After some more trying I found out that only the speech recognition task works, whose buffer comes first. That means if I do it like this: self.speechAudioBufferRecognitionRequest1.append(buffer) self.speechAudioBufferRecognitionRequest2.append(buffer) then the speechRecognitionTask1 works. If I do it like this: self.speechAudioBufferRecognitionRequest2.append(buffer) self.speechAudioBufferRecognitionRequest1.append(buffer) then the speechRecognitionTask2 works.