digitalsoapbox's Forum Posts

  • Proof positive that native is significantly faster, even when it has to deal with interpreting data created for another (non-native) runtime.

    Thanks for starting this thread — Hopefully this helps cancel out any future related arguments before they begin.

  • I only see a momentary pause on my dev machine (which is high end) - not sure why it can't loop seamlessly though. Could you test r256 as well? It's always useful to know the specific release that started exhibiting an issue.

    I'm on a fairly powerful machine as well (i7-7700k 4.2GHz, 64GB RAM) so I also find the pause surprising.

    In r256, I'm seeing a SLIGHT pause in the longer/larger video, but nothing as severe as r257.3. Both r255 & r256 pause for about half a second, r257.3 for far longer, with the test .capx above.

    It's possible that it's due to the compression settings the videos use, but won't have time to test other settings until this evening EST, as .ogv & .mp4 are exhibiting much less of a delay than .webm, and the .webm file is much larger. Will update here with more info then.

    UPDATE: It's looking like the issue may be due to using large, long video files. I don't have time to update the .capx right now but you can probably close this out as a minor issue; re-exporting the longer video with different, lower-quality (but not significantly lower visually) settings reduces the delay quite a bit, though there is still a slight delay when it loops that causes a visual stutter throughout all on-screen elements.

    An interesting side effect of the initial delay before the video starts playing is that it also prevents streaming audio from starting until the video has started.

  • Are you sure this does not reproduce in r256? The video changes in r257 should only affect how video starts playing, not any aspect of playback beyond that. The browser handles the actual playback including the loop-around so there might not be much we can do about this. Also it would be helpful if you can provide a shorter video since it will make it quicker to test any changes.

    I don't recall trying it in r256, I was previously on r255. As stated above, the video also takes awhile to start the first time through in projects with other things in the layout - a typical scenario on a menu screen, with the video as the background, some spritefont objects, keyboard/gamepad input, streaming audio playing.

    The video is longer to illustrate the issue more clearly - a shorter video does not pause for as long, and depending on length, restarts without pause. I've updated the .capx above so that if you press the spacebar, a shorter video clip will load. I've also added in a streaming audio clip in case that is having some effect on the video.

  • Problem Description

    Looping video freezes on last frame for a LONG period of time before restarting.

    Attach a Capx

    https://www.dropbox.com/s/bu3fwrw6fjpcz ... .capx?dl=0

    Description of Capx

    Plays looping video

    Steps to Reproduce Bug

    • Run Project
    • Wait for video to reach end
    • See how long it takes to restarted

    Observed Result

    10+ second delay to restart video, meanwhile the video is stuck on the last frame. In larger projects with streaming audio and/or player interaction/input, this delay also occurs before the video starts playing the first time.

    Expected Result

    pre-r257 behavior where looping pause was very short/unnoticable, no/short delay for videos to start playing the first time.

    Affected Browsers

    • NWjs: YES
    • Chrome: YES
    • FireFox: YES
    • EDGE: YES

    Operating System and Service Pack

    Win10 Enterprise 1803 17134.1

    Construct 2 Version ID

    r257.3

  • Hi,

    I am looking to license some games. If you've created a killer game and you're interested in licensing it please drop me a line with the link and how much you want.

    I'm only looking for well put together, engaging, bug free games, not simple games like 'here's a physics test' etc.

    Thanks

    Mark

    I'm sure spamming the same message in multiple forums will get you exactly what you're looking for.

    Kyatric, want to deal w/ this?

  • Try Construct 3

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

    Try Now Construct 3 users don't see these ads

  • I used C2 to make a simple RSS news feed as a test to run on a Pi 3, which works fine, since because it's text it doesn't need a steady 60fps - which is our only realistic target since we can't lock C2 to 30fps with predictable or reliable results, nor should anyone really consider 30fps a target with simple 2D projects or games - but you don't really need C2 to do something like that either.

    If you're doing something basic, like an endless runner, it MIGHT perform okay on Pi 3. But even a simple retro arcade-style game like a Pacman clone is going to have some performance issues. The Pi hardware isn't really built for HTML5 games, but considering its low cost it definitely can't hurt to give it a try on your own.

  • Hi guys, any update about this? I was hoping to buy a raspberry just for run my games but I would like to know what's the best export to use for better performance after one year of updates.

    Thank you

    Performance on Pi 3 is still subpar. The GPU seems capable - it'll run Quake 3 just fine at 720p - but WebGL support is questionable and the CPU just isn't very good at running any heavy HTML5/JS content (games, for example) created with C2 at this time. Could be drivers, could be the low-power CPU, could be the way C2 does things.

    My takeaway from this thread is that Scirra is upset they can't control the discussion off-site, have worked themselves into a hole in terms of community trust, and are unhappy that a clearly unofficial forum for discussion (seriously, it's said it's unofficial since day 1) provides a better and more welcoming outlet for open and honest discussion of known issues than the official forums, with its vague and draconian rules designed to shut down discussion Scirra doesn't like, where it's much easier for them to shut down such discussion, or sweep it under the rug.

    This thread provides a very clear and compelling reason to keep the Discord unofficial and independent, as well as illustrate why it's great resource for Construct users interested in helping each other and solving issues without the constant spin or misdirection in the process of doing so.

  • Hey, thank you to all who replied!

    digitalsoapbox, thank you for the length reply, I really do appreciate it! I was kind of in a bad mood when I wrote my rant, and I might have changed some opinions from that time

    Yes, I am talking about the new DOOM! It' was made by only a handful of engineers? Wow, that's amazing!! Did you see this article talking about it :http://www.adriancourreges.com/blog/2016/09/09/doom-2016-graphics-study/

    So many things happen, it's amazing how they designed the engine!

    And yes, thanks for the points you have given about performance, and about the relationship between C2 and C3. To be perfectly honest, I'm kind of disappointed that Scirra knows that the Android export module is problematic for C2, and they claim that it's fixed in C3 by doing the conversion on their servers, and that you can convert C2 games with this method, but in order to access it you need to subscribe to C3!! I mean, I know it's a business, but if you know that you have a problem with your exporter, wouldn't you want to offer a solution to C2 owners too? If it couldn't be done (no resources, not worth it), maybe I could understand, but not like this. Perhaps they could charge a small fee to convert C2 games.. But heck, what do I know? I think it's bad PR for people to know that your engine has a faulty exporter, but you have a solution, but you keep it behind a paywall..

    Again, I was in a bad mood when I wrote the initial post, and I probably should have taken heed of what others have said, mainly because they are C2 users and have had issues with the software, versus me, someone who never worked with it. So, if I've offended someone, sorry!

    Thank you again to everyone who replied!!

    No worries, I'm pretty hard to offend, and I'm with you on disappointment in C3 as well as Scirra's more recent behavior connected to it and fixing software-breaking bugs in C2. Their behavior points to things maybe not going so well for them right now and not having the experience to know how to deal with it in the manner we all expect from a business.

    As for the behind-the-scenes on the newer DOOM - yep, read it awhile back. It's pretty amazing to see that they can apply the kind of techniques developed for film and TV visual effects to a realtime game with performance as impressive as they've been able to achieve. It's the kind of stuff that would take hours upon hours to render out just a few years ago, and far above and beyond what even $500,000 SGIs were capable of back when I started working with 3D professionally all those years ago. Even on lower-end hardware or Intel embedded GPUs, their work will run rings out, say, the work of people developing a far more basic 2D game engine with developers who continuously blame the hardware for their tool's shortcomings. But I guess hardware is...third party.

  • Can you develop a functional game on the nintendo switch by building 2? Thanks

    Nope.

  • First of all I would like to say that I think that Construct 2 is pretty damn awesome from what I've played with it so far! It's very intuitive and it makes the process of making games very fun!

    I am thinking of buying Construct 2 but I have some questions (I apologize if they were previously answered in a forum thread or a news comment, I unfortunately couldn't find the information I was looking for):

    a) will Construct 2 be supported after Construct 3 is released? Technically I know that Construct 3 is released and Construct 2 is obviously still being updated, but I don't exactly follow the technology about wrappers for JavaScript, but from what I understand some get deprecated (this was done for Cocoon.io, I believe), and I would like to know if I can use Construct 2 to export games, even if support would be dropped for it. Sorry if this seems like a stupid question.

    b) this is a mini rant: from what I could find, performance sometimes can be troubling for games (mobile games in particular). The thing is, I don't really trust most users with their expectations and with the scoping of their projects. For instance, I once saw that someone was having huge framerate issues, but the problem was that he was emitting thousands of particles every second, which of course would slow down almost any engine (not trying to be mean to that user, I know that we're all learning here, sorry that I singled you out).

    On this performance subject I admit that I am having some reservations. The thing is that I know too well how some software/ game developers want to seem hardcore and brag that they learn other engines (now the craze is with Godot, you're especially cool if you work in Godot -- not saying it isn't a great engine) or that they believe that every single thing should be coded in C++. I saw that some users threw around a test about jankiness and another user was comparing freaking DOOM with a game engine. I think that's downright unfair and to be frank disingenous and mean, considering DOOM is a single game that was optimized by a horde of engineers, arguably one the best hordes in the business (iD Software, come on!), and while I believe that the engineers that made Construct 2 are especially talented, you can't really beat ONE game that was optimized to perfection versus a game engine. I don't understand how the developers/ people who work at Scirra can have so much patience with some of the accusations some people throw at the engine, but damn if they don't deserve a medal.

    When I was looking at Unity people kept saying that it's kind of overkill for smallish 2D games, Gamemaker Studio also is reported to having issues when porting to mobile and HTML5... I'm no expert, but is it safe to say that it's a matter of skill and imagination when it comes to performance? I just saw that Environmental Station Alpha was done in Multimedia Fusion, not even Clickteam Fusion 2.5, and it's an amazing game! Construct 2 also has great games built with it, and the ones on the Showcase page are especially impressive!

    The question I am asking, as I am stuck in an analysis paralysis loop, can I reliably make Construct 2 games for browser games and mobiles, maybe desktop, too, without having huge problems with performance?

    Sorry if my questions seem stupid, I'm not that knowledgeable in game engines, I tried Unity but that seemed kind of too much for the type of 2D games I want to make and I think that Construct 2 fits better for my skills.

    Thanks in advance!

    A couple things:

    a) C2 is being sold with software-breaking bugs. This is discussed elsewhere on the forum, feel free to take a look.

    b) C2 has a lot of performance issues, on desktop and elsewhere, when you try to do anything visually complex or expect things to behave the same way every time (for example, issues with jump predictability using the platform behavior in C2, that Scirra has said they will not backport fixes for from C3). This is not a developer or hardware limitation, regardless of earlier comments in this thread - it's the way C2 handles rendering, which is very inefficient, and little improved in C3 at this time. Promises of a new runtime can't be counted upon for commercial releases, especially when the switch from C2 to C3 has been so fraught with issues from a user perspective and has alienated many experienced users. It's why most of the more successful developers have moved on to other engines, and why some of those developers have ported their games originally made in C2 to other engines, including at least one of the games listed on the Showcase page.

    Do you mean new DOOM or old DOOM? Because old DOOM's engine (and it was developed as an engine, not just for a single game) was developed mostly by a single software engineer, and new DOOM's engine was developed by only a handful of software engineers - probably a smaller team than you think, considering its complexity, but they were smart on how they built upon previous existing tech. It's a perfect example of what's possible when a game engine is designed to be easily ported to a large variety of platforms as a native application by extremely skilled developers. C2/C3 can't be wrapped as a "native" app, not sure how anyone thought that was true, so that's something to keep in mind should you be interested in porting to widely popular platforms like XBox One where performance can be, to put it politely, somewhat disappointing (not to mention the lack of required features to do a release outside of the XB1 indie program, and even then IDvkt@XBox requirements have yet to be addressed).

    It's not your imagination that there are some issues with HTML5 performance, regardless of engine. There are, however, less issues with HTML5 using certain development tools over others. Phaser is a good example of a high-performance HTML5 engine geared towards the kind of quality and features that developers expect an HTML5 game engine - or, really, any modern game engine - to have.

    Unity has done a lot these last few years to improve their 2D toolset. I wouldn't recommend ruling it out as an option. Unlike some other subscription-based engines, they are very good at justifying their subscription cost.

    If you're looking to do 2D games on mostly desktop, C2 is probably okay if you keep things very simple graphically (C3 is still a mess in its current state and not worth the cost - it's early beta at best). Mobile performance is iffy for anything past very, very simple games, and the same is true for web-based games. Generally speaking I'd suggest using it ONLY for prototypes and porting to a more capable/feature-full engine for a commercial release or for very, very simple games.

    Well, it at least means it's not due to a Construct 2 update, and nobody else noticed it until late 2017, so again it points towards some kind of Windows Update.

    MS rolled the Intel CPU patch back about a month ago:

    http://www.zdnet.com/article/windows-em ... ectre-fix/

    And, again, the issue occurs on systems that never received the patch (and which worked fine previously, and have not received OS updates since working fine previously), which unless I'm mistaken, is what was initially blamed for the issue?

    Thanks, digitalsoapbox for the support. Which windows version and build are you using?

    In my case "Don't show unique icons" option reduces the lag significantly. But this can not be considered a solution or workaround, as not having unique icons HUGELY affects the usability.

    Is anyone wants to test if your system is affected, here is the batch script that I used.

    Download this capx, create a text file in the same folder, name it RunAgain.bat :

    start test_lagging_win10.capx
timeout /t 40
taskkill /IM Construct2.exe
timeout /t 10
RunAgain.bat
[/code:1g5bkakw]

Run the bat file. Take a note how many seconds does it take for C2 software to open. Leave the script running and go have some lunch or dinner 
Check again in 30-60 minutes to see if C2 startup time has increased. And then test if there are any lags with opening event dialogs, context menus, sprite editor etc., similar to this [url=https://youtu.be/8thpozIBnUU]video[/url].

    My PCs are on this build:

    October 10, 2017—KB4041676 (OS Build 15063.674)

    It's stable and since I need to use them for work, I don't have these PCs receiving or installing any updates (if it can run 3DSMAX, Photoshop, After Effects, various mocap software and more it should definitely be able to run C2). This would be before the patches that are supposedly causing the issue. I've also reported this before but based on Ashley's reply, he's either not seen this info (though I have mentioned multiple times in other threads on this issue that I have not installed the updates the issues are allegedly cause by), but that's unlikely as he commented on one of the previous mentions in another thread.

    There is no and has never been an accepted definition of "third party" in which a piece of software that is marketed as running on a specific OS is the first party and the OS is the third.

    The OS is the host. The software is the guest. With no OS, the software cannot run. From the disturbingly limited POV of Scirra, Microsoft may be *a* third party, but you are still a guest in their house.

    If the software does not run on the OS, that is a software issue, regardless of if it ran on the OS okay previously, and regardless of that being the fault of the OS or software. The OS makes the rules. Period. This has always been the case. It is not a new development or concept. It predates the birth of every single person on this forum.

    Construct is not a special case. It is not an exception to established rules. Microsoft provides a platform, and developers create software to run on their platform, not the other way around. You are running on their OS at their discretion. If they want to make changes, and it breaks something you did (assuming that's what happened here), it's on you to make it work again. Not them. This has always been the case, outside of extreme circumstances. Once again, Construct is not an exception to the rules. It is a tiny software fish in a much, much larger OS ecosystem. It is not the top of the food chain, or the arbiter of change, on the OS.

    In the end, none of this argument matters. Fix the software so it works properly or sunset it. One thing that can't be explained away regardless of shifting definitions of terms or the placement of blame is the effects of selling faulty software, to both Scirra's reputation, or potential violation of, say, Steam's policies surrounding the sales of software with known, breaking issues. So if Scirra doesn't want to fix their software, I can't imagine there won't be any (more) consequences to that decision.

    Maybe Scirra can also not try to alter the definition of trolling to pointing out that it is a software developer's responsibility to ensure that the software they produce works properly, and no one else's? Because it isn't. It's on Scirra to support the software they develop and sell. Transparent attempts to handwave away this responsibility, or place blame, does not make the software work as it should. Vague, conveniently-applied "rules" on a forum do not change this, and I would think it would be better for Scirra's reputation that it's dealt with on their forum instead of say, the rest of the internet where, as a reminder, Scirra has zero control of what content can be posted or read.

    But hey, it's your business. You can not support the products you sell if you don't want to, despite that approach going against previous announcements related to ongoing support. But that's going to cause issues past the hurt feelings of a Scirra developer that can't admit when they may be wrong about something.

    NOTE: Edited for overall clarity, spelling, snark.

    For the record, this is categorically an issue with Windows. Unless the dialog still takes a long time to appear with "Don't show unique icons", then any delay is entirely in Windows, and there is no question about whether it's in C2. I also can't resist pointing out the great irony that people accuse us of depending on third parties too much with HTML5, but you can't get away from third parties on native either!

    The dialog does, in fact, still take a long time to appear with the "Don't show unique icons" option set in preferences. And, as mentioned previously, it gets worse over time until a full restart.

    Interesting that this occurs only with C2 and not with any other, more complex software that features heavy icon/generated thumbnail usage that I've tested, but is categorically a Windows issue due to recent patches, despite occurring on multiple systems here (as in, directly in front of my eyes), where Windows updates have not been installed in months.

    If I'm reading this correctly, the "third party" you're referring to is Microsoft. In which case, reality is that they're the first party. Because it's their OS that software is expected to run on, and not the other way around. It is the software developer's job to ensure that the software is fully functional on the OS it's designed to run on - especially when that software is still for sale, and even if the blame can "uncategorically" be placed on an OS update. This fact - and it is a fact - of software development doesn't seem all that difficult to understand. The mental gymnastics to avoid the obvious probably take more effort.

    You mentioned upon the release of C3 that Scirra would be fixing any bugs that appear in C2 - this is a bug, regardless of whose fault it is. If Scirra will no longer be fixing known, proven issues that occur when using the software they have for sale, please let users know so that the appropriate steps can be taken.