mediar-ai
automatic device switching
[2024-08-14T16:41:54Z ERROR screenpipe_audio::core] An error occurred on the audio stream: The requested device is no longer available. For example, it has been unplugged.
e.g. i turned off my airpods
kinda tricky, what if someone does want to use specific devices?
seems like macos audio output audio becomes "disconnected" when closing the lid also:
[2024-08-20T09:26:04Z INFO screenpipe_server::core] Finished record_and_transcribe for device Display 1 (output) (iteration 68)
[2024-08-20T09:26:04Z ERROR screenpipe_server::core] Error in record_and_transcribe for device Display 1 (output) (iteration 68): Audio device not found, stopping thread
[2024-08-20T09:26:04Z INFO screenpipe_server::core] Exiting audio capture thread for device: Display 1 (output)
we should something so that when lip is up again it restart recording
also learning:
i was recording iphone mic and that prevented me from listening to music (but i tihnk we cant do much here, its the native thing of apple) should prob implement an app at some point for more control
should also work with monitor i guess?
say you work somewhere w second screen, unplug and go somewhere else, then it should automatically have swiched? (atm would crash)
not sure how the UX would be
maybe user say
"always record these devices: x y z user can select even if disconnected now" and should only record when connected
It would be great if you could dynamically switch to whatever is the active sound source
@m13v been thinking about this
my use case i think would be fine:
- say i want to record macbook mic + display + headphone + iphone whenever connected (i dont think audio uses lot of storage / compute anyway)
- it would just poll for audio device every seconds and when a device is unplug just stop transcribing it and when another one in my list comes up it would start recording it
i think this is the best balance of UX / easiest to implement (prob 1-2h for me)
(likely will be useful for use case where they record from server too)
sounds good, in the future we can listen to notifications instead from the OS libraries
On Tue, Nov 12, 2024 at 4:28 PM Louis Beaumont @.***> wrote:
@m13v https://github.com/m13v been thinking about this
my use case i think would be fine:
- say i want to record macbook mic + display + headphone + iphone whenever connected
- it would just poll for audio device every seconds and when a device is unplug just stop transcribing it and when another one in my list comes up it would start recording it
i think this is the best balance of UX / easiest to implement (prob 1-2h for me)
(likely will be useful for use case where they record from server too)
— Reply to this email directly, view it on GitHub https://github.com/mediar-ai/screenpipe/issues/170#issuecomment-2472012070, or unsubscribe https://github.com/notifications/unsubscribe-auth/AY62CDG4Q2IBJXNGZOUMTUT2AKML7AVCNFSM6AAAAABMQXUC22VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINZSGAYTEMBXGA . You are receiving this because you were mentioned.Message ID: @.***>
