Release notes
Videorc 0.9.78 Beta 1
Unplugging a microphone mid-recording used to end the whole session. Now the recording keeps going, the audio track is padded with silence from the moment the mic vanished, and a warning stays on screen explaining what happened — so you finish the take and keep the video. Recordings also survive brief encoder backpressure that could previously cut a session short.
A recording no longer dies when the microphone does
- Losing a microphone mid-session no longer stops the recording — it keeps rolling, pads the audio with silence, and tells you what happened.
- The warning about a lost microphone stays put and survives a reconnect, so you are not left guessing why the audio went quiet.
- Short bursts of encoder backpressure no longer end a recording that was otherwise healthy.
Fixed
- Microphone loss no longer ends the session. Native microphone EOF or a
stall was indistinguishable from an intentional stop or a closed pipe, so
FFmpeg treated it as end-of-session. The three cases are now classified
separately: a lost microphone pads the audio track with silence until the
video ends, recording and streaming continue, and one durable,
session-correlated
microphone-input-lostwarning is raised that survives renderer reconnects. - Transient encoder backpressure no longer kills a recording. The FIFO writer reused an access unit's queue timestamp as its delivery deadline, so a complete, still-valid frame could arrive with no time left and terminate the session. Each complete access unit now gets a fresh, bounded write window.
Known limitation
FFmpeg-owned microphone fallbacks (the macOS AVFoundation fallback and Windows
DirectShow) receive the silence-padding policy but do not yet use the native
source-loss monitor, so a clean end-of-stream there raises no
microphone-input-lost event, and a fatal DirectShow driver error can still
end the FFmpeg process. The macOS CoreAudio native path — the default — is
fully covered. Windows was not verified on physical hardware for this change.