3

The createSpeechRecognizer() method was added in API level 8. The createOnDeviceSpeechRecognizer() method has been added in Android S. But their descriptions are same (according to the official site). Can anyone clarify the difference between them? Is there any possibility in the near future of the createSpeechRecognizer() method being deprecated?

Lukasz Blasiak
  • 642
  • 2
  • 10
  • 16
Samudra Ganguly
  • 637
  • 4
  • 24
  • 2
    In the `SpeechRecognizer` class description, it says that _"The implementation of this API is likely to stream audio to remote servers to perform speech recognition"_. The name `createOnDeviceSpeechRecognizer` sounds like a recognizer created by that method won't stream any data to a server - i.e. the recognition is performed on-device. That's just my guess though. – Michael Apr 28 '21 at 16:47
  • Thanks a lot for the clarification. Your guess seems very logical. – Samudra Ganguly Apr 28 '21 at 16:49
  • But if the recognition is performed on device, how can ERROR_TOO_MANY_REQUESTS occur? This error has also been added in Android S. – Samudra Ganguly Apr 28 '21 at 17:04
  • That doesn't automatically mean that the new error code is related to the new method. – Michael Apr 28 '21 at 17:07
  • Yeah, this seems logical too. Thanks man. – Samudra Ganguly Apr 28 '21 at 17:25

0 Answers0