How to deliver at the loudness a platform expects
Tell Vizard Agent the loudness target and the peak ceiling you deliver to. It measures the speech and the music separately, cross-checks the speech figure a second way before relying on it, normalises the master to your number, and applies the same treatment to every episode in the set.
What do you type into Vizard Agent?
Give Vizard Agent both numbers rather than one. An average level with no peak ceiling lets a single loud moment clip the delivery, and a peak ceiling with no average target produces a file that is technically safe and audibly quiet in a feed. The numbers come from wherever you deliver to, so they're yours to supply.
Prompt
What does Vizard Agent actually do?
Platforms turn your video up or down to their own level, so hitting the number yourself keeps control of the sound. But loudness measurement depends on what you include, such as a musical tail, a silent gap or an intro, so the speech figure is cross-checked before a whole series is mastered to it. In the session behind this article it:
- Measured the loudness of the music on its own, then of the speech and music together, and mixed the first segments of a branded podcast series.
- Measured the speech loudness a second way to cross-check the first figure, then mixed with proper loudness control rather than a flat gain.
- Checked the peak levels against the ceiling and fixed the peaks that exceeded it.
- Mixed all eight segments to the same settings, then balanced a new speaker who sat differently in the mix and re-checked the levels.
- Measured the loudness again before uploading, then mastered the audio and muxed the final cut.
When does this not work well?
Normalising to a number does not improve a recording; it only sets its level. Where the source is noisy or already damaged, hitting the target makes those problems louder, and Vizard Agent will say so rather than quietly deliver a louder fault. Already-clipped audio keeps its distortion, and a wrong mix needs balancing before it is mastered.
Common questions
What number should I use? Whatever your destination specifies. Vizard Agent cannot know your spec.
Why measure speech separately? Because a music-heavy section can carry the average while the voice sits too low. Vizard Agent measures each on its own.
What if two measurements disagree? Vizard Agent tells you and says which one it used.