Heads up for 1.25 release

191012141525

Comments

  • ArmellineArmelline Member, PRO Posts: 5,327

    Viewer more than Preview, but noticed in both. The project I noticed it most strongly does use control camera. However, like some others, I made a super-basic test project and experienced it. I cannot reliably reproduce it, and the projects I've been working on for the past couple of weeks have been ones that wouldn't exhibit it anyway. Afraid I really can't offer anything to help narrow it down.

  • IceboxIcebox Member Posts: 1,485
    edited March 2016

    @Hopscotch im sorry i didnt read your second message, I work on the mac version

    The results with me are always like this :

    GS Preview : No stuttering except when i stay on a project for a long time.

    Gs viewer iphone 5s/6/3 ipad devices : No stuttering
    Adhoc same devices: there is stuttering

    GS viewer android ( samsung galaxy s5,s4,prime): stutters alot
    Apk android same devices stutters but less than the viewer

    EDIT: The frame rate is always from 53-60 fps although before it used to run 58-60 fps

  • HopscotchHopscotch Member, PRO Posts: 2,782

    Good info @Icebox

    If you say the android apk stutters sometimes, can you guess what your app is doing at that time? E.g. stutters for a while after a scene starts, or stutters when new actors spawn or new sounds play, etc. Try to think what your app is doing at the time. Or is the stutter less, but consistent throughout?

  • IceboxIcebox Member Posts: 1,485

    @Hopscotch yea the stutter is less on apk but consistent , it never stops. On viewer it stutters every second but on an apk it takes about 3 seconds to stutter and goes on , i tried switching things off here and there but it never stopped, i thought its because i had many constrains on the scene switched them all off but it remained the same , i made a new scene with nothing on it except one actor with no sound or graphics and it still had the same result.

    I decided to make a new project and test again , thats when i knew its a bug. The new project only had one actor on it . The actor only has a rotate behavior i switched the physics density/ friction /bounciness to 0 and kept moveable on placed the actor on screen and test on three android devices it stuttered . So I guessed from that , that the problem is not with my logic.

    I tested again on another project with an actor, i switched moveable off and had it interpolate its x position , the stuttering happened again. Thats the reason why i said it has nothing to do with logic this time , i had logic problems before and when i cleaned my project the results were great. So thats why i kept on saying the problem is only with this build. I dont know if your results earlier were tested on android devices or not ? if they are then i think something is wrong with my devices and your right, but judging from the same experience other people are having i think its a bug.

  • BitvorkBitvork Member, PRO Posts: 119
    edited March 2016

    OSX and iMac are both able to show serious stuttering itself sometimes. I have iMac 2013 with latest El Capitan and you don't want to see that terrible (disaster) stuttering user experience sometimes (random stuttering). I always keep all of my computers clean. Don't like any garbage. But I am still experiencing terrible stuttering sometimes across all tabs/desktops/screens on my iMac (even after complete re-install). Working with Creator is a disaster then but not because of GameSalad. I noticed this terrible stuttering after El Capitan installed (everything was fine with Yosemite). I am not saying that GS is super fast without stuttering (their code is probably oversized and it is harder and harder to keep great performance) but also be aware of Apple possible performance issues. It is really weird that someone see that stuttering on desktop and adhoc and someone doesn't.

  • The_Gamesalad_GuruThe_Gamesalad_Guru Member Posts: 9,922

    @Bitvork said:
    OSX and iMac are both able to show serious stuttering itself sometimes. I have iMac 2013 with latest El Capitan and you don't want to see that terrible (disaster) stuttering user experience sometimes (random stuttering). I always keep all of my computers clean. Don't like any garbage. But I am still experiencing terrible stuttering sometimes across all tabs/desktops/screens on my iMac (even after complete re-install). Working with Creator is a disaster then but not because of GameSalad. I noticed this terrible stuttering after El Capitan installed (everything was fine with Yosemite). I am not saying that GS is super fast without stuttering (their code is probably oversized and it is harder and harder to keep great performance) but also be aware of Apple possible performance issues. It is really weird that someone see that stuttering on desktop and adhoc and someone doesn't.

    I agree. I made the mistake of installing el capitan and after two days ended up time machining back to Yosemite. It's horrible.

    I have been doing test builds with old games in both viewer for ios and android as well as adhoc and apk builds and I encountered no stuttering at all. Hmmm

  • HopscotchHopscotch Member, PRO Posts: 2,782
    edited March 2016

    Thanks @Icebox, I am not denying it is a bug, something is obviously amiss. Though to help GS to cover all possibilities we should try to be as specific as possible as it is specific to SOMETHING.

    I installed a 1.24.42 android apk of my heavy game, up top, on my super slow and handy-capped Samsung Galaxy Tab 3 and it still ran OK.

    The rotating squares app also stayed completely smooth.

  • IceboxIcebox Member Posts: 1,485

    @Hopscotch I dont get it really , then i guess its something with my device , thanks for your time ! ill just wait for the next release and see how it goes

  • HopscotchHopscotch Member, PRO Posts: 2,782

    @Icebox, since the GS team flagged the issue as confirmed, let us hope they have a notion of what it could be and it is an easy fix for them.

  • BigDaveBigDave Member Posts: 2,239
    edited March 2016

    I do interpolate the position of the camera as well as give additional direct increments on the speed based on the situation.

    with my new game I experienced stutter on the newest and strongest iPhone
    and was quite loud about it in the forum but..

    I realised I still was still 1.24.41
    after updating to the newest 1.24.42 it was fluent not sure if it was related to the app already being cached in the device, if this is the case?
    (patch notes said minor general improvements and it had an effect)

    So it ran fluent on the iPhone6s+ but for Android i do not know.
    Would be nice if someone here could test and state for Android devices.

    The main issue I have now is chartboost related is it can block your game totally right now forcing the player to quit the app. But this is confirmed to be fixed in 1.25.

  • The_Gamesalad_GuruThe_Gamesalad_Guru Member Posts: 9,922

    We ask and post because we are part of the beta team and we are trying to recreate these bugs for the team. If we can recreate them, then the team can narrow down where the issue is. This is essential to debuging.

  • SocksSocks London, UK.Member Posts: 12,822
    edited March 2016

    @Lost_Oasis_Games said:
    We ask and post because we are part of the beta team and we are trying to recreate these bugs for the team. If we can recreate them, then the team can narrow down where the issue is. This is essential to debuging.

    Can you check something out for me . . . . a bug was introduced in Creator 1.24.39 onwards (and is still in the current 1.24.42) with Angular Velocity, which is somewhat broken now.

    Angular Velocity works as normal in that you can set a value of - for example - 500, and your actor will rotate at 500°/sec - as it should . . . but if you try to read an actor's Angular Velocity the value you see will be 1.745% of the correct value - so for an actor spinning at 500°/sec it will show an Angular Velocity value of 8.7266 rather than 500.

    In older versions (pre-1.24.39) - if you were to use a change attribute to change an actor's Angular Velocity, to 100, and use a Display Text behaviour to display Angular Velocity's value, you would get - obviously - '100'. But since 1.24.39 - Angular Velocity will always return 1.745% of the correct value, so in the case of a value of '100' Angular Velocity will return 1.745 instead of 100.

    So, imagine you have an actor that has two behaviours, firstly a change attribute Angular Velocity to 100 and then a Move behaviour with the speed set to 'Angular Velocity' . . . what should happen is that the actor should rotate at 100°/sec and move at 100px/sec . . . which is what happens prior to this bug appearing . . . . instead in 1.24.39 onwards it rotates at 100°/sec (correctly) and moves at 1.745px/sec (incorrectly).

    Lots of other aspects of Angular Velocity are screwed up too, another example would be where you have an actor spinning around, and you want to bring it to a gradual stop, in the past you could simply interpolate the Angular Velocity to 0 . . . that doesn't work now, rather than the Angular Velocity slowly coming to a stop (let's say you set the interpolate to work over 4 seconds) it will instead stop instantly rather than gradually over 4 seconds.

    Another method I often use for an interpolate that you can interrupt, would be to say:

    When X >0
    --Constrain X to X-1

    So if X was 600, this would interpolate X from 600 to 0 over 10 seconds (as Constrain fires on every code cycle, so 60 times a second).

    This works as normal, but if you try it with Angular Velocity . . .

    When Angular Velocity >0
    --Constrain Angular Velocity to Angular Velocity-1

    . . . it does the same as Interpolate with Angular Velocity, it jumps instantly to 0, rather than taking 10 seconds to go from 600 to 0.

    . . . . . . .

    So, my question, could you, in 1.25, throw an actor into a scene, use Change Attribute to change Angular Velocity to some value like 100, and then add a Display Text to display the actor's Angular Velocity . . . to see if you get '100' or 1.745 . . . ?

    Thanks if you got this far !

    Also @Hopscotch and @Armelline and @GeorgeGS and anyone else with access to the beta and a minute to spare - thanks !

  • jonmulcahyjonmulcahy Member, Sous Chef Posts: 10,408

    @Socks

    I did the test slightly differently before I got to the bottom. let me do it using change attribute next.

  • NNterprisesNNterprises Member, PRO Posts: 387

    totally off the stuttering/bug topic... but this update does not include Google Play Services yet, right? I remember it saying sometime Q1 2016

  • SocksSocks London, UK.Member Posts: 12,822
    edited March 2016

    @jonmulcahy said:
    @Socks, looks like it's still in there

    Thanks for testing it out Jon, much appreciated :) could you throw it onto the 1.25 debugging list ?

  • RThurmanRThurman Member, Sous Chef, PRO Posts: 2,879

    @Socks @jonmulcahy

    It is returning radians, rather than degrees. (1 degree = .01745329252 radian)

    To convert it back to degrees, use the formula
    radians x 57.295779513

    so for example, 1.7453292519943 x 57.295779513 ≈ 100

  • The_Gamesalad_GuruThe_Gamesalad_Guru Member Posts: 9,922

    @RThurman said:
    @Socks @jonmulcahy

    It is returning radians, rather than degrees. (1 degree = .01745329252 radian)

    To convert it back to degrees, use the formula
    radians x 57.295779513

    so for example, 1.7453292519943 x 57.295779513 ≈ 100

    Yes this seems to be in concert with the update to the box2d engine. They may have missed exporting the proper metric to that variable.

  • RThurmanRThurman Member, Sous Chef, PRO Posts: 2,879

    @Lost_Oasis_Games said:
    Yes this seems to be in concert with the update to the box2d engine. They may have missed exporting the proper metric to that variable.

    Yes... Box2D does everything in radians. But its simple to convert.

  • SocksSocks London, UK.Member Posts: 12,822
    edited March 2016

    @Lost_Oasis_Games said:
    Yes this seems to be in concert with the update to the box2d engine. They may have missed exporting the proper metric to that variable.

    @RThurman said:
    @Socks @jonmulcahy

    It is returning radians, rather than degrees. (1 degree = .01745329252 radian)

    To convert it back to degrees, use the formula
    radians x 57.295779513

    Cheers ! Good detective work too ! :)

    Hopefully we can return to degrees, as degrees are used everywhere else in GameSalad, it's odd to use degrees for things like an actor's Rotation attribute, the Rotate behaviour, Angular Velocity and so on - to then have that attribute stored/returned as radians is pretty confusing (and screws up quite a few pre-1.24.35 projects of mine!).

    And a radians to degrees multiplication will always involve an irrational number so this may well introduce rounding errors somewhere along the line (?).

    P.S. Not sure if this explains that you can no longer interpolate Angular Velocity as well ?

  • vikingviking Member, PRO Posts: 322

    Maybe this is why the golf ball in my game stopped spinning back consistently after the new version of Box 2D was merged into the latest version of GS...? My physics were working 100% for a year and then suddenly 90% of the time the ball stopped spinning back. I hope they will fix this in 1.25.

  • RThurmanRThurman Member, Sous Chef, PRO Posts: 2,879

    @Socks -- Cheers! Yes on rounding errors. Yes on asking GS to convert it back to degrees. (Or at least be consistent one way or the other.)

    @viking -- I bet your right about your guess. I bet your wrong about your hope. (But maybe its a quick fix in GS to convert it back to radians before they release 1.25 to the public.)

  • The_Gamesalad_GuruThe_Gamesalad_Guru Member Posts: 9,922

    Did anyone post a bug report on this?

  • SocksSocks London, UK.Member Posts: 12,822

    @Lost_Oasis_Games said:
    Did anyone post a bug report on this?

    I thought it was best if it came from one of the 1.25 beta team . . . so it would be good if someone could report it. Thanks in advance !

  • SocksSocks London, UK.Member Posts: 12,822
  • DuesDues Member Posts: 1,159

    Ok, so whats really wierd is that I created a completley new project with 1.23.1 and it's still not smooth... :confused:
    Haven't done adhoc on it yet but still..

    I will do a test with an even older build and see if I get the same result.

    Could it be something that GS adds to the project even though im running separate builds? Not sure that's even possible...

  • The_Gamesalad_GuruThe_Gamesalad_Guru Member Posts: 9,922
    edited March 2016

    @Dues said:
    Ok, so whats really wierd is that I created a completley new project with 1.23.1 and it's still not smooth... :confused:
    Haven't done adhoc on it yet but still..

    I will do a test with an even older build and see if I get the same result.

    Could it be something that GS adds to the project even though im running separate builds? Not sure that's even possible...

    It would be more productive to begin turning off code to find the offending issue at this point. Are you running El Capitan?

  • DuesDues Member Posts: 1,159

    @Lost_Oasis_Games
    That won't help me much (although I already tried that) since even a new simple project with a control camera and a change attribute still isn't smooth. I'll send you a project file so you can test on your computer if you want.

  • The_Gamesalad_GuruThe_Gamesalad_Guru Member Posts: 9,922

    @Dues said:
    @Lost_Oasis_Games
    That won't help me much (although I already tried that) since even a new simple project with a control camera and a change attribute still isn't smooth. I'll send you a project file so you can test on your computer if you want.

    I'm starting to wonder if this issue is also related to the box2d update?

Sign In or Register to comment.