How to fix uneven levels without the voice sounding processed
Give Vizard Agent the levels you need and say the voice must still sound like the person. Levelling works by pushing quiet parts up and loud parts down, and pushed far enough that makes any voice sound gritty, so the check that matters is a comparison against the original, not a meter reading.
What do you type into Vizard Agent?
Say "cosmetic" and mean it literally. Asking Vizard Agent for professional-sounding audio invites the entire processing chain, such as compression, de-essing and an exciter, when what you want is the same voice with only its level problems taken out. Give the numbers too: a peak ceiling, and a floor for the quiet passages.
Prompt
What does Vizard Agent actually do?
Levelling applied to the whole file has to be aggressive enough for the quietest moment. Applied only where the speech is quiet, it lets everything else be left alone. And the test is not whether the file meets the target, but whether the voice still sounds like the person when the two are played back to back. In the session behind this article it:
- Rendered a revision of a talking-head clip with broader cuts and more room at the speech boundaries, and checked every speech join and the last seconds for clipped words.
- Found the ending left the story hanging and re-cut to a natural close, confirming the audio stayed below the peak limit and the last word had room to finish.
- Measured the quiet speech passages and checked a phrase boundary, so removing one word would not clip the start of the next.
- Applied speech-aware levelling with extra attention to the quiet passages, then checked the spoken peaks after the lift.
- Compared the before-and-after waveform, then the original voice with the processed audio, to judge how much peak control was actually needed.
When does this not work well?
Some recordings cannot be levelled gently. If the original is clipped, recorded far from the microphone, or has a noise floor close to the speech, the processing needed to make it even is the processing that makes it sound processed. Two voices at different distances cannot share one treatment, and aggressive targets force hard processing.
Common questions
Can it remove mic pops? Yes, and Vizard Agent treats that as a cosmetic fix rather than a change to the voice.
Will it use compression? Vizard Agent uses as little as it can when you ask for a natural voice.
Is a re-record better? For clipped audio, yes. Vizard Agent will tell you when it is.