How to apply this guide
Separate a detected USB device from a working capture signal using mute, gain and a control recording.
Start with
Unmute any inline or desktop switch and speak into the microphone side identified by the maker. Check input volume for the device in macOS. Use the displayed device name, especially with a headset that has separate input and output profiles.
Check before you finish
If both Sound and a separate recorder stay flat, the speech model cannot fix it. If the other recorder works and Utterwove does not, check its permissions and chosen input.
Evidence to keep: Record the expected sentence, actual result, microphone, language, model and app build before changing settings. This page’s key limit is: Separate a detected USB device from a working capture signal using mute, gain and a control recording.
Start here
A microphone can be listed in Sound Input and still deliver silent samples. If the input bar stays flat while you speak, troubleshoot the signal before adjusting the dictation language or model.
Check the physical controls
Unmute any inline or desktop switch and speak into the microphone side identified by the maker. Check input volume for the device in macOS. Use the displayed device name, especially with a headset that has separate input and output profiles.
Try a simple connection
Connect directly instead of through a hub, then reselect the input. If Sound shows signal, use the recorder’s microphone test. Record at a normal distance and verify the meter moves during speech, not just when tapping the desk.
Keep the finding narrow
If both Sound and a separate recorder stay flat, the speech model cannot fix it. If the other recorder works and Utterwove does not, check its permissions and chosen input.




