Give us as much detail as possible regarding the issue you're experiencing:
Unity Editor version : 2022.2.12f1
ML2 OS version : 1.12.1
Unity SDK version : 2.2.0
Host OS : Windows 11
I was trying to incorporate speech to text using Whisper AI. However, after trying to build to the headset, I encountered build errors saying I needed to change my target architecture to arm64. It seems that I cannot change the target architecture to arm64 if I want to build to the headset. Is it true that in order to build to the magic leap the target architecture must be x84-64? I trued reimporting my assets and deleting my library folder but am still encountering the same issues. If it is true that I cannot use arm64, does anyone know of other open source tools I can use for speech to text?
Hi @kritig,
the ML2 is not an arm device so you will need to have your build target set to x86-64. Many android libraries can be re-targeted to x86-64, though I am not familiar with whisper ai so others may have more insight there. I will ask if there is a preferred speech to text api.
-Sidney
Thanks! I'm looking into other open source unity packages I could use and found this one: GitHub - yasirkula/UnitySpeechToText: A native Unity plugin to convert speech to text on Android & iOS --the developer says that the plug in uses the device's build-in speech recognizer. Does the magic leap have support for something like this?
I think it would work. It is important to note that our OS is based on android 10, so it will not be able to work offline. I know that we have voice commands and voice intents (Voice Commands | MagicLeap Developer Documentation).
As for that specific plugin, I would suggest trying it and seeing if it works.
I think this was answered pretty well in this other post Speech-to-text service issues and recommendations - #4 by mrushton
and I am going to close this thread.