HomeFeaturesDetect Audio Problems
Detect Audio Problems

Every Audio Problem
Found in One Pass.

A single analysis pass detects clipping, loudness that's too low, background noise, poor signal-to-noise ratio and A/V sync drift. Every issue timestamped and ranked so you fix the right things first.

Premiere Pro Final Cut Pro DaVinci Resolve 5 audio detectors
Four Issue Types

Every audio problem that ruins a deliverable.

Audio Clipping

Sections where the audio waveform has hit 0 dBFS and the peaks have been hard-limited, creating harsh digital distortion. The detector measures peak levels and inter-sample peak (ISP) values. Reports the number of clipped samples, maximum peak level and a clipping density score per section.

Loudness Too Low

The integrated LUFS value for the section falls below the recommended target for the configured delivery platform. YouTube targets –14 LUFS, broadcast –23 LUFS, Spotify –14 LUFS. Sections that are too quiet are annoying for viewers who set their volume for your dialogue and then have to turn it up for other content.

Background Noise and Poor SNR

The signal-to-noise ratio between the dialogue signal and the background noise floor falls below the acceptable threshold. The detector estimates the noise floor during silent sections and the signal level during speech, computing the SNR in dB. Sections with poor SNR are flagged for noise reduction treatment in post.

A/V Sync Drift

When audio and video tracks in a multi-camera or multi-track recording drift out of alignment. The detector cross-correlates the audio waveform with the reference track to measure drift in milliseconds. Even a 150ms drift is noticeable to viewers as a visible lip-sync mismatch.

What the Numbers Mean

LUFS, dBFS and SNR explained for editors.

Audio measurement has its own vocabulary that not all video editors are familiar with. EditLint AI reports numbers but also tells you what they mean in plain language for your specific delivery target.

  • dBFS (decibels full scale) — measures peak levels. 0 dBFS = maximum before clipping. Target peaks: –3 to –6 dBFS for most delivery.
  • LUFS (loudness units full scale) — measures perceived loudness over time. YouTube targets –14 LUFS integrated. Broadcast (EBU R128) targets –23 LUFS.
  • LRA (loudness range) — measures dynamic range. High LRA means dynamic content with quiet and loud sections. Target 8–15 LU for most programme material.
  • SNR (signal-to-noise ratio) — dB difference between dialogue signal and background noise. Below 20 dB is problematic. Above 40 dB is broadcast quality.
  • Sync drift in ms — how many milliseconds the audio is ahead of or behind the video. Visible to viewers at +/–150ms.
Delivery Platform Targets
YouTube –14 LUFS · –1 dBFS peaks
Broadcast (EBU R128) –23 LUFS · –3 dBFS peaks
Spotify / Podcast –14 LUFS · –1 dBFS peaks
Apple TV+ –27 LUFS · –2 dBFS peaks
Netflix –27 LUFS · –2 dBFS peaks
How It Works

One pass. All five detectors.

1

Audio Extraction

The audio tracks from your sequence are decoded and buffered in memory for analysis. For multi-track sequences, each track is analysed independently so you can see which track has which problem. All five audio detectors run in a single pass through the decoded audio data.

2

Per-Section Analysis

The audio is analysed in rolling windows. Peak levels, integrated LUFS, LRA, ISP, SNR estimate and sync correlation are all computed per window. Windows where any metric falls outside the configured thresholds for the minimum duration are recorded as candidate issues.

3

Issue Log and Markers

Each detected issue is logged with its type, measured value, threshold, severity and a plain-language suggested fix. Markers are placed on the audio track at the start of each section. The Generate Report feature includes every audio issue in the PDF and CSV output.

Audio problems found. Before the client hears them.

Every clipped section, every noisy room, every sync drift logged and ranked in under 60 seconds. Fix the right things first.

No credit card required · Works inside your NLE · Cancel any time