-
Notifications
You must be signed in to change notification settings - Fork 128
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
cubeb.h: Document device switching policy. #582
base: master
Are you sure you want to change the base?
Conversation
Following the discussion in mozilla/cubeb-pulse-rs#54 and elsewhere, document the interaction between cubeb_stream_init parameters and cubeb's intended device switching policy.
CUBEB_STREAM_PREF_DISABLE_DEVICE_SWITCHING on @ref | ||
cubeb_stream_params.prefs. | ||
|
||
@note If @p output_device or @p input_device is a valid non-NULL @ref |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ah, it seems we do support output_device
selection, but I never tested it.
If the stream needs to be reinitialize to default output device when the specified output_device
is no longer available, then we do need to listen kAudioDevicePropertyDeviceIsAlive
for output device like what we do for the input side.
On the other hand, if the stream using the specified output_device
doesn't need to be reinitialized when the specified output_device
is no longer available, it seems audiounit-rust
meets the spec here.
Anyway, my problem is: what is the intended behavior when the specified output_device
or input_device
is unplugged while they are being used by some active streams?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Anyway, my problem is: what is the intended behavior when the specified
output_device
orinput_device
is unplugged while they are being used by some active streams?
This is in flux at the minute. The current behaviour we want in gecko:
- If a output device has been opened with an explicit device or a default device, and is no longer available, switch to the new default device
- If an input device has been opened, it's always with an explicit ID. It it is not available anymore, turn this stream into a output-only stream
When @achronop finishes setSinkId
, we'll reevaluate this (or as part of his effort).
I had a design a long time ago that would allow enough flexibility to implement whatever we need: https://github.com/kinetiknz/cubeb/issues/167. I think it still works
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think in general the behaviour @kinetiknz documents here is simpler and works just as well.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If a output device has been opened with an explicit device or a default device, and is no longer available, switch to the new default device
If I am not missing something the output device is always non-explicit at the moment in Gecko so @kinetiknz 's document covers our needs. This will be changed when sinkId is in place but we can re-evaluate then. I have already some thoughts about it.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think my question is here:
If @p output_device or @p input_device is a valid non-NULL @ref cubeb_devid, cubeb will only ever use the specified device and never follow device changes automatically
We do change the output device if the specified output device
is no longer available but the document states: cubeb never follow device changes automatically.
I am ok to reword these once setSinkId
is done. Just want to make sure this issue is one someone's radar.
|
||
@note If @p output_device or @p input_device is a valid non-NULL @ref | ||
cubeb_devid, cubeb will only ever use the specified device and never | ||
follow device changes automatically. In this case, the application may |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
audiounit-rust
backend will automatically change the input device when the specified input_device
is unplugged (cannot be used anymore). So this behavior need to be changed? Should we just return error instead when the specified input_device
is gone?
In addition to the above, we need to document what happens when an explicit-defined device is not available anymore. In rust PA this situation results in a state change to |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The current behavior is slightly different from the updated descriptions. I am ok to reword these after setSinkId
is done or update our code to meet the descriptions.
CUBEB_STREAM_PREF_DISABLE_DEVICE_SWITCHING on @ref | ||
cubeb_stream_params.prefs. | ||
|
||
@note If @p output_device or @p input_device is a valid non-NULL @ref |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think my question is here:
If @p output_device or @p input_device is a valid non-NULL @ref cubeb_devid, cubeb will only ever use the specified device and never follow device changes automatically
We do change the output device if the specified output device
is no longer available but the document states: cubeb never follow device changes automatically.
I am ok to reword these once setSinkId
is done. Just want to make sure this issue is one someone's radar.
Any thoughts regarding the state change (to error) when the explicit device goes unplug? PA exhibits this behavior, at the moment, and I would like to change it if it is something that we don't want. |
Following the discussion in mozilla/cubeb-pulse-rs#54
and elsewhere, document the interaction between cubeb_stream_init parameters
and cubeb's intended device switching policy.