About the jerkiness on the movement...

  • Thanks Tylermon! Fingers crossed...

  • Colludium

    Yeah, thanks for posting the bug reports, they look great.

  • Just want to say great job with those graphs, they've turned up surprising data and it's very valuable to include that type of thing in bug reports since it points to an objective thing which can be improved in a measurable way. Hopefully this test helps browser makers fix their vsyncs!

  • Colludium , I tried the tests without overclocking. Still the same results on Chrome and Canary.

    Tylermon , at first I didn't get what to look in your test so I overlooked the results. Having read more since then, I understand that there is a connection between the results I get from Colludium's test and your test. IE and Firefox provide very random float numbers at higher decimal places and the only difference between them is the fps fluctuations: IE stays close to 60 pfs and dt numbers that are close to 0.016 (in respect to all orders of magnitude) while Firefox fluctuates greatly in fps from as low as 45 to 60 or so (and the dt values fluctuate also, but still have random values in all decimal places).

    In contrast, Chrome (and Canary) stay rock stable framerate wise but never leave integer values (dt wise) or when they do (in a big order of magnitude) they produce ridiculous floats, like 160.0000000012, or 159.0000000053, or 170.0000000023 etc...

    Here are some pictures that capture this behavior in all scaling levels.

    Also, your rig is in deed a beast, it should render anything 2d while been turned off

    TiAm , I have the latest nVidia drivers installed. Last time I checked I was able to run Crysis 3 at 1980x1080 with High settings (except from motion blur, I hate motion blur) with no noticeable fps drops, hangs, stutters etc. My only problem was that I didn't like the game at all... |

    I'll go and add my signature of frustration to Colludium's reports now

  • Colludium - thanks and good work--I hope it'll get something done! I commented on the bug report.

  • Try Construct 3

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

    Try Now Construct 3 users don't see these ads
  • Happy to do a bit to help the community - I'm glad my geeking proved worthwhile! I'm happy to report that that both bug reports have generated some interest - and rightly so: if a browser can render to 144 Hz then it should be able to do 60 Hz without any frame drops whatsoever. Just saying, Mr Google and Mr Mozilla....

  • BTW everybody: the Firefox bug report is getting some action too. Firefox is arguably the most problematic of the lot, so we ought to keep on them about this, and provide whatever data we can.

    Specifically, they are requesting affected users post the results of their about:support page. If firefox is working badly for you, especially compared to other browsers, bring up a new tab in firefox, paste in:

    about:support

    When this page comes up, click "Copy text to clipboard", and post your result on this bug report:

    https://bugzilla.mozilla.org/show_bug.cgi?id=1093454

  • I can only think of a couple possibilities.

    1. Motherboard bios. Are they up to date?

    2. CPU and/or memory issue. Potentially corrected by better bios.

    3. It could be display drivers/your screen(s) If you have a different screen to try that would be interesting.

  • Tylermon , my motherboard bios is as up-to-date as they can be. It's an old motherboard after all If this is actually related with how old the developer's hardware is, then I'm afraid that for the moment I'm in trouble! However, I would expect the performance of a system to be tied with the actual specs, and not by it's age. But of course, you could be right and for the time been I can't confirm/disprove your theory!

    I tried all possible monitor combinations (disabling them and trying running the tests on each one separately) and the integer values problem persisted every time. What was indeed improved however, is the jerkiness, especially in IE in which was practically non existent. You have a point there. However, I find working only on one monitor to be very counter intuitive and I believe that this shouldn't stay as a "solution". This should also be addressed somehow.

  • So, just to be clear, to fix this we are depending on browser makers? Has there been a response from them and does it seem like it's a priority? The jerkiness is really a deal breaker for me even though I am a huge fan of Construct.

    I just played the beta for a Stencyl game called Ghost song. It runs incredibly smooth with a res of at least 720p if not higher. I really don't want to learn a new program, so if this seems like it will be fixed in a reasonable time, I'd be happy to stick with Construct.

  • UberLou,

    I hear you. I've been watching both of the bug reports and the problem has been acknowledged by Mozilla and Google. Mozilla devs think that 'project silk' will rectify this problem for Firefox - time will tell. The Google devs are looking into what's become broken in the later releases of Chrome - it is manifesting as an inconsistent adherence to vsync which causes the dt variations talked about in this thread and some other issues (the bug report makes a good read). I'm confident that this will get fixed, especially for Chrome. When I lose heart I export to Node Webkit, which is based on an older version of the chromium engine that was not broken - it's as smooth as a smooth thing... Chrome Canary is pretty darn good as well (but not perfect...).

  • Unfortunately, the initial fix to Canary (lock to system v-sync on windows) has been reverted.

    Performance currently seems inconsistent...when I first tested a couple projects today, it was halting and jerking worse than ever...almost as bad as firefox. Now, after running a couple more times, performance seems pretty good. I suppose, being alpha, that's to be expected.

    UberLou

    The way I look at it is this: many of the movers and shakers in tech are moving toward 'run anywhere' applications, cloud apps are becoming more and more common, and the web is getting visually richer all the time. The issues we're dealing with may have been peripheral in the past, but not anymore.

    Browser makers are becoming very focused on performance, and as such, I think issues like these will continue to take on increasing priority. In respect to 'jerky' movement, IE is currently kicking Chrome and FF in their metaphorical arse. That won't last.

    tl;dr: Project Silk was not a response to a few C2 devs complaining that their 'Flappy Jack' epic was not up to speed.

    Even talented teams working with high-budget engines have to deal with glitches and bugs...seen the recent coverage on Assassin's Creed?

  • When I lose heart I export to Node Webkit, which is based on an older version of the chromium engine that was not broken - it's as smooth as a smooth thing...

    thanks for this, I'm trying to make a video of my game and the screen stutters are so incredibly annoying. NW runs so much better!

  • Colludium

    Node-Webkit doesn't fix the issue. Putting a simple 1 frame sprite with platform behavior and running back and forth causes random jitter. Tested on multiple PCs.

    TiAm

    We differ from the Assassins Creed dev because they chose not to fix the issues. We have to wait for a possible fix from a third party.

    I'll wait and see what happens, but I'll be checking out Stencyl in the meantime. It's tough to wait for a fix that might never come while another engine runs high res games perfectly right now.

  • Oh wow, the lastest node webkit update (coming with C2 r188 beta) = Hiccups everywhere, ugly ugly ugly! T__T

    Both in preview or exported mode.

    Rolling back to r187 , but how do I get the previous NW build? (I don't know the name of the NW previous build, which was working nicely for my game)

    http://www.scirra.com/node-webkit only provides the lastest one : )

    Thanks!

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