Bug 236961

Summary:	iOS Safari 15 can't hearing voice after notification sound from another app while connected bluetooth headset
Product:	WebKit	Reporter:	PhungDuong <ddp.bkdn>
Component:	WebRTC	Assignee:	Nobody <webkit-unassigned>
Status:	NEW ---
Severity:	Major	CC:	daginge, ddp.bkdn, webkit-bug-importer, youennf
Priority:	P2	Keywords:	InRadar
Version:	Safari 15
Hardware:	iPhone / iPad
OS:	iOS 15

Description PhungDuong 2022-02-21 00:32:59 PST

Problem:
We have encountered an issue that during a WebRTC conference, when we connected an iPhone with third-party Bluetooth (e.g Inpod12, Xiaomi, etc...) and joined a call, we can't hear the voice after notification sound from another app (e.g Messenger, Facebook, Gmail, etc..)

Test environment:
- Device A, iPhone 12 Pro iOS: 15.1, Safari mobile browser
- Device B, whatever desktop/mobile browser

Steps to Reproduce:
1. Connect a Bluetooth device, I reproduced using Bluetooth Inpods 12.
2. Open our product link: https://it.dev-akameet.link/video/room/demo
3. Provide permissions to choose media devices.
4. Click the button "Join".
5. During the call on safari, when notifications sound from another app is coming on Device A, Device A can't hear the voice from Device B even though the state of Bluetooth is still connected. 

*Note: We also test on https://meet.jit.si/

Expected results:
The voice should be normal after notifications sound stoped

This has been tested and reproduced on:
- Safari on iOS 15.2 on an iPhone XR

We have not yet found any work around this, so please let us know if we have to provide more information or if you have any suggestions.

Comment 1 Radar WebKit Bug Importer 2022-02-23 15:36:26 PST

<rdar://problem/89382340>

Comment 2 PhungDuong 2022-03-01 02:16:31 PST

For this case, we add more actual phenomena that's when notifications sound from another app is coming.
The voice is come out of the phone speaker, even though the Bluetooth state is connected.

Comment 3 PhungDuong 2022-03-09 00:39:09 PST

Dears.
For this bug, we can't reproduce the case described in comment #2, so please ignore this comment. Thanks.