IssueTracker

Because Media2 is using the Media1 media session. A media button event coming from bluetooth is converted into an onPlay/onPause command by the Media1 MediaSessionCompat support library. I'm pretty sure that BT commands arrive via media1 at least for the three BT headsets that I'm usually testing with.

We can not support you 1:1 with your app I'm afraid. So you would have to narrow down to some more specific questions.

I would wonder whether you can repro this on that phone with your BT headphones with the session demo app of Media3. The delivery of BT events is exactly the same as we are using the support library under the hood. If it works there, it seems to be an app issue. If it doesn't work I would be interested to hear the BT headset you are using.

Can you tell us in a bit more detail how you fixed the problem?

The SOF post says that the media button is not properly routed to the app as expected. Did your fix change something in the priority? When you check with

adb shell dumpsys media_session

do you see a change in the section Audio playback (lastly played comes first) with your fix?

I don't know what is the difference between TTS player and traditional players that would cause this, but your feedback would be appreciated in the sense that we would learn what you actually did to receive the media button events.

Sorry I lost a lot of time on this issue, I can't invest more for the moment.

I just played a silent sound for less than 1s using MediaPlayer class and when done I start using TextToSpeech.

You have my sample project demonstrating the issue, just add the play of silence, on completion start speaking using TextToSpeech and it will work.

I hope this can help you fix the issue. 😉

Based on the analysis above, it sounds as the TextToSpeech playback is not set up in a way that media buttons can be routed correctly to the app.

Assigning to the TTS team for further triage. Is this a known issue? The linked SO post in #5 suggests this problem exists since Android Oreo.

(Apologies if you are not the team maintaining the TextToSpeech implementation, please reassign as needed)

TTS playback doesn't inherently handle the media buttons - quite simply because it cannot be paused or rewound like a traditional media source.

If you do want such functionality, please use the .synthesizeToFile() method (instead of .speak()). This will give you a wav file which you can then put through a media player, which will then handle the media playback buttons.