MikeNet / Soundcheck
Hear the noise. Change one thing.
A quiet opening gives you a reference. The words that follow tell you whether the fix also changed your voice.
Make a recording with room tone.
During the first three seconds of MikeNet’s soundcheck, stay quiet and leave the microphone in its normal position. Keep the keyboard, desk and cable still. Then read the prompt normally. This captures both the background and your voice under the same settings.
Listen at a comfortable, fixed playback volume with headphones. Silence does not need to look perfectly flat to be usable. Turning playback far above your normal listening level can make a minor background sound seem like the main problem.
Match the change to what you hear.
| What you hear | A useful next comparison |
|---|---|
| Fan, traffic or other room sound | Keep gain fixed. Move the mic closer to your voice, or remove one source of background sound where practical. |
| Steady hiss | Compare a modest gain reduction with closer placement. Keep a record of each separate change. |
| Steady electrical-sounding hum | Check cable seating and compare an approved alternative cable or USB connection. Do not defeat electrical grounding. |
| Hollow or reverberant speech | Compare a closer microphone position or a less reflective part of the room. |
| Words cut off, watery sound or pumping | Compare capture processing settings, then check the equivalent controls in your actual calling app. |
These are test directions, not diagnoses. A recording alone cannot identify which cable, circuit, microphone component or software stage caused a sound.
Compare the voice and the quiet section together.
Suppose A has room tone at −50 dBFS and speech at −25 dBFS. B has room tone at −60 and speech at −35. Both dropped 10 dB; the speech-to-room difference is still 25 dB. B is quieter, but those numbers alone do not show better separation.
If speech stays at −25 while room tone falls to −60, the difference rises to 35 dB. That is useful evidence to listen alongside the recordings. It is not a laboratory signal-to-noise ratio: you made the quiet and speaking measurements at different times, with a real room and a voice that can vary.
Noise suppression has a tradeoff.
A browser may apply noise suppression, echo cancellation or automatic gain. They can help a call but also change the sound you are evaluating. MikeNet shows the settings the browser actually reports; a request to turn processing off is not proof it turned off everywhere in the device or operating system.
A very low quiet-section reading with missing beginnings of words may be gating rather than a quieter physical room. Listen to soft words and pauses in both takes. Do not choose a setting solely because its number is lower.
Keep the improvement repeatable.
Change one variable, repeat the same phrase and name the change. If the difference is small, record again before deciding. Compare in the app that matters afterward: its processing may differ from the browser’s.
Run a before-and-after soundcheck
Shure’s discussion of direct and ambient sound; MDN on reported capture settings; MDN on echo-cancellation requests. The comparison protocol and dBFS examples are original illustrative calculations, not measured microphone reviews.