1186135 - [Camera][Video] Recording a video, no chime indicator that recording has started or ended

Status: VERIFIED FIXED

(Firefox OS Graveyard :: Gaia::Camera, defect)

Description

[Oliver Nelson [:oliverthor]]

Attachment: logcat_20150721_1225.txt

Description:
When a user begins to take a video capture from the Camera app, they will observe that no chime can be heard from when recording begins or when recording is ending. This is a regression from previous builds and is not sure whether this is by design.


Repro Steps:
1) Update a Aries to 20150721140411
2) Open the Camera app
3) Switch to video
4) Start recording a video

Actual:
Camera is silent when video recording begins and ends

Expected:
Audio chime is heard when beginning to record a video


Environmental Variables:
--------------------------- 

Device: Aries 2.5
Build ID: 20150721140411
Gaia: 4fe0507781f3ed56c8ae5e66dd9489165d1ff68e
Gecko: d7a9e44717b7
Gonk: 2916e2368074b5383c80bf5a0fba3fc83ba310bd
Version: 42.0a1 (2.5)
Firmware Version: D5803_23.1.A.1.28_NCB.ftf
User Agent: Mozilla/5.0 (Mobile; rv:42.0) Gecko/42.0 Firefox/42.0

Device: Flame 2.5
BuildID: 20150721010202
Gaia: 4fe0507781f3ed56c8ae5e66dd9489165d1ff68e
Gecko: 3a4bfa5d2d02
Gonk: 41d3e221039d1c4486fc13ff26793a7a39226423
Version: 42.0a1 (2.5) 
Firmware Version: v18D
User Agent: Mozilla/5.0 (Mobile; rv:42.0) Gecko/42.0 Firefox/42.0
************************

Issue DOES NOT REPRO on 2.2 for flame devices
Results: Audio chime is heard when beginning to record a video

Device: Flame 2.2
BuildID: 20150721002506
Gaia: e1e6317f17a840b19af9dbb25f5a771d8d9fa161
Gecko: 1bc400014f71
Gonk: bd9cb3af2a0354577a6903917bc826489050b40d
Version: 37.0 (2.2) 
Firmware Version: v18D
User Agent: Mozilla/5.0 (Mobile; rv:37.0) Gecko/37.0 Firefox/37.0
--------------------------- 

Repro frequency: 5/5
See attached: 
video- https://youtu.be/7pCgNigwLyo
logcat

Comment 1

[Peter Bylenga [:PBylenga]]

[Blocking Requested - why for this release]:
Noticeable regression.

Requesting a window, I highly suspect this to be related to the other audio management issues.
bug 1130350 and dependencies

Comment 2

[KTucker [:KTucker][Inactive 3/4/2016]]

b2g-inbound regression window:

Last Working
Device: Flame
BuildID: 20150720073224
Gaia: a47e0c58dd1340a17be7cc96ccd90fcad15a922b
Gecko: f7d9b77a7afb
Version: 42.0a1 (2.5 Master) 
Firmware Version: v18D
User Agent: Mozilla/5.0 (Mobile; rv:42.0) Gecko/42.0 Firefox/42.0

First Broken
Device: Flame
BuildID: 20150720074629
Gaia: a47e0c58dd1340a17be7cc96ccd90fcad15a922b
Gecko: e9b4a2b76fb3
Version: 42.0a1 (2.5 Master) 
Firmware Version: v18D
User Agent: Mozilla/5.0 (Mobile; rv:42.0) Gecko/42.0 Firefox/42.0

Gaia is the same so it's a Gecko issue.

Gecko pushlog:
http://hg.mozilla.org/integration/b2g-inbound/pushloghtml?fromchange=f7d9b77a7afb&tochange=e9b4a2b76fb3

Caused by changes made in Bug 1184055.

Comment 3

[KTucker [:KTucker][Inactive 3/4/2016]]

Alastor, can you take a look at this please? Here is another issue that might have been caused by the landing for Bug 1184055.

Comment 4

[Alastor Wu [:alwu]]

In preliminary analysis, the BrowserElementAudioChannel didn't get the active notification from the AudioChannelService, so the system didn't open the audio chime sound.
Still finding more details.

Comment 5

[Alastor Wu [:alwu]]

I found that the BrowserElementAudioChannel (BEAC) would not receive the event from the AudioChannelService(ACS), when we are playing multiple audio channel sound in the same process.

The first sound can work well, but the second sound would get stuck, for unknown reason.

In this case, we have two types sound, "content" and "notification".
The "content" channel is used to paused other sound when the recording starts (in DOMCameraControl.cpp), and the "notification" channel is the start sound of the recording.

So, the situation happens like that,
(1) "content" starts, register agent, send active event to content BEAC
(2) BEAC gets event, and pass it to the system app
(3) system app open the "content"
(4) "notification" starts, register agent, send active event to content BEAC
(5) nothing happens

Here is the problem! The BEAC can't get the "notification" active event from the ACS , so the system app would not open it!

----

To verify the problem (can't play multiple audio in same process), I use Dominic's test app [1].

If you play the normal audio first, then the you want to play content one, the later audio can't be playback and vice versa. The later audio can't be playback.

However, after the above steps, if you pause the content audio, the BEAC can receive the inactive event! And then every operations would be back to normal! 

Steps like that,
(1) Start normal  - ok
(2) Start content - fail, BEAC can't receive the active event
(3) Pause content - ok, BEAC can receive the inactive event
(4) Start content - ok, BEAC can receive the active event

Comment 6

[Alastor Wu [:alwu]]

Hi, Baku,
Do you have any idea about comment5?
Very appreciate :)

Comment 7

[Andrea Marchesini [:baku]]

Attachment: bug2.patch

I cannot reproduce this issue. I wrote a test. Am I doing something wrong?

Comment 8

[Alastor Wu [:alwu]]

Hi, Baku,
Your test logic is correct, if the issue happens, that you can't playback audio2.

However, I'm not sure whether the issue can be reproduce on the desktop. Since the mochitest didn't have a gaia competing logic, they are all playable. The problem we got is that we muted the audio in initial, then we can't get the open command from the system app.

And one questions,
Where you declare the permission "audio-channel-alarm" in the test? 
The issue can only be reproduced when we playback different audio channel types audio.

If you have any idea or question, please discuss with me :)
Thanks!

Comment 9

[Hema Koka [:hema]]

Regression - blocking

Comment 10

[Alastor Wu [:alwu]]

Attachment: Independent audio channel active flag

[Bug issue]
We can't playback multiple audio channel type in the same window.

---

[Root cause]
Because we only used one variable to store the active status of the audio channel in TabChild.cpp, it should be stored independently by per channel.

---

[Issue detail]
The |mAudioChannelActive| is used to store the active status of the audio channel of the tab. It is false by default.

Think about this case, 
> Playback the "content" type first, and then "notification" type.

(1) When we playback the "content" type channel, we send the active event (active=true).
    Then, the |mAudioChannelActive| would be changed to true. 

(2) When we playback the "notification" type channel, it can't be playback. 
    Because the |mAudioChannelActive| is already set to true, we can't send the notification again.

In this bug, we would playback "content" first, then "notification" during the recording.
The "content" type is used to pause other sound, because we can't playback other sound during the camera recording.
The "notification" type is used to playback the chime sound.

Comment 11

[Alastor Wu [:alwu]]

Try-server result.
https://treeherder.mozilla.org/#/jobs?repo=try&revision=edd891d10ba2

Comment 12

[Alastor Wu [:alwu]]

Attachment: Independent audio channel active flag

Modify the compile warning, and new try-server result.
https://treeherder.mozilla.org/#/jobs?repo=try&revision=85906e4543fc

Comment 13

[Alastor Wu [:alwu]]

Attachment: Independent audio channel active flag

New patch and try-server result.
https://treeherder.mozilla.org/#/jobs?repo=try&revision=b1c4e24126be

Comment 14

[Alastor Wu [:alwu]]

Comment on attachment 8642792 [details] [diff] [review]
Independent audio channel active flag

Hi, Ehsan,
Sorry to bother you,
Could you help me review this patch?

The root cause analysis is in the comment10.
In this patch, I stored the active status per channel so that the TabChild can send the correct audio channel notification event.

Very appreciate :)

Comment 15

[(no longer active)]

Comment on attachment 8642792 [details] [diff] [review]
Independent audio channel active flag

Review of attachment 8642792 [details] [diff] [review]:
-----------------------------------------------------------------

r=me with the below addressed.

::: dom/audiochannel/AudioChannelService.h
@@ +25,5 @@
>  #ifdef MOZ_WIDGET_GONK
>  class SpeakerManagerService;
>  #endif
>  
> +#define NUMBER_OF_AUDIO_CHANNELS (uint32_t)AudioChannel::System + 1

Please use (uint32_t)AudioChannel::EndGuard_ instead.  That is guaranteed to always remain up to date.

::: dom/ipc/TabChild.h
@@ +640,4 @@
>      bool mAsyncPanZoomEnabled;
>      CSSSize mUnscaledInnerSize;
>  
> +    nsTArray<bool> mAudioChannelsActive;

Nit: please use an auto array with a buffer size of 10.

Comment 16

[Alastor Wu [:alwu]]

Attachment: Independent audio channel active flag. r=ehsan.

Try-server result.
https://treeherder.mozilla.org/#/jobs?repo=try&revision=5d7e8f899979

Comment 17

[Pulsebot]

https://hg.mozilla.org/integration/b2g-inbound/rev/8ea5c69ae101

Comment 18

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/8ea5c69ae101

Comment 19

[Cola, OOO till Oct. 8th]

This bug has been verified as "pass" on latest Nightly build of Flame master and Aries master by the STR in Comment 0.

Actual results:Audio chime is heard when beginning to record a video.

See attachment: Aries v2.5.3gp

Reproduce rate: 0/5

Device: Aries KK 2.5(Pass)
Build ID               20150810003528
Gaia Revision          09dea2d5ff21cdb56da35fe4aa5bf4c90cf1da7f
Gaia Date              2015-08-09 17:11:47
Gecko Revision         https://hg.mozilla.org/mozilla-central/rev/0e269a1f1beb
Gecko Version          42.0a1
Device Name            aries
Firmware(Release)      4.4.2
Firmware(Incremental)  eng.worker.20150810.000017
Firmware Date          Mon Aug 10 00:00:24 UTC 2015
Bootloader             s1

Device: Flame KK v2.5(Pass):
Build ID               20150810150206
Gaia Revision          9a8880a95ee4a4aea7895d4e2bcab31bc49ea281
Gaia Date              2015-08-10 16:01:11
Gecko Revision         https://hg.mozilla.org/mozilla-central/rev/8cba870a352c
Gecko Version          43.0a1
Device Name            flame
Firmware(Release)      4.4.2
Firmware(Incremental)  eng.cltbld.20150810.183734
Firmware Date          Mon Aug 10 18:37:46 EDT 2015
Firmware Version       v18D v4
Bootloader             L1TC000118D0