How to Fix a Clipped Interview Recording: Recovering Distorted Audio

You recorded an important interview — a key subject for a documentary, a newsworthy source, an important client meeting — and the audio clipped. The signal hit the maximum level and the waveform is flat-topped, producing that characteristic harsh distortion that makes voices sound like they're being shouted through a broken speaker.

Can you fix clipped audio? Sometimes, to a useful degree. Here's an honest guide to what's possible, what tools work, and when to accept limitations.

What Clipping Actually Does to Audio

Digital clipping occurs when a signal exceeds the maximum representable level (0 dBFS). The waveform, which should naturally curve at peaks, is instead cut flat — "clipped." The flat tops represent lost information that was never recorded.

Two types of clipping:

• Occasional brief clips: Short peaks that hit 0 dBFS for a few milliseconds. These appear as isolated flat tops in the waveform. Most recoverable.
• Sustained heavy clipping: Extended sections where the gain was simply too high. Long stretches of flat-topped waveform. Limited recovery possible; severely clipped sections may be unrecoverable.

What clipping sounds like:

• Harsh, buzzy distortion on louder passages
• Characteristic "crunchy" quality on consonants and fricatives (S, T, P sounds)
• Loss of natural voice character — speech sounds tense and unnatural
• In severe cases, speech becomes completely unintelligible

How De-Clipping Tools Work

De-clipping software attempts to reconstruct the "true" shape of the waveform using the audio on either side of the clipped region.

The algorithm:

1. Identifies clipped regions (flat tops in the waveform)
2. Analyzes the shape and level of audio approaching and leaving the clip
3. Mathematically extrapolates what the missing peak likely looked like
4. Fills in the reconstructed waveform

This works well for brief clips because a short clipped transient has predictable waveform behavior — peaks tend to follow smooth, curved shapes. The reconstruction is plausible.

For extended clipping, there's simply too much missing information. The algorithm is guessing over longer periods, and results become unreliable.

Method 1: iZotope RX Declip (Best Available Tool)

iZotope RX's Declip module is the professional standard for clipped audio recovery. It's significantly more effective than any free alternative.

Steps:

1. Open the clipped recording in RX
2. Select the module Declip (under the Repair section)
3. Run the Analyze function first — RX detects clipped regions and shows statistics
4. In the Settings:
   • Threshold: Where RX considers samples "clipped" — start at -0.3 dBFS
   • Quality: Higher quality = better results but slower processing; use Maximum for important recordings
5. Preview a clipped section before applying to the full file
6. Apply

Realistic results with RX Declip:

• Brief clips (under 10ms): Often fully transparent recovery — you won't hear where the clip was
• Moderate clips (10-50ms): Good improvement, reduced distortion; may still sound slightly processed
• Heavy extended clipping: Noticeable improvement in intelligibility; distortion character changes but doesn't fully disappear
• Severe sustained clipping (seconds of flat tops): Limited improvement; underlying distortion quality changes but remains

Method 2: Audacity Clip Fix (Free Alternative)

Audacity includes a basic clip fix function:

1. Select the clipped region (or the whole file)
2. Effect > Clip Fix
3. Set Percent Clip to Restore: 95-100%
4. Apply

Audacity's algorithm is simpler than iZotope RX and produces noticeably less natural results, but it's free and better than nothing for minor clipping.

Method 3: Manual Waveform Drawing (For Individual Clips)

For isolated, clearly visible peaks in the waveform view:

1. Zoom in until the clipped region fills the screen
2. Use Audacity's Draw tool (pencil icon)
3. Manually redraw the waveform to create a natural-looking curved peak
4. This is tedious but can produce transparent results for isolated very brief clips

This approach is practical only for a handful of isolated clips — not for recordings with extensive clipping throughout.

Processing After Declipping

Even after declipping, processed audio often benefits from additional treatment:

EQ correction:
Clipping introduces harmonic distortion that adds harsh high-frequency content. After declipping, a gentle high-frequency roll-off (2-3 dB shelf cut above 8 kHz) can reduce residual harshness.

Light compression:
Compression after declipping evens out level inconsistencies that appear when loud peaks are reconstructed at lower levels than the original clip event.

De-noise:
Some declipping algorithms introduce subtle noise artifacts. A light noise reduction pass after declipping can clean these up.

Honest Assessment: What Can and Can't Be Fixed

Fully recoverable:

• Brief isolated clips on peaks where surrounding audio is clean
• Recording that's slightly over 0 dBFS by 1-2 dB
• Single-word or syllable clips in otherwise clean speech

Significantly improved:

• Moderate clipping where the recording is otherwise usable (SNR is good, just level too high)
• Recordings where the content is clearly intelligible despite distortion

Partially improved but not fully fixed:

• Extended heavy clipping throughout an interview
• Simultaneous clipping plus other problems (background noise, room echo)

Very limited recovery:

• Severely distorted passages where the waveform is nearly entirely flat
• Recordings where the clip level was so far above the limit that harmonic distortion is deeply embedded

For an important interview with moderate clipping, professional declipping at WefixSound is often worth pursuing. The free 60-second sample will tell you honestly what's achievable — upload the most distorted section and see the result before committing.

Prevention: How to Avoid Clipping

Set gain appropriately:

• For interviews, aim for average speaking level at -18 to -12 dBFS
• Leave 12-20 dB headroom for louder moments
• Test levels before beginning with the actual subject, not just yourself

Use a limiter:
A hardware or software limiter prevents signals from exceeding a set level. Set a true-peak limiter at -3 to -6 dBFS as a safety net during recording.

Record at 24-bit:
24-bit recording has much greater headroom than 16-bit. Even if signals go slightly above your target level, you have more room before real distortion occurs.

Monitor while recording:
Using headphones and watching levels during an interview lets you catch and adjust gain before a brief clip becomes an extended problem.

Use a backup recorder:
For critical interviews, record on two devices simultaneously — one at normal gain and one 12 dB lower as a safety net. If the main recording clips, the backup preserves the content.

Related Articles

• How to Fix Audio Clipping and Distortion
• Podcast Audio Cleanup Guide
• Audio Restoration Service: What to Expect

Clipped interview audio is frustrating but often improvable with the right tools. Brief clips recover cleanly; extended heavy clipping presents real limitations. For important recordings, WefixSound's professional declipping service applies the best available tools to maximize intelligibility from every second of your recording.