PlayStation Move lag analysis

By Richard Leadbetter Published 17 March, 2010

Updated: DF counts the frames.

(This piece has been updated! Scroll down for additional clarification of our measurements.)

Bold claims have been made for PlayStation Move in terms of its latency, or so-called controller lag. At both the GDC press and developer events, Sony itself pegged the lag of the new motion controller at "under one frame".

Previous information from the guys who created the PC drivers for the PlayStation Eye camera also indicated that the mechanism is extremely low latency. In that case, a lag of one frame was measured, and the chances are that in this case it was probably impossible to measure if it was actually any faster.

At the nuts and bolts level, then, it's difficult to argue with the case being put forward when both official and independent sources corroborate each other so closely. In use, however, on actual PlayStation Move software, it's clear that there is some level of latency.

Regular readers of Digital Foundry will know that lag is basically inevitable. The creation and flipping of a framebuffer within the console causes lag, and in measurements (not just ours), the very fastest response has come from the PS3 XMB, which has been pegged at three frames, or 50ms.

On top of that, of course, the proliferation of flatscreen panels means that we have display equipment where latency can change radically even between screens made by the same manufacturer. Anything from two to five frames' additional delay (33ms to 84ms) is fairly common.

Muddying the waters still further is that the PlayStation Move is such a precise implement that game-makers may feel the need to smooth the input, reducing natural player jitter, but again introducing further latency.

So, with all that in mind, I set out to get some idea of the lag in Sony's motion controller while at GDC. The chance to get hands-on with Sony's augmented reality demo (used by the firm itself in demonstrating the controller's tech credentials) was probably the best possible test, and presumably the screens Sony used would be the very fastest consumer-level panels it has bearing in mind just how much "precision" was mentioned at both of the company's GDC events.

A cheap and cheerful Kodak Zi6 handycam was used to film the proceedings. It's an unremarkable camera, but it can film at 60 frames per second, making it very useful for measurements like this where temporal resolution is far more important than the number of pixels.

Hanging out with the Move waggle homies, testing motion controller response.

The augmented reality demo is a worthwhile test for a number of reasons. Firstly, actual game-level processing is at a minimum. The demo is simply displaying the camera feed (in itself used for some of the motion control calculations), and overlaying a couple of fairly simply 3D shapes. Make no mistake, this isn't some Killzone 2 level workout of the PS3 hardware.

While exact latency measurements aren't possible in these conditions, a ballpark idea of the level of response isn't a problem at all. The methodology is remarkably straightforward. Keep your hand as steady as possible, then make fast motions with the controller. Count the frames between your hand moving, and the motion being carried out on-screen.

Equally illuminating is to stop your movement suddenly, then count the frames necessary for your on-screen counterpart to catch up. While not 100 per cent accurate, repeat the process enough times and the frame difference becomes fairly evident.

Bearing all of that in mind, and recognising that we don't know how much latency the display itself is adding, I'd say that a ballpark figure of around 133ms of controller lag (give or take a frame) seems reasonable, certainly not the ultra-fast crispness of response we see from games like Burnout Paradise or Modern Warfare, but fine for most of the applications you would want from such a controller.

Also worth remembering is that the motion control itself is only a part of the mechanics within the Move. Much of the internal mechanics, along with the button presses, will all be sent to the PS3 via the Bluetooth connection: lightning-fast and perfectly equivalent to the response levels of the DualShock 3. In terms of a more definitive test, a 60FPS game with comprehensive support for both Move and the traditional controller could be interesting...

Update: Some interesting feedback to this piece, so I thought an update with clarification on the measurements would be in order.

First up, I chose to use this demo as a best-possible example of Move operating in ideal conditions - no game logic running and basically in "tech demo mode", so as close as conditions allowed to what we would hope to be gameplay at optimum conditions. This is what the video itself is showing - the time taken between input by the player and it being recognised, processed and displayed by Move.

This is the generally accepted definition of controller lag (certainly what's been used in our previous pieces on the subject, and that includes the 200ms pre-production Natal measurement), although we do like to filter out the display lag in the measurement where possible - and in this case, we can't. So hopefully this element of the piece is now clear.

So, to put it plainly, based on frame-counting of the motions shown in the video, and others, the measure of 133ms overall is an indication of the gameplay experience as the camera captured it. So, this includes controller and camera input, processing, display of the completed frame plus the additional latency from the LCD (probably in the region of one to two frames on a set like this).

Interestingly, this measurement puts us in the ballpark area of the PlayStation Eye test I did a couple of weeks ago. As mentioned at the beginning of the piece, there is independent corroboration for Sony's sub-frame latency claim at the mechanical level, but the key measurement comes from gameplay.

So, this is indeed a tentative analysis with one big unknown remaining (the lag of the display itself), and others still undefined, for example whether the demo itself is smoothing the data and introducing further latency as a result.

However, as an indication of the response level of the games played at the GDC reveal (and I played them all), this overall measurement seems sound in the meantime. Once the controller is here in the office, we can get it running on properly calibrated equipment to see how close we are.

Picture of Richard.

About Richard Leadbetter

Rich has been a games journalist since the days of 16-bit and specialises in technical analysis. Commonly known around Eurogamer as the Blacksmith of the Future.

Read more articles from Richard by visiting the Richard Leadbetter archive page.

You may also like...

Comments (83) Latest comment 2 years ago

Comments threads automatically close after 30 days, but please feel free to continue chatting on the forum!

  • jamhead #1 2 years ago

    "...but fine for most of the applications you would want from such a controller"

    or rather, ...fine for most casual game players. Sadly, from the look of the video it does look like enough to annoy me, and prevent suspension of disbelief. Ultimately, games won't feel as immersive as I want this to make them. I doubt this is a Move specific problem - could be that it's just not solvable with the technology available today (at least not in an affordable fashion).

    Most of the games I like - your arcadey types - will require the input mechanism to feel extremely responsive. Until someone solves this issue it will be nothing other than a novelty consigned to post-pub party games... :-(

  • Roarster #2 2 years ago

    I don't think you can call the 133ms seen here controller lag. You've no idea what the game is doing in the background and what processing it's doing on the image before it appears on the screen. The only conclusion that you can draw is that this game has a response time of approx. 133ms meaning that the absolute maximum lag introduced by the controller is 133ms but could be far less depending on the game.

    The same goes for all of the Natal tests you've done as well...

  • mcmothercruncher #3 2 years ago

    Jesus, that was LOADS of exposition and then a "ooo, about yay big" single line at the end. Bit of a waste of time.

  • cyacomini #4 2 years ago

    As a bit of a Quake expert - I can tell you that 125ms lag (give or take!) is enough to mean the difference between being at the top of the leader board or sitting in the corner with a grenade up your ass.

    If I joined a session with 175+ms lag, I'd just leave.

  • Nazo #5 2 years ago

    Noticeable lag, but is that really being caused by the controller? It looked to me as though the video image as a whole was lagging behind, the graphic itself remaining remarkably well rooted to the hand.
    If it were controller input lag wouldn't you see your hand move then the generated graphic afterwards?

  • jonbwfc #6 2 years ago

    Dear God..
    Can we please put an end to these 'finger in the air' pieces? I don't mind at all the DF anaylsis that actually uses appropriate tools and reference kit to come to an objective result but basing a piece on a bunch of footage taken with a consumer camera at a trade show with an unknown display on what may be little more than alpha software? Really? That's honestly worth an article now?

  • Der_tolle_Emil #7 2 years ago

    Of course it's worth an article. The only two things that matter is precision and delay. You cannot increase the precision all that much my interpolating/extrapolating movements if the timeframe the raw data needs to be registered is already too big.

    This is definitely no scientific proof at all but certainly a good measure of where things are going.

  • americorloliveira #8 2 years ago

    133ms is very good, is in line with a 30fps game.

    And if part of the lag is from the TV it is very good in deed

    congratulations SONY

  • abigsmurf #9 2 years ago

    "Also worth remembering is that the motion control itself is only a part of the mechanics within the Move. Much of the internal mechanics, along with the button presses, will all be sent to the PS3 via the Bluetooth connection: lightning-fast and perfectly equivalent to the response levels of the DualShock 3. In terms of a more definitive test, a 60FPS game with comprehensive support for both Move and the traditional controller could be interesting... "

    I'd consider it is an incredibly big problem if the button presses aren't synced to match the lag.

    Imagine you're playing Call of duty Move. You're aiming at the left of the screen, see an enemy on the right side and flick the controller to the right and press the fire button in one lightning fast manoeuvre. If the buttons aren't synced to allow for lag, you'll end up firing to the left or centre of the screen rather than at the enemy on the right where you were aiming.

  • Murton #10 2 years ago

    There's quite a few things wrong with this article.

    1: you quote a figure for controller lag but you don't know the latency of the VDU so you can't possibly know how much lag is down to the controller, it would have been more correct to quote the total lag accounting for an unknown display latency.
    2: the demo you are using is quite clearly debug code, it's therefore not unreasonable to assume that it is unfinished and not fully optimised, making it unsuitable for this sort of test
    3: you used a cheap handheld camera to get your footage, it's capable of 60fps but did it actually achieve that and was it steady?

    The entire article reads as an attempt to refute Sony's claims of a one frame latency in the controller by any means necessary. Come back when you are:

    1: using a known display and can account for its part in overall latency
    2: using press demo or full release code and not pre-beta/debug code
    3: using the right tools for the job

    It would also be nice if in the proper report on Move's performance you could write it impartially rather than in the usual DF fashion, though that might be asking a bit much of you.

  • Mkwone #11 2 years ago

    These numbers are great but how does it companre to the Wii and Natal? And more improtantly to 'traditional' game controllers?

    I don't know about anyone else but millseconds is a bit difficult to grasp without comparison.

  • jonbwfc #12 2 years ago

    @Der_tolle_emil "The only two things that matter is precision and delay. "
    Yes, and they've chucked out the former to avoid the latter.
    You're entirely missing the point. Whatever the estimated (more like guessed) precision and delay discerned in this article, it's an enormous assumption to make that things won't change significantly before launch. Right now nobody should be doing quantitative analysis on the performance of PS Move (or Natal for that matter) because they're six months away from retail. It doesn't matter what particular aspect of the technology is more or less important, it's simply too early to judge any of it.
    Edited by 1 at 17/03/10 @ 11:12

  • Bigglesworth #13 2 years ago

    DF's scientific methodology is particularly impressive here.

  • robg #14 2 years ago

    Come on people, it's just a one-page article that gives us a rough idea. The controller isn't going to add on loads of lag, which is a good result, and worth reporting on.

  • TonyCB #15 2 years ago

    @robg

    yes mate that's fair comment if it was on eurogamer main site or another gaming site, the fact this is on DF means I kind of expect a little more due diligance

  • americorloliveira #16 2 years ago

    The normal lag, on a normal dualshock3 for a 30fps game is 100-133ms, so 133ms is very good

    http://ww w.eurogamer.net/articles/digita...

  • GamesConnoisseur #17 2 years ago

    There is no anti or biases, as all new product, be it Natal or Move are equally measured and weighted, so that we as gamers can get pass ALL the bulls from the companies involved.

    They will ALWAYS talk up their product, afterall Sony is perhaps the famous drummer! Emotion Engine, HD starts... and all that.

    Still one page article with little to report other than 133ms supposed response time, which is actually pretty good for 30fps games but what if 60 fps and this would perhaps be noticeable?

    STILL I really fancied Move to Wii Remote, as I m playing the Sky Crawler on Wii by Ace Combat team and the lack of 1 to 1 precision is very affective and enough to revert to classic controller instead.

  • pollux #18 2 years ago

    @TonyCB : this article is reported on eurogamer.net main page, it's very visible. Therefore, being a blog or not is irrelevant, it's just as if it originatedfrom the main site.

    Guys, the lag is under 1 frame (16.6ms), that's scientific, it's already been proven.
    The only thing we're seeing here is the lag of the application + the lag from the tv-set.

    Of course this article is so well made it doesn't explain this well.

    Oh yeah, Halo @30fps = 100ms lag, what's the framerate of this application btw?
    Edited by 1 at 17/03/10 @ 12:41

  • TonyCB #19 2 years ago

    @pollux

    I am sure many other pages will link to this article but it is hosted by and was conducted by DF. That was my point.

    Df apparently pride themselves on getting their figures absolutely right under strict lab conditions. Which makes this article a strange piece indeed

  • cyacomini #20 2 years ago

    "Unless you've played Quake on a 28.8k-56k modem against people on a T1 Uni"

    Yes, it was 1996 - happy days!

    [god that was 14 years ago - I'm old]

  • SpaceMidget75 Verified Senior Software Developer, Minerva Computer Services #21 2 years ago

    Well I for one was quite interested and happy to see a video of the current controller lag.

    If a bunch of you want to have a hissy fit about the 133ms figure at the end, be my guest.

    Thanks for filming it RL.

  • StooMonster #22 2 years ago

    Can't be arsed to look it up, how does this compare with Natal lag and Wii lag?

  • ronuds #23 2 years ago

    You guys will only be happy when DF names Sony the company of the millenium. The guy says very clearly it isn't a perfect test, so the whole thing should be taken with a grain of salt. DF doesn't have that much to report on, so they have to do something while they're in the office!

    They did the same thing with Natal, and I don't remember the comments for that article talking about Sony bias.

  • ronuds #24 2 years ago

    @ StooMonster

    If my recollection is correct, the latency was the same with Natal. Or Natal was at most 166ms.

  • Murton #25 2 years ago

    "You guys will only be happy when DF names Sony the company of the millenium."

    Not me, I only ask for impartial reporting. EG/DF seem to delight in partisanship these days and they even go as far as to deliberately bait readers into flame wars to drive hit counts. I know I'm not the only one who has noticed and commented on this recently.

    The first three paragraphs set the tone of this piece perfectly "Sony claim latency of under one frame" "the guys who did the PC driver for the camera have confirmed the claims" and "DF thinks otherwise and will now prove it" the rest of the article outlays their rather crude methodology in order to achieve that end, while they do say that their method here wasn't perfect they claim it is "good enough" what you're seeing here are people saying, "No, actually it isn't good enough. Try again when you're ready to do it properly"

  • TonyCB #26 2 years ago

    @Murton

    Totally agree with most of your comment. Have all systems - looking forward to Natal tremendously

    only thing I am not a fan of is Bad Science

  • StooMonster #27 2 years ago

    @TonyCB: actually, I am a big fan of Bad Science, but only the book by Ben Goldacre. ;)

  • Tiger_Walts #28 2 years ago

    Interesting to point out that 133ms is also the latency expected for the OnLive service, for users at the edge of the service area. So there are those that think it's an acceptable ammount.

    It's also very close to the input lag that Killzone 2 players experienced before it was patched. The response was a mix of players who adjusted and were fine with it and those who abhorred the sluggishness.

    Murton, to add to your points...

    1. True, the display is an unknown quantity but I think since this was set up for display we should assume that Sony chose a display with the smallest latency. A figure in the 15-20ms ballpark is expected for most displays of the type used.

    2. Again, this is the tech demo for display purposes but it is also performing the fundamentals of the technology and little more. If any software using Move is optimized, you'd expect it to be this be this. For now, let's give them the benefit of the doubt.

    3. The PS3 outputs the display at 60Hz, the camera they used captures at 60 frames per second. That means their measurements can only be as accurate as 1/60 of a second. Even if their camera was faster, they'd only be able to take measurements at points where the PS3 does a frame refreshanyway, which is also 1/60 of a second. For their method, their camera is accurate enough. Because of this, DF should have stated such an error margin.

    So, with the error margin and the display latency, a value of (with nice rounded figures) 105-135ms processing lag should have been suggested.

  • TonyCB #29 2 years ago

    @stoomonster

    :o)

    lets just hope DF dont start looking at our poos next mate.....
    Edited by 1 at 17/03/10 @ 14:11

  • Bigglesworth #30 2 years ago

    @ronuds
    Nobody's bothered about bias, other than the joker at #1. What rankles is yet another page of DF speculation and handwaving passed off as technical analysis.

  • ronuds #31 2 years ago

    @ Bigglesworth

    I understand, but is it really something worth getting so worked up over? You'd think DF stole people's children. These are video games here... not life changing events.
    Edited by 1 at 17/03/10 @ 14:46

  • dfish #32 2 years ago

    i think people haven't considered an important point here.
    133ms with a thumbstick is one thing 133ms with a motion capture device is another. The line drawn by the sticks moving from full deflection to full deflection is very small.(3cm?) the arc of movement drawn by a swinging hand is measured in feet. This means a larger distance can be traversed per unit time before the player starts to notice an onscreen response, this is going to lead to an increase in the perception of lag imo.
    though i accept the DF initial experiment is not definitive by any means
    Edited by 1 at 17/03/10 @ 14:52

  • Roarster #33 2 years ago

    I don't think there's any bias in what DF have said here, it's just that their methodology is completely flawed. He's measured the response time of a game and has labelled this "controller lag" which it isn't. Some of that amount is undoubtedly down to the system processing the controller input but there's no way of knowing how much.

    This is the same as concluding the controller lag of a dualshock is 133ms due to the response time of Killzone 2 which is clearly wrong.

    I can't help but think that DF don't really know what they're talking about here.
    Edited by 1 at 17/03/10 @ 14:53

  • HandOfBeadle #34 2 years ago

    pollux

    Where is the proof that it is a 1 frame latency? Not saying it aggressively - I'd like to see it.

  • pollux #35 2 years ago

    @Roarster: couldn't agree more.

    Then DF shouldn't talk about "playstation Move lag" but more about "augmented reality demo lag", see?

    Just with the title, people expect to see the lag of the device itself, you can see that on the reactions too, people are now confused. We already know the hardware lag is under a frame (16ms), so now it's up to the developers to fine tune their algorithms, might have to keep some stutter and so on ...

  • pollux #36 2 years ago

    @HandOfBeadle: 3 sources
    1) Sony announced this 1 frame lag
    2) PSEye was verified by pc users to be 1/60 sec lag
    3) A real-time debug demo at the GDC showed the lag (screenshot showing : "latency: 22ms";), this appears to show the real input information. I assume the control input is direct (wireless), just as fast as the DS3, and the latency would be the delta between the input modification (just by moving) and the camera (which would be a bit slower).

    Ok finally managed to find the picture on the internets, [link url=htt p://img696.imageshack.us/img696/9048/datam.jpg
    ]htt p://img696.imageshack.us/img696/9048/datam.jpg[/link] (granted it's not 1/60 sec, but more like 1/50)
    Edited by 4 at 17/03/10 @ 16:42

  • super_monty #37 2 years ago

    Anyone remember the controller lag on the speccy? You had to compress 2 cm of rubber to get a response 1 second later!

  • HandOfBeadle #38 2 years ago

    Pollux:

    1) Not really enough

    2) I've noticed lag when using the PSEye though. Not on my PC, but my PS3. It's not TV lag either - my TV is supposedly pretty good for latency.
    I'm not sure the camera itself will be the quickest part of this anyway - Sony said they were using all the other sensors for quick movements and the camera as a constant recalibrator to maintain precision over longer periods of time. That was the interview on here, I think?

    3) Now this is more like it - have you got links at all? Would be nice to see something more concrete than this article which,as has been pointed out by most, is attributing the overall latency to just the controller, rather than the sum of all the parts.
    Edited by 1 at 17/03/10 @ 16:35

  • pollux #39 2 years ago

    @HandOfBeadle:
    1) I'm sure they said this quite a few times!

    2) You're seeing the lag of the whole application, not just the PSEye! It's exactly the point of our discussion.
    Yep that's about how it works, PSEye is used to correct the drift, it's very responsive and accurate but it must see the sphere, when it doesn't, then only the internal sensors are being used, and they'll give an accurate position for a few seconds (like the wii).

    3) I tried to post the link

    4) Looks like "133 ms lag" quickly spread on internet, many websites thinking this is now the device lag, great job DF!!
    Edited by 1 at 17/03/10 @ 16:44

  • dfish #40 2 years ago

    i think the article makes clear the flaws in its methodology and is not trying to spread misinformation at all. its an honest appraisal derived from the tools at hand. its only those with an agenda to push who are taking the estimates as facts and spreading the word of the gospel about the place.

  • HandOfBeadle #41 2 years ago

    Pollux -

    I meant that Sony saying so isn't really scientific proof!
    Your link doesn't work straight away but I've fixed the http and it looks like sufficient proof to me..

  • man.the.king #42 2 years ago

    In my opinion, this article cannot be considered a scientific measurement, more like a layman's estimation of Move lag.

    That being said, I presume Rich must have to show some output to justify his salary (or hours charged, if he is contracted). You can't expect tech analysis subjects (which seems to be DF bread-and-butter) to pop up everyday. So RL goes with whatever topic is available.

  • des #43 2 years ago

    If this thing has worse lag than dualshock it will be total fail for core shooter games..nobody will care,competitive multiplayer players will hate it.

  • Josh128 #44 2 years ago

    Agree with the person that questioned that the entire image, not just the virtual sword, was obviously lagging. Now, it remains to be seen whether the video is purposely lagged by the PS3 hardware to sync with the known lag of the Move, or if its just lag caused by the camera feed itself. It in no way , shape, or form can be caused by the display itself. Why would anyone even think that? What kind of display lags >100ms behind its input signal?? Theres nothing on the market even close to that kind of input signal to display latency.
    Edited by 3 at 17/03/10 @ 21:44

  • dfish #45 2 years ago

    fwiw i think move will be the most responsive motion control device on the market whether it will be enough for the core is another thing.

  • Drpwnage #46 2 years ago

    Interested to see the rabid comments in this thread, especially when it seems that some people fail at reading comprehension. Richard couldn't have signposted the intent and limitations of the article much more clearly he even italicises what he was setting out to do (get some idea ) and proceeds to use caveats and indicate unknowns in his 'ball park' latency guess.

    I wonder if some of you read the same article I did, or just lept straight to the comments section after a cursory glance, frothing at the mouth.
    Edited by 2 at 17/03/10 @ 19:48

  • Jayke #47 2 years ago

    Ok, a few things
    First, the latency of the controller has nothing to do with the latency of the game. essentially, the controller responds exactly the same as the camera, that's why it sincs up nicely. This is like playing killzone and saying the ps3 controller has a latency of 150ms, thats the game, not the controller. Also, lat5ency is built into some games to add a sense of weight to the movement

    as for quantifying latency with motion controlsm its VERY different from button controls. having to avtually move yourself makes you almost completely unaware of game lag, because of physical movement lag. they slow it down to shw the lag, you know why? cause you cant tell when its normal speed.

  • ronuds #48 2 years ago

    @ Drpwnage

    Well, the crying SDF have reached levels of epic proportions. Every single DF or EG article related to the PS3 has a bunch of crying and whining fanboys in the comments section talking about "bias" and "teh DF is teh Sony hatererz!" I'm sure they're blind to this, though, and it's everyone elses fault.

    It couldn't be anymore akin to when a child doesn't get his or her way. Every article's comments section devolves into a discussion about how Sony isn't treated fairly because EG and DF aren't constantly splooging all over themselves to compliment them. And then they wonder why the comments section is considered a complete joke.

    Go on, negate away! Someone will still read it - and negating won't make it any less true.
    Edited by 1 at 17/03/10 @ 21:06

  • Josh128 #49 2 years ago

    Again, its interesting to note-- its not the virtual sword alone thats lagging, its the entire video display. Looking at the display alone, the sword moved 100% accurately with the image of the guy, with NO discernable lag /difference between the two. Say what you will, but I thought it was pretty impressive...

  • MeBrains #50 2 years ago

    the comments here are beyond belief. how the hell can you be against the article which you have all profoundly read? you crave for information and you blame the guy who gives it? what gives?

    RL / EG / DF has nothing to go with but a meagre Kodak #dunno. Sony's not giving out any press kits for the thing, yet the guy / site / subsite draws some conclusion of it. I think that is freeginn incredible and yet some here are being the Experts and telling what is wrong with the news?! Can you do any better or is it a classic story of "the best boatsmen stand ashore"?

    give the guy / the site / the sub some credit. He has nothing to run on and still manages to get a hint at what it all could mean. I'd say: respect...

    no?

  • KDR_11k #51 2 years ago

    As a bit of a Quake expert - I can tell you that 125ms lag (give or take!) is enough to mean the difference between being at the top of the leader board or sitting in the corner with a grenade up your ass.

    Remember that online lag behaves different from input lag, for one thing it's a lot less predictable. With input lag you see the exact lagged state of the game, with online lag you see a mixture of the state and some prediction mechanisms that mean if your crosshairs are on top of a dude the bullets still won't necessarily hit.

  • funkateer #52 2 years ago

    The article analyzed overall lag, not controller lag. (the title of the article notwithstanding)

  • Ivan_R #53 2 years ago

    Guys who did this test lack some elementary knowledge. You can't test lag in augmented reality app or anything that processes video and then outputs it.

    The real test would be something that ps3 renders in 3d without any acceleration or ingame items' weight added.

  • kangarootoo #54 2 years ago

    @Der_tolle_Emil

    "This is definitely no scientific proof at all but certainly a good measure of where things are going."

    Its only a good measure if it is a reliable test. There is no scientific theory being discovered here. All that is being discussed is controller lag. And so all that matters for the user in controller lag is the final number.

    If the demo being recorded for the atticle contains too many unknowns, the final number cannot be relied upon. And if the final number cannot be relied upon (jonbwfc's post puts it all very simply), there is no "good measure" of anything.


    based on this article, can we safely assume that lag in games will be less or more than observed in this test? This is of course a trick question, as the true answer is "we can't say either way, at all". A good measure still?

  • Bigglesworth #55 2 years ago

    Oh well, we managed to get in a page-worth of normal discussion before the ad hominems and logical fallacies arrived.

  • kangarootoo #56 2 years ago

    @ronuds

    "and negating won't make it any less true"

    Given that it is already untrue, you are quite correct.

  • TonyCB #57 2 years ago

    @bigglesworth

    being new to all this forum stuff (however not at all new to games or bad science), is it a general rule on eurogamer that the second page is for toys to leave the pram as the logical argument is lost? :)

  • Sunyavadin #58 2 years ago

    "Unless you've played Quake on a 28.8k-56k modem against people on a T1 Uni"

    Yes, it was 1996 - happy days!

    [god that was 14 years ago - I'm old]

    I was one of the guys on the university's T1 playing it.

    We had a whole room devoted to Quake.

    OMG I'm pushing 30...

  • Bigglesworth #59 2 years ago

    @TonyCB
    EG comments threads as a whole generally constitute a whole big region marked "Abandon hope all ye who enter here"...
    I'm sure you'll soon formulate your own rogue's gallery, though =)

  • Arwin #60 2 years ago

    (Correction). The article is saying that 133ms is likely to be the complete lag from the whole input, rendering and display pipeline, and does not say much about what the controller adds.

    What we are looking at here is this:

    1 The camera sends its picture data to the PS3 over the USB connector
    2 The wand sends movement and position data to the PS3 over the BlueTooth wireless signal (based on its magnetometer, gyroscope, accelerometers and whatnot.]

    (1) and (2) happen in paralel

    3 The video input from the USB camera is streamed to the SPE that analyses the signal for the Move controller's led
    4 The results of the analysis is used to enhance the precision of the signal from the Move controller

    (4) is extra, and is what Sony is talking about when it says it can typically do this within one frame, I'm fairly sure

    5 The video input is streamed on towards a framebuffer on the graphics memory
    6 The graphics engine receives the controller input from the SPE and uses it to determine the location of the sword
    7 The graphics engine renders the sword

    (5) is probably done in paralel to (6) and (7)

    (6) is the location where controller smoothing will typically be implemented, potentially adding some lag

    8 The sword and video input are combined into a framebuffer
    9 The framebuffer is switched to become the active displayed image on the screen

    How the lag that is purely added by the Move controller is experienced will eventually depend on the functionality that is used inside the Move controller and the time it takes for the camera to grab an image to the SPE returning positional data from it that can be combined with the Move controller's input (1),(3) en (4)

    Additional lag can happen in the smoothing fase (6)

    The minimum amount of lag from the Move controller is probably in the 22ms ballpark as suggested by the screenshot posted earlier on in the comments.

    How much this differs from, say, getting input from the analog stick of a dualshock3 controller I don't know off the top of my head, but there will be a difference, maybe 16ms or something similar?

    Also interesting to note is that the button inputs may not be influenced by the input of the camera. This depends on how the threading and data stream model of the driver works, and that is currently anyone's guess I think.
    Edited by 1 at 18/03/10 @ 14:30

  • Josh128 #61 2 years ago

    @ Arwin

    Good breakdown of how *likely* this process works. My best guess is that the Camera > USB >PS3 video feed is the cause of ALL the lag seen in this article, and basically has nothing to do with the controller. Anyone with a similar camera can to USB it to their PC (IF its in fact using USB and not firewire-- which could make all the difference in the world too) and play around with it and any lag should be noticable immediately.

  • Beano #62 2 years ago

    @des : Dual Shock lag?... are you drunk?

  • pollux #63 2 years ago

    So we all agree, this is not just the "controller lag", but the lag of all : controller + display + processing + ...

    That's why I originally said the title shouldn't be "PS Move lag" but "Augmented reality demo lag", which is a very nice piece of information, just it doesn't really mean anything, it's like I say: "Dual Shock 3 lag" and I use killzone 2 to say the DS3 gets 150ms!!!

  • pollux #64 2 years ago

    @Arwin: great analysis, but I agree with Josh: the lag is probably the difference between the camera and controller input times; the controller being bluetooth sends information just as quickly as the DS3 = close to 0ms, this includes sensors and button inputs. But the camera needs to do some processing+send images to the ps3, this takes more time and that's probably the 22ms figure.

    Uhm that means that to combine both picture and controller, the controller must be delayed by that 22ms to be completely in sync. However in a real game we could gain an extra 22ms on the lag by not waiting for the camera and only use its input for the next frame, see? :)

    Really it's in the hands of the developers now. With a pad you don't really notice lag because there isn't any 1:1 mapping visible on screen, so your brain adapts to it, here it's more visible, so any lag cutting solution will be welcome:
    - high framerate, or variable with triple buffering (like gow3)
    - just less buffering, more direct processing, less input smoothing (or intelligent one, like only do it when there is slow motion to reduce stuttering, otherwise keep it)
    - don't wait for the camera
    - extrapolation (= prediction) can help too, can be improved by using properties of the human body

    Basically a lot of work, but it'll pay off!

  • Bigglesworth #65 2 years ago

    Firstly, credit to Richard for taking time to respond to us internet troublemakers.

    Following the link trail back to the Gamasutra article that provided some of the baseline figures shows that this whole subject is still far too woolly. Unfortunately, judging response to a button press to be the first visible effect of an on-screen animation tells you little about the actual input response, even leaving aside the whole 'zero-frame' problem.

    What seems clearer though is that lag from the display hardware is a primary factor here, followed closely by software framerate and frame buffering method.

  • Arwin #66 2 years ago

    Yes, pollux I agree that Josh has a point. USB cameras do tend to have some lag. However, testing the PS Eye myself at home, although I don't have precise measurement equipment right now, it can't be a lot especially if my TV (a HD Ready Samsung 32" in its fourth year) also adds some lag of its own.

    But presuming that the USB camera is responsible for most of the lag, then every Augmented Reality game for Move that we've seen is going to have lag similar to the 133ms that we have seen. The Move input will then in fact be delayed somewhat simply to match up what the GPU renders (the sword) with the camera input.

    It may not matter too much - if we look at what the best response time in most games currently seems to be, then its 66ms for a 60fps title like Burnout and 50ms for the XMB, as the DF article mentions. Assuming that we also reserve about 22ms for what Move adds in total (combining the input from the Move's 3-4 completely different internal sensors to produce a coherent stream of input data, and in paralel receiving and processing the camera's image to enhance precision of the controller data), that leaves about 40ms. We already know more or less for certain that step (6) is implemented in most games, to smooth out the input from the controller to prevent jittering from your hand never being completely stable, and if a game triple-buffers we also know this adds some.

    So an augmented reality game could well end up being 133ms typically. Which is fine, as it is still just within that magical 150ms barrier.

    What will be interesting to see is how games that use pointer functions stack up. If you look at the tech-demoes from Sony's hardware developers, there's almost no lag to be seen at all particularly in those demoes that don't involve augmented reality. These can be doctored/resynched of course, but it may also very well be that 1:1 tracking in 3D space can still happen in 66ms "controller-to-display" time.

  • prettyboytim #67 2 years ago

    But what's the minimum lag on a standard controller? I estimate the following bast-case scenario for a 60fps game:

    0 - 8ms: Dualshock 3 detects user input and transmits it to PS3
    8 - 24ms: PS3 reacts to input and renders reaction to backbuffer
    24 - 40ms: PS3 swaps backbuffer to front to display and transmits to the TV
    73ms: TV displays the image.

    For 30fps game, it'd be something like

    0 - 8ms: Dualshock 3 detects user input and transmits it to PS3
    8 - 42ms: PS3 reacts to input and renders reaction to backbuffer
    43 - 75ms: PS3 swaps backbuffer to front to display and transmits to the TV
    103ms: TV displays the image.

    I'm assuming an input lag to the LCD of 33ms.

    I think these are pretty close to best cases - if you've got the render happening the frame after the simulation, you've got another frame of lag, and you may get almost a frame of lag if the display is triple buffered. For a 60fps game, that could push the actual input lag up to 100ms or so and 160ms at 30fps.

    To me then, 133ms really isn't a bad input lag at all, and really probably about the same as many 30fps games.
    Edited by 1 at 18/03/10 @ 22:12

  • smelly #68 2 years ago

    Yet another article where DF shows their lack of understanding - while trying to sound clued up.

    meh

  • Arwin #69 2 years ago

    @prettyboy: DF has already mentioned that the minimum lag measured so far is 50ms on the XMB, and 66ms in Burnout Paradise. These are indeed both then running at 60fps.30fps adds more lag, and so does triple-buffering.

    And no, 133ms definitely isn't bad for a device like this. I doubt very much that WiiMotion+ and Natal will be able to match it, let alone beat it.

  • Transcendent #70 2 years ago

    133ms holy fucking shit. I'd go back to NES controllers before that.

  • vizzini #71 2 years ago

    If Sony (with over half a decade or R&D in electronics) has said it is responsive, then I'll take their word for it; and try it my myself to do my own DF. Same for Natal; if Microsoft state the accuracy level of the controller and state the system is responsive I'll trust them.

    Why? well unlike DF, when they make a definitive technical statement it is legally reqired to be true. Advertising laws are reasonable good the worldover in combating miss-selling of goods.

    As always in this world, it is far easier to tear other people's good things down than build good things up, yourself. The article feels like they were itching to undermine the new motion system, even after Sony's statement had already made their job redundant.

  • drxym #72 2 years ago

    It would have been pretty stupid if Sony had forgotten to set their TVs to game mode but stranger things have happened. On my TV, not enabling game mode adds 3-4 frames of latency which account for most of the delay reported just by itself.

    Aside from that, there are several obvious ways latency may have been introduced:

    1. Sample smoothing. People don't hold their hands steady and the game might be averaging out several samples to dampen the jiggling of the controller. That makes the capture steadier but at the expense of lag.
    2. Sample rate. The PS Eye camera can capture at 60fps but that doesn't mean the motion capture software is processing all 60 frames a second. Some games may well discard every other frame, or even more for performance reasons. If its discarding frames, it could add latency.
    3. Unoptimized code. These games and the motion capture software have at least 6 months of development and QA testing ahead of them.

    Aside from the camera aspect, I doubt the actual technical gubbins in the controller itself introduce any latency at all. I imagine that many games will mostly rely on dead reckoning to ensure responsiveness and will only periodically (or possibly never) use the camera to sync the user's position. Twitchy games like Socom 4 probably don't need any positional tracking because the user will be making small movements and not waving their arms arcs or whatnot.
    Edited by 1 at 19/03/10 @ 13:38

  • HandOfBeadle #73 2 years ago

    Vizz

    Ask about the accuracy of official TV/Monitor contrast ratios and response times on an AV forum.

  • vizzini #74 2 years ago

    @HandOfBeadle
    “Ask about the accuracy of official TV/Monitor contrast ratios and response times on an AV forum.”

    Maybe you could review the performance of my last 2 flagship screens (in game mode), to tell me they had problems I know first hand they don't.

    On a student placement a decade ago, I had world leading expert work colleagues, who actually create pro A/V technology, so naturally I'm unimpressed by so-called expert opinion.
    Edited by 1 at 19/03/10 @ 18:52

  • TheNinkyNonk #75 2 years ago

    "I'd say that a ballpark figure of around 133ms of controller seems reasonable"

    Dude. Get. Out. More.

  • HandOfBeadle #76 2 years ago

    @Vizz

    "@HandOfBeadle
    “Ask about the accuracy of official TV/Monitor contrast ratios and response times on an AV forum.”

    Maybe you could review the performance of my last 2 flagship screens (in game mode), to tell me they had problems I know first hand they don't.

    On a student placement a decade ago, I had world leading expert work colleagues, who actually create pro A/V technology, so naturally I'm unimpressed by so-called expert opinion. "

    What does any of that have to do with the price of chips?

    I was pointing out that, when contrast ratios or screen response times are quoted by LCD screen manufacturers, they can do so in a number of different ways, some of which give the impression that one model of TV outperforms another for depth of colour or suitability for fast moving images. It's hardly a new concept, and not one exclusive even to the electronics industry.

  • Xerx3s #77 2 years ago

    "If Sony (with over half a decade or R&D in electronics) has said it is responsive, then I'll take their word for it"

    You mean the same sony that has everything to win or lose with this product and who has lied to the consumer time after time again? Pardon me for being extremely sceptical towards any new product until it is actually proven to work.

  • vizzini #78 2 years ago

    @HandOfBeadle
    “What does any of that have to do with the price of chips?”

    It was a correct response, to the poor, condescending advice you initial tried giving about TV info.

    The statement by Sony was quite simple, no more than 1 or 2 frames of lag from the controller and as responsive as the Dualshock3 (In full HD 1frame ~=17msec). This is something which they must be able to easily prove and have confirmed by other developers. These technical statements by Sony have nothing to do with the TVs they used or DF's misguided experiment.

  • HandOfBeadle #79 2 years ago

    Vizzini

    Going for the dying alone option, eh? Fair enough.

    I'm not even disagreeing with you that the controller is extremely responsive. I just found your initial post crushingly naive.

  • Arwin #80 2 years ago

    Yeah, they confirm the 22ms figure again, noting that the rest of the lag is outside of their control, caused by game rendering/display and TV lag.So I think my analysis stands pretty much as is. The camera control thing looks fun and like something that's probably going to be used a bunch.

    And it's also something that could be easily patched into existing games - some already have some form of camera control. I think Uncharted's cinema features could be a lot of fun with this.

    And of course the multi-touch stuff was fun too.

    I think it's interesting to note comments like 'I don't want motion stuff in my games'. People used to think the same about the analog stick. Nintendo came up with it for the 64, and all Sony basically did was put two of them on the Dualshock. Not that big a difference on the surface of things, but these days it's all 'hardcore' console gamers seem to want. Will be interesting to see if this type of technology is the same and move is perfecting what Nintendo started, or whether things will go in a different direction altogether.

    I do wonder though if the 'subcontroller' is really necessary, or if we're not just better off with two Moves for everything.
    Edited by 2 at 21/03/10 @ 20:26

  • HandOfBeadle #81 2 years ago

    Personally I don't see how you could move,say, an FPS character using a Move to replace the left stick. Certainly not in a way that improves on the existing system. Any suggestions?

  • TaoJay #82 2 years ago

    from NeoGAF: After show demo of Move party & some chat with Anton.

    http://ww w.engadget.com/2010/03/23/plays...

  • denialmaccullum #83 2 years ago

    Looking at the list of games back here, I can not change my preorder from Shoot Time Crisis is probably more fun. And I would not mind trying to play Socom, I never tried, but that series. PC games have always been the more I think about a game on the PS3 if the fps is used to move and works well as I hope it does. I still do not have to be very good analog stick lol.

    Package holidays Gumbet