R190 - jerking on IE, Chrome, Firefox, Opera

0 favourites
  • IE 11 is "jerking/lagging" the same as chrome.

    Steely - I haven't noticed this myself with IE 11. (Windows 8.1 64 bit).....

    Ashley, So is this a Chrome specific bug only, or an underlying problem with browsers\html5, or what.....?

    Anybody else noticed this on IE?

  • have win7 64 bit, but the 195 release runs very smooth, ie, chrome and node-webkit so I'm very happy.

  • Here is a related issue I have in another thread,

    viewtopic.php?f=151&t=122834

    It is likely to be related. Here is the captured movie of the platform starting to jitter, when pixel rounding is turned on in the second part of the movie (also repeated in slow motion).

    https://dl.dropboxusercontent.com/u/297 ... truct2.rar

    Ashley Please take a look, this has been ever since I use Construct 2, around v.162. Now using v195, still there, always has been. In Chrome, Firefox, IE, yes, also v 11. Web-Nod does not seem to help either. Sometimes some stuff seems to help, to get it less disturbing, but it never got perfect. (Also, I have no idea, what exactly makes it occur less <img src="{SMILIES_PATH}/icon_e_sad.gif" alt=":(" title="Sad"> )

  • Chrome 40 just went stable. Does this improve v-sync for anyone affected by jank?

    Even if not then I think Google have a v-sync patch in for Chrome 41.

  • No discernible improvement, according to my test capx. Here's a screen shot result from Chrome . As you can see, over a 120 sec period, there were a few hits of vsync values up to nearly 25 ms... Chrome v38 was arguably better....

    Latest stable Version 40.0.2214.91 m:

    [attachment=1:b4knyzzf][/attachment:b4knyzzf]

    A screen shot from version 38:

    [attachment=0:b4knyzzf][/attachment:b4knyzzf]

  • Chrome 40 just went stable. Does this improve v-sync for anyone affected by jank?

    Even if not then I think Google have a v-sync patch in for Chrome 41.

    No more jank for me in Chorme 40, even with Aero off.

  • Colludium - how about Chrome Canary which ought to have the v-sync fix?

  • Ashley, Canary I tested last night as well - it was no different to the current release (perhaps slightly worse, but it's Canary...). Those 25ms vsync his all happened in the first 20 sec of running the test as well.

  • Ashley, here's the image from the very latest Canary - updated this morning! I'm guessing that spike is at around 20 ms vsync...

  • Colludium - can you put this info on the official bug report at https://code.google.com/p/chromium/issues/detail?id=422000? Your data on the latest chrome canary is probably useful to the Google engineers, and shows there's still an issue. (As with all bug reports remember to cover the exact Canary version and your basic system spec).

  • Hello, once again, since I had no reaction at all to my previous post:

    I have jittering issue aswell, but related to the platform behavior. PLEASE TAKE A LOOK, IT MIGHT BE RELATED, possibly providing some hints to the cause of the jittering/jerking-problem.

    I have experienced this since I use Construct 168, and it has been always there. (Eventually, I manage to get a smoother platform movement, but I don't know exactly how... also never 100% smooth, which is annoying.) In this example case, it is just a plain platform behavior inserted, and pixel rounding turned on. (Please note: the cam-scroll is not important, it also happens to the platform object, when the cam is not following it.)

    http://www.exitthegame.eu/test

    I have tested it on different PCs and Macs, all with the same jittering result, in Chrome, IE (11), FF, Safari. This also happens, when exported via Web-Nod.

    Oddly, sometimes the shaking/jittering is horizontal, sometimes vertical, but always there.

    Do you experience a shake there? If not, here is a video of what I mean: https://dl.dropboxusercontent.com/u/297 ... truct2.rar

    Also the file test.capx is attached below.

    One more thing to add: when I add PlatformSprite > Set position to (round(self.X, self.Y), the platform object is always snapping one pixel down and right, when I press any platform controlling buttons. When I release them, the platform snaps back 1 px to -x and -y. You would expect it not to do that at all, since rounding of X should not be influenced by moving only Vertically and so on. See test-rounding.capx This is important, since the same effect might cause the Platform shaking in the other example, since pixel rounding is on (meaning that the displayed sprite is at a rounded position, whereas in my example I also round the object itself to a rounded position, which seems to soften the shaking, but still not working properly, as described.)

    Since the effect seems to be very similar, I think it could be very related...

  • Just a quick question:

    Is this discussion about the hideous vertical/horizontal jerks I see (on two different hardware-accelerated PCs with different brand everything) when playing a game with fast moving graphics?

    I recently developed a physics game that uses an object that is meant to dodge walls and such. As these walls move by in a horizontal direction, I have noticed a "tearing" that, while it doesn't effect every single tile, it creates a pattern of tearing.

    Like I noted earlier, this visual effect is obvious, and effects my game on both my PCs. On the one with an even more powerful everything, on a larger resolution it is nearly unplayable.

    I have yet to export or try more than Chrome as a test subject, however, it is of note that when I use Chrome, it is either ugly at best and hideous at worst.

    Is this what we are discussing? In my game in particular it looks unrelated to the game itself and more like a display error because the whole screen is never affected at one time. Just part of it.

  • Ashley - they were my thoughts too, so I added a comment right after posing the above. I'm in the notice distribution list as well and that bug report has been quite quiet for a while...

  • Try Construct 3

    Develop games in your browser. Powerful, performant & highly capable.

    Try Now Construct 3 users don't see these ads
  • MultipleChoice - I don't think your issue is related to this thread. This thread is basically about the v-sync quality of browsers, whereas your issue looks like it's to do with rounding in the C2 engine. You already posted a report for it here so it will be handled from there instead of this thread.

  • Ashley : Thanks.

Jump to:
Active Users
There are 1 visitors browsing this topic (0 users and 1 guests)