Gain applies a chosen level change to the whole signal. Peak normalization chooses a gain from the highest peak. Loudness normalization measures a loudness target. AudioCut’s Volume Changer applies fixed gain; it does not match files to a LUFS target.
Fixed gain: choose the amount yourself
A +3 dB gain raises every part of the signal by the same ratio. Speech, breaths and background noise all rise together. A predicted peak helps you decide how much room remains before full scale.
For a source peaking at −6 dBFS, +3 dB predicts a peak near −3 dBFS. Adding +9 dB instead would exceed full scale. AudioCut clips samples that exceed the output range, so lower the gain if you want to avoid introducing that distortion.
−6 dBFS + 3 dB = −3 dBFS
The same adjustment raises quiet passages and background noise too.
Peak and loudness targets answer different questions
Peak normalization bases its adjustment on the largest peak. Two recordings can reach the same peak and still sound different in overall loudness: a single transient can dominate the measurement.
Loudness normalization considers a loudness measurement and a target. It is a different workflow from choosing +3 dB by hand. A platform’s delivery requirements may include a loudness target and peak constraints; use a tool that actually measures those requirements.
| Method | How gain is chosen | Fits this task |
|---|---|---|
| Fixed gain | You choose a dB adjustment | Make one recording louder or quieter |
| Peak normalization | From a peak and a target | Bring the highest peak to a chosen level |
| Loudness normalization | From measured loudness and a target | Match an applicable loudness requirement |
What to check in AudioCut
Adjust the gain between −12 and +12 dB, inspect the predicted peak and listen to a loud passage. Also check quiet passages: increasing gain may reveal distracting background noise.
If your job is matching a set of episodes to a specified loudness level, do not assume a safe peak proves that target is met. AudioCut’s peak display is a level warning, not a loudness compliance measurement.
Questions about this task
Does a safe peak mean every passage is equally loud?
No. It says nothing about matching the quiet and loud parts to one another or matching separate recordings.
Does lowering gain repair clipped source audio?
It lowers the level of the distortion too, but does not reconstruct the missing waveform detail.