pause media during transcription

[emmgr23]

I think it would be good to have an option to enable or disable the "pause of any media during transcription". I use Voice Ink extensively and also the open-source application FluidDVoice. They are not perfect, but each has its advantages. This one is also very promising.

[vladborovtsov]

Perhaps this should exactly be an "option". MacOS offers "voice isolation" feature. It works really really well when turned on. I have zero issues using type whisper while listening to music or some YouTube video.

[SeoFood]

Thanks for the suggestion! We actually had media pause/resume implemented in an earlier version but had to remove it because of macOS limitations:

• MRMediaRemoteSendCommand (the private API for media control) doesn't reach Chromium-based browsers (Chrome, Brave, Arc, etc.) - so YouTube, Spotify Web, etc. wouldn't pause
• Simulating the media play/pause key (NX_KEYTYPE_PLAY) as a workaround would often open Apple Music instead of pausing the actual audio source

There's unfortunately no reliable, universal way to pause "whatever is currently playing" on macOS.

As @vladborovtsov mentioned, macOS Voice Isolation (System Settings → Sound → Microphone Mode → Voice Isolation) is the best alternative - it does a great job of filtering out background audio so the transcription stays clean even with music or videos playing.

[emmgr23]

Try looking at Voice InK or Fluid Voice. The media pause works perfectly on both of these applications and in any web browser.

[vladborovtsov]

Actually, @SeoFood I even tried to implement voice isolation switch inside the app ,but found that its private api and macOS/iOS both designed to let user choose it on system level. So I gave up on this.

On iOS there are different modes available for audio session, including ducking. https://developer.apple.com/documentation/avfaudio/avaudiosession/categoryoptions-swift.struct/duckothers

I dunno if there is equivalent option for MacOS. So instead of pausing other media sources, maybe it's possible to make them less loud utilizing the OS capability.

[SeoFood]

This is now implemented in PR #160 and will be available in the next daily build for testing.

Toggle is in Settings > Recording > Media Pause. Known limitation: only targets the active "Now Playing" session, so if multiple apps are playing simultaneously, only one will be paused/resumed.

[emmgr23]

awesome... That was quick 😂

[SeoFood]

This is being worked on in #163. The implementation uses mediaremote-adapter (same library used by FluidVoice and VoiceInk) to reliably detect play state and pause/resume media during recording.

Will be available in the next daily build for testing once merged.

Known limitations:

• Only works with apps that register with macOS system media controls (most media players and browsers)
• When multiple media sources play simultaneously, only the active Now Playing session is paused
• Not available in App Store builds (uses private MediaRemote.framework)