED at 60fps, unlike mvc3 which was found to indeed be dropping frame rate on ps3. If anything we see a more varied input timing on 360 due to v sync problems. This matches the huge amount of dropped 1 frame links this evo. If someone can show me a concrete example of the ps3 dropping below 60fps (offline) more than a 360 I will be convinced that I was wrong. Also I am going to look into getting arcade footage (actual cabinets) showing the timing matching pc (v sync) and ps3 to put this to rest. I am that sure of my information and analysis.

110

it doesnt even matter really. even if the 360 was the console that dropped frames or had slowdown i think people would still play on it more just because the netcode in general is better, and its a lot more intuitive for playing with friends

111

There’s also the whole ordeal of everyone hating PS3 at EVO for like 3 years in a row. Going back to PS3 would be a mess at this point.

112

Good point.

At first I thought you were wrong with this conclusion but you’re right. With more input lag you need to press your buttons sooner. If you’re used to this and then play with less input lag your button presses will be too soon.

113

I played the real arcade game on Taito X2 hardware (it was AE not ultra) again this week in Chinatown.
Managed to get 20 wins in a row and the timings and combos matched with the PC and Xbox 360 to me. It was an easy transition.
Not scientific I know but my PS3 version I have at home feels different to the arcade in terms of timing combos.

114

Exactly. Most mid to high level players practice links in the training room using the same concept. If a link doesn’t come out at all you pressed it too early, if it comes out but is auto blocked then you pressed it too late. Using this knowledge we can figure out when Daigo, Sako, and Kazunoko keep dropping links due to moves not coming out that it’s happening because they’re pressing buttons too early. This means that either they’re having an off day (it’s a possibility) or that the input timing is faster than what they’re used too at the arcade (this makes more sense to me seeing the consistency of this lately on arcade players switching to 360s on lagless evo monitors)

115

found this if anyone is interested, regarding the supposed PS3 delay:

During performance of any Super and Ultra moves, along with the real-time pre and post fight intro and ending sequences, the PS3 game no longer drops resolution down from 720p to 1120x630 unlike in SFIV. It seems that through optimisation, that Capcom have managed to solve some of the bandwidth issues that may arise from the fact that PS3’s RSX GPU has access to less overall bandwidth than either the 360, or the Taito Type X-2 board the original SFIV runs upon. Essentially, all the transparency effects that are displayed onscreen during a Super or Ultra move vastly eat into each system’s bandwidth. However, it just so happens that this time around, that capcom have found a way of maintaning full 720p resolution on both platforms at all times.

Perhaps the most noticeable difference comes in the form of texture detail, or more specifically, from the observation that the 360 version has slightly more detailed textures, which are used in some of the background scenery found in the game. These, along with some of the background objects are indeed rendered in 1120x630 instead of 720p on the PS3 build. You can see this happening clearly in the screenshot below, just look at the trees in the top right hand corner.

Texture filtering on the other hand looks to be identical on both versions of the game, which is somewhat surprising, considering the PS3 usually gets the exclusive advantage of having almost free use anisotropic filtering. This time around, both PS3 and 360 versions feature equal amounts of AF, with detail being visible far off into the distance. Yet another sign that the game isn’t perhaps pushing the 360 as much as it is the PS3, with all its use of alpha transparency effects sucking away potential performance.

*When it comes down to it, Super SFIV is pretty much equal on both platforms, with the PS3 game becoming even closer to the 360 one compared to last year’s SFIV. Some differences remain, like the lack of any anti-aliasing on the PS3 game, along with one or two missing effects and the occasional lower resolution texture. The use of self-shadowing on the PS3, and equal amounts of texture filtering balance out any differences to the point that when seeing the game in motion it doesn’t really matter at all. *

116

This means nothing unless you are positive that you are playing your 360 on a monitor with the same input lag as arcade, and pc. If you are playing your 360 on a typical HDTV with 3-4 frames of input lag than you would be about even with the arcade version on a typical arcade vewlix monitor with 1 frame input lag. This is why I did all my tests on the same monitor. If I were to switch my ps3 to the HDTV in the living room than it would have been running a few frames behind the pc and skewed my results. If you can verify for me what monitors you play your 360 and PC on then I could calculate where they should fall compared to the arcade version and we can see if that matches what you are saying.

117

I dont own the 360. The 360s ive played are at the game center and were set up using CRT monitors.

But I do know my PC monitor has 0 frames of lag 14ms.

Not sure what monitor the Taito X2s were using. My PS3 is using a different monitor Ill give you that.

118

Hmmm. Seeing that Capcom has gotten the hang of developing the SF IV series on PS3 in general, the differences of the 360 & PS3 versions of Ultra Street Fighter IV today shouldn’t exactly be too far off at all. Kind of makes you wonder why EVO, etc. are using the Xbox 360 to play this game for anyway.

119

It really is ridiculous how many different ways input timing can be affected between the different platforms and versions of SFIV.

Just from Arcade SFIV: AE2012 to Arcade Ultra SFIV, it’s possible timings did not stay exactly the same. But nobody knows, because it hasn’t been (and probably wont ever be) tested. AE2012 runs on Taito Type X2, and USF4 runs on Taito Type X3.

Let’s take a look at the hardware for each, from wikipedia:


Taito Type X²

    OS: Microsoft Windows XP Embedded SP2 [3]
    CPU: Intel LGA 775 CPU. Supported CPUs include Celeron D 352, Pentium 4 651, Intel Core 2 Duo E6400
    Chipset: Intel Q965 + ICH8 (dg31pr +ich7)
    Video output: 640×480 (VGA), 1280×720 (HDTV 720p), 1920×1080 (HDTV 1080p)
    RAM: 667/800 MHz DDR2 SDRAM. Supported capacities 512MB, 1GB, 4GB.
    GPU: PCI Express ×16-based graphics. Supported GPUs include ATI RADEON (x1600Pro, x1300LE) or NVIDIA GeForce (7900GS, 7600GS, 7300GS)
    Sound: Onboard Realtek HD 7.1 channel Sound (supports add-in sound cards)
    LAN: 1000BASE-T 10/100BASE-TX
    I/O ports: 1x JVS, 4× USB 2.0, 1× serial (max 2), 1× parallel port, 2× PS/2, 2× SATA
    Audio inputs: AKG C535EB Stage Microphone, line-in (Surround 7.1)
    Audio outputs: 7.1, SPDI/FX
    Expansion Slots: 1× PCI Express ×16 (used by video card), 1× PCI Express ×4, 2× PCI
    Storage: SATA 3Gbit/sec Hard Drives


Taito Type X³

    OS: Windows Embedded Standard 7 64bit / Windows XP Embedded SP3 32bit
    CPU: Intel Core i5 2400
    Chipset: Intel Q67 express
    Graphic: AMD RADEON HD 6770
    Sound: 7.1ch HD AUDIO
    Memory: DDR3 2GB
    Storage: HDD 160GB (2.5 inches)
    USB: USB3.0 × 2 group + USB2.0 × 2 group
    Network: 1Gbit/s × 2 port
    Serial: 1 port
    Power Supply: 600W

In addition, in Type X³, hardware configuration changes are possible for each game title, following the lineup is as an optional part.

    CPU: Intel Core i3/2120, Core i7/2600
    Graphic: NVIDIA GeForce GTX560Ti
    Memory: Up to 16GB
    Storage: Up to 3TB HDD
    SSD: 16GB

Let’s check out just a few, of the many, things that could very easily affect input timings from Arcade AE2012 to Arcade USF4:

Game software: From AE2012 to USF4. This should be immediately obvious to anyone. It’s a WHOLE NEW VERSION of the game, and made for different hardware. To just automatically assume, without any testing, that the input timings stayed exactly the same is ludicrous. Almost any of the game’s changes could have unintended consequences.

Operating System: Whether it switched from Embedded XP SP2 to Embedded XP SP3 or Embedded 7 64bit, the change is still a massive one. XP SP3 has 1,174 fixes, and a change to 7 64bit would be drastic.

Chipset: From Intel Q965 to Intel Q67 express. The chipset handles USB ports, which I believe is how the JVS I/O board connects (If I’m wrong, you can ignore this). Anyone who’s a hardcore PC gamer should know first hand how your chipset can affect your USB’s performance. Worse yet, if they’re using a USB 3.0 port to plug the JVS I/O into for the Type X3, it’s even more of a possibility for input timings to have changed, due to backwards compatibility with USB 2.0 being buggy as fuck with a lot of drivers.

GPU: I haven’t been able to find exactly what graphics cards AE2012 and USF4 use on their respective arcade machines, but it doesn’t matter, as they are certainly not the same. Let’s assume AE2012 uses the nVidia 7900GS and USF4 uses the nVidia GTX 560 Ti. This is a major area where input timings can be altered. They both use different driver versions, each with it’s own set of bugs, fixes, and features that are not shared with the other version. Take a look at the changelogs of the drivers from when the 7900GS came out all the way up to when the 560Ti came out and you will most likely be amazed at the kind of shit goes on with graphical drivers. Major features such as the way anti-aliasing and vertical sync (to name a couple) are implemented get broken/fixed/changed, sometimes on a game by game basis and sometimes overall.

Just between arcade versions, the possibility for different input timings exists. Yet here we have seph13, who hasn’t tested either arcade version, and is running an AE2012 rip with emulated JVS I/O on his PC.

Let’s look at a few of the ways his PC setup may differ from the arcade:

Game software: He’s comparing an Arcade rip of AE2012 with console USF4. Read above.

He says he has a more powerful graphics card. Refer to the GPU section above to see why this may cause discrepancies. He also almost certainly is not using the same GeForce driver version that is on the Type X3.

Emulated JVS I/O: As said above, the JVS I/O connects via USB (I believe) to the system. He doesn’t have an actual JVS I/O, and is emulating it. He most likely has a different chipset, with different drivers, and who knows what method of input he’s even using. Is he using a USB game pad, stick, or keyboard? Maybe he’s using a PS/2 keyboard? All of these devices could have very different input timings (even within each category: many sticks have different input timings), and who knows what kind of fuckery could happen with emulated JVS I/O.

Monitor: I don’t know what monitor they use in the arcade setups, and I believe seph13 said he is using the evo monitor, which if it’s still the same, would be the ASUS VH236H. The VH236H has 10ms of lag according to http://displaylag.com/, and who knows how much delay the arcade monitors have.

Other software: Any software which interacts with operating system at a low level (such as drivers) can mess with input timings. It’s even been speculated (but completely unproven as far as I can tell) that things such as Sun’s Java and Adobe’s Flash can increase input delay: http://www.overclock.net/t/1433882/gaming-and-mouse-response-bios-optimization-guide-for-modern-pc-hardware-2013-r0ach-edition Again, practically nothing in this overclock.net thread has been scientifically proven, but it’s something to think about as a possibility for discrepancies between an actual Arcade machine, which would not have such extraneous software, and a regular PC, which typically does. But even if you were to format and run a clean install of whichever embedded OS either arcade system uses, you would still have a litany of other differences, some of which are listed above.

So in the end, whether seph13 is right or wrong, nobody knows, and his testing is 100% flawed and should be completely disregarded. The only conclusion you can draw from seph13’s testing, is that he has no understanding of what the fuck he is doing.

120

I said I wouldn’t respond to you any further but since you made a somewhat technical sounding post I will oblige for a moment.

A lot of the concerns you bring up would make sense and could be legitimate factors to affect input lag, however for the millionth time I’ve already tested all of this stuff. I’ve run ae 2012 (pc version) on both windows 7 and windows 8, same input timing, graphic settings on max and graphic settings on low again same input timing, I have tested ae 2012 on at least 4 different versions of the drivers and many different internal driver settings, again same input timing, tried on both usb 2 and usb 3 (using Hori and Mad Catz Sticks) and guess what…same input timing. I’ve even tested ae 2012 on a different pc with different chipset and gpu on my monitor (yes, VH236H) and guess what? SAME input timing. It always matches the ps3 and always matches the arcade version. The only thing that does indeed change the input timing is the V-Sync and max pre-rendered frames setting. So guess what? The arcade version should also be running the same ( and it does as you see in my video demonstration, that you’ve clearly never watched) and a transition to Taito Type X3 (just like the pc version) has no reason to change the game engines input timing.

You can scream all you want that my testing is 100% flawed and should be disregarded (which you’ve been doing since I started this) but that does not make it so just because you want it to be flawed. Almost everything you mentioned in your previous post is stuff you could have just asked me to clarify in my methodology and to explain in further detail. My main system is a Windows 8.1 64 bit Intel Core i5-3570k, 8gb ram, Nvidia GTX 780 gpu running on a Z77 intel chipset. I am a 31 year old Computer Science major but I guess I have no understanding of what the fuck I’m talking about.