I realise that this query is probably outside the scope of Phaser knowledge, but thought I’d post this issue as someone might be able to suggest things to try.
I am using microsoft.cognitiveservices.speech.sdk to convert speech to text via the microphone. This works fine with phaser running within all operating systems except on the IPAD safari. When I run the continuous recognize speech with Phaser running on the IPAD there is large lag, incorrect results or doesn’t work at all. When I destroy the Phaser game object it works fine - on the IPAD.
I had tried many different gameconfig options with no luck. WebAudio off, noaudio, context = new webkitaudiocontext. Still the same issues. I also noticed when using the window.speechSynthesis.speak sdk - the oneend callback is very delayed and sometimes doesnt work only on the IPAD.
I have tried a complete barebones Phaser project and I still get the issue only on the IPAD.
Any suggestions would be HUGELY appreciated.