NetOne's Forum Posts

  • FYI Google are now tracking this as a Chrome issue here: https://bugs.chromium.org/p/chromium/issues/detail?id=1014614

    Oh... I raised my bug on the Android tracker but they closed it to track the issue against this bug in the Chrome tracker instead. But they stated that the bug is in Android ?? (reporting too high latencies to chrome) not in Chrome?. Unless they have maybe resolved the issue in latest Android?

    Hmm , I dont know. But it is good that Google have at least acknowledged the issue and are looking into it. (Although you would think a company the size of Google would be all over these sort of issues long before joe public starts to raise them, makes me a little nervous about the organisation there.).

  • Ok Constructors. More information on Android + Audio + Chrome / Webview audio lag.

    So, according to Google.

    Chrome assesses the device capabilities based on Android OS returning a value for baseLatency

    this value

    BaseLatency TEST http://rtoy.github.io/webaudio-hacks/tests/osc.html

    if the value is "high" Chrome adds a ton of latency into the audio path to avoid audio glitching.

    apparently the value returned is erroneously too high (far too high) on most of the Android phones currently out there. even on powerful phones capable of much lower latency.

    This is where the 300-400ms lag is coming from.

    Ive noted on the two phones I have direct access to

    Cubot power. Very little lag on the app.website , returns a baseLatency of 256 frames

    On the MOTO G5. 300ms lag on the app/website/ returns a baseLatency of 3200 frames.

  • OK, just to close the weird pitch issue on IOS

    (especially for anyone thinking about trying to make synths instead of games like silly me)

    The bug has been closed with a solution.

    Im not going to get too tech here as Im not too tech on the matter. Have a look at the bug link for more info.

    But... The sample used for the test is known as a single cycle waveform and was taken from a well known sample library. Problem was that single cycle waveforms are very short. Like milliseconds short. Too short for the IOS audio minimum sample buffer is seems. The minimum buffer was being looped but the sample was smaller than the minimum buffer size. And this makes total sense with what is being heard.

    This is not an issue on Android, but who would want to do anything with audio on android with all that lag !!!.

    So this can be solved easily by using a longer sample or looping the sample a few times in an audio editor and re-sampling it.

    Thanks everyone for their input. especially tarek2 and Ashley. Looks like i will indeed be joining the Apple fraternity now (something you would have never heard me saying a few days ago!!!) as I would like to explore this music instrument side of things a bit more. Cheers.

  • Cheers tarek2 youve been super helpful there.

    ive raised a bug report. https://github.com/Scirra/Construct-3-bugs/issues/3457

    Ive linked to your vids but if you want/need to delete those video let me know and I can stick them on my gdrive no problem.

    Im thinking I could get an Ipod for now so at least I could test websites on IOS apparently they have the same chipset as an iphone 7. But will just wait to see what Ashley has to say regarding this. As it may be again outside his control. If so then its back to desktop apps or I have to learn to program with JUCE.

    edit< just been down the apple store, ipods are tiny! Too small for major gaming or music apps. Maybe just good as music player. So I await see what Ashley say.

  • tarek2

    What?!?!?

    You give such hope showing that beautiful low lag on IOS but then....

    ... You take it all away with that horrible mutant sound..!

    What is that? It is lower pitch and distorted on Ios across 3 devices...?

    Man i was just about to get a mac mini and iPhone for testing but i cant justify if there are further issues on IOS.....

    Maybe it's just a bug with the frequency change??? It wouldn't be that common to use in games. Do other games exhibit different sound between IOS / android?

    Hmm.. thanks again , this is very revealing... I think i will raise a bug later linking your vid if you dont mind.

  • OK then cheers, I will try to raise with the G men for FWIW.

    I still love construct BTW.

    Blizzcon / Overwatch 2 OMG ! 😮😮😮

  • No offence Ashley but you didn’t really answer anything there.

    Im not blaming Construct, Im pretty confident that you have done everything in your power to get sound lag down within the scope of the engine. I get that there are multiple parts to this equation. However there must be a reason that Geometry Dash and COD et all can get down to ~50ms sound lag on the same devices that Construct games are giving 400ms and if it something we can bring to Google then isn’t it worth doing so even if they’ve heard it all before?

  • Ok, this is my first venture into Android mobile land with Construct 3 but I must admit Im a little shocked.

    Before we kill this thread off only to be resurrected by the next poor sucker who ventures down this road.

    Please Ashley if you could confirm the following.

    1) You are essentially saying that 300-400ms sound lag is normal and accepted on android games on what must be over 90% of devices?

    a. If this was the case then I don’t think anyone would be playing games on android. I have just played geometry dash on 2 of the offending devices and there is no perceptible sound lag when dying or when entering a portal. (I would estimate it is under 3 frames at 60fps)

    b. The superpowered website is talking about a general RT lag of 30-50ms for 99% of android devices , this would be acceptable and almost non perceptible for most gamers however this is 10 times less than what we are seeing with the Construct exports something dosent add up?

    c. If you are saying that 300-400ms is generally the accepted norm with Construct 3 games or mobile javascript games in general (and only native can give you low sound lag) then you are admitting that Construct 3 games on android are not actually viable. Is this so?

    2) The superpowered website are selling Javascript audio libraries that boast very low latency.

    a. Can these be uses with construct scripting?

    b. Could similar tech be used for Construct audio?.

    3) The superpowered website database shows that IOS has low sound lag on all devices. Like less than 12ms on most of the range as compared to the 30-50ms on android. Which is why it is the choice for audio applications I assume.

    a. But do construct games also display this low latency on iphones. Or is there an extra 250-350+ milliseconds added into the path like on android? (I ask As I am ready to ditch Android and embrace apple if this was the case)

  • so far pretty much exact same lag showing in 7 , 8 , 8.1 and 9 on devices tested.

    Currently showing lag on 9 on S8 (as shown in tarek2s demo above and linked on the bug report)

    so unless someone can show otherwise Im going with the current evidence.

  • Have just quickly raised a bug report.

    github.com/Scirra/Construct-3-bugs/issues/3455

    But will continue investigation here on weekend

  • Ok Halloween duty over back to s**t posting on the net..... hehe.

    anyhew....

    WackyToaster

    So your app is saying as far as construct goes it thinks the audio is playing and stopping within a few milliseconds which is good but not what is actually happening in the hardware?

    also

    2 more videos

    Samsung S7 Android 8 = lag

    drive.google.com/open

    Hawaei P10 Android 8 = lag

    drive.google.com/open

    all of these devices are lagging between 300-400 ms (I have checked in video editor)

    its a bit late and im knackered so I wont update the app to include animation right now but will do on the weekend

    however I think there is enough evidence here to clearly show that there is a major lag issue.

    edit > Tom I think there is far too much space between new lines (paragraph breaks) on desktop

  • Try Construct 3

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

    Try Now Construct 3 users don't see these ads

  • how are we defining wiggle here?

    this sounds like something you could do very easy with sine behavior.

  • No its a good test. I will try to update tonight. Wont be till much later tho.

    edit > WackyToaster already indicated that sprite animation reacted instant but the sound still lagged. But yea will test later. Need to go out and scare some kids first tho.

    🎃🎃🎃🎃🎃🎃🎃🎃🎃🎃🎃🎃🎃🎃🎃🎃

  • fredriksthlm

    I am happy to update the experiment to include an animation of some sort and we can see how it compares,

    but my primary here interest is in touch to sound lag for real-time music instruments and right now the evidence points to a serious sound lag issue on android.

    There may be thousands of construct users but ive been here long enough to know that I am one of a very very small handful of users who regularly test and post bug reports as I really want construct to be the best it can be.

  • Asmodean

    no not using worker as advised by Ashley

    tarek2

    Really appreciate you taking the time to do that.

    What you have demonstrated there is some serious lag. Especially on the S8.

    That is around 400ms.

    Not only will that make a playing a real-time instrument impossible.

    But if it is happening in Construct games that would also be unacceptable.

    It’s at this point I would love someone to tell me there is some issue in my events but I think we are exposing serious touch lag issues on android here. (at least with regard to sound and at least on Moto and Samsung devices)

    Slightly off track. Anyone with Apple I phone?, are you getting similar lag on website? (Wondering is this actually just a Android issue)