Time-coded degradation evidence
Clicks, short clipping runs and dropouts sit on the same timeline as full-file DC, hum-family and high-frequency noise measurements.
Complete restoration workflow · local processing
Old Audio Restoration Studio
Turn a vinyl, cassette, reel or field transfer into an explainable restoration. Map common defects, review every time-coded repair, compare Source / Restored / Difference and keep the untouched recording.
No uploadNo fixed 200 MB workflow capExact frame count preservedOpen WAV + reports
- 01Analyze transfer
- 02Review & restore
01 · Build the evidence
Map what changed in the recording
Presets tune detector bias. Nothing is repaired until you review the recipe.
Audio remains on this device
Visual degradation map
Waveform evidence plus aligned defect lanes.
Analysis will place time-coded issues hereUse the deterministic demo to see clicks, a dropout, clipped samples, DC offset and a 60 Hz hum family.
The canvas is supplementary. After analysis, the adjacent marker list contains the same time, type, severity and confidence information for keyboard and screen-reader review.
Analyzing decoded PCM…
02 · Listen before changing
Review markers and build the restoration
Keep safe candidates, ignore musical transients and render one aligned result.
Select a marker
Choose evidence from the review queue.
Detector evidence—
Proposed operation—
| Metric | Source | Restored |
|---|---|---|
| RMS | — | — |
| Peak | — | — |
| Maximum DC | — | — |
| Detected markers | — | — |
Review evidence before renderingThe original decoded source is immutable.
RESTORATION MAP
Reviewed Old-recording Restoration
Separate the damaged event from healthy context. A repair is safest when the affected span is narrow, the surrounding signal is trustworthy and the difference audition contains mostly the defect.
Damaged vinyl or tape transferMap, review and attenuate defectsRestored master and receipt
Open the full signal breakdownBest source, processing logic, limits and three-pass listening check
- Damaged vinyl or tape transfer
Start with the right material
Reviewing a complete old recording when clicks, clipping, short dropouts, DC offset, mains hum, rumble and high-frequency noise need one coordinated workflow rather than unrelated passes.
- Map, review and attenuate defects
Understand the transformation
The studio measures global and time-coded evidence, applies only accepted short-region interpolation, then runs bounded DC, hum, rumble, hiss and output-protection stages while retaining Source, Restored and Difference audition.
- Restored master and receipt
Verify the usable result
An exact-frame lossless WAV restoration master, optional MP3 listening copy, Difference WAV and JSON, CSV and standalone HTML evidence packaged locally.
Interpolation is not recovery
Where this workflow stops
Interpolation cannot recover unknowable missing samples, aggressive reduction can erase wanted detail and the workflow does not equal specialist spectral restoration, mechanical alignment or archival engineering. Keep the raw transfer and listen to every Difference.
THREE-PASS CHECK
Make the final decision by ear
- Review every detected event near drums, consonants and other sharp attacks.
- Use Difference audition to hear what the repair removes or invents.
- Prefer several small repairs over one aggressive global pass.
Quick answer: AudioWrench scans an old recording for common measurable defects, puts time-coded candidates on one map and applies only the repair markers and bounded global stages you keep. It preserves duration and provides Source, Restored and Difference audio; it does not claim to reconstruct an unknowable original.
One complete restoration desk
Keep or ignore every local repair
Fast drums and consonants can resemble defects. Confidence is evidence for listening, never permission to destroy a transient silently.
A medium-aware signal chain
Vinyl, cassette, reel and field presets choose detector bias and conservative starting strengths while every stage remains visible.
Source, Restored and Difference
The aligned Difference monitor makes hum attenuation and interpolated samples audible instead of hiding the transformation behind a score.
Exact timeline and untouched source
Short repairs replace samples in place. The restoration keeps channel count, sample rate and frame count and never overwrites your file.
Open archival receipt
Export lossless WAV, a listening MP3 and JSON, CSV and standalone HTML evidence in a package you can keep beside the raw transfer.
How to restore vinyl, cassette and tape audio
- Add the untouched transfer.
Use the highest-quality browser-decodable capture available. Keep that raw file separately before any repair.
- Choose the physical medium.
The preset changes detector sensitivity and starting strengths; it does not process audio or claim to know the recording history.
- Read the map and listen.
Jump to severe sections first, compare detector confidence and ignore any musical transient incorrectly marked as damage.
- Render a bounded recipe.
DC correction runs before short-region interpolation, then hum, rumble, conservative hiss attenuation and output protection.
- Inspect the Difference and receipt.
Confirm that removed energy sounds like damage, keep the lossless master and document every setting and accepted marker.
What each detector and repair actually does
| DC offset | Measures the arithmetic mean of each decoded channel. The correction subtracts that mean; it does not alter timing or infer missing audio. |
|---|---|
| Clicks and crackle | Finds unusually abrupt impulses that return toward local context. Accepted short regions are replaced by a smooth interpolation between surrounding samples. |
| Clipping | Looks for short near-flat runs close to full scale. Only bounded regions are interpolated; severe or long distortion needs a specialist spectral repair. |
| Dropouts | Finds short near-silent runs surrounded by active audio. The local repair bridges endpoints and cannot recreate the content that was never captured. |
| Mains hum | Compares 50 and 60 Hz harmonic families with nearby frequency evidence. Strong suggestions use narrow mixed notches rather than a broad tonal cut. |
| Hiss and rumble | Reports high-frequency energy as a candidate, then offers bounded high-band attenuation and a simple subsonic high-pass. Both require listening. |
Choose the right starting preset
Vinyl record
Biases the map toward short impulsive damage and subsonic rumble while keeping high-frequency attenuation gentle.
Cassette tape
Starts with stronger hum and hiss controls and keeps dropout review visible. It cannot correct a badly aligned tape deck after capture.
Reel-to-reel
Uses conservative broadband settings for transfers where preserving transients and room character matters more than silence.
Field or video transfer
Raises rumble and mains-hum attention for portable recorders, VHS audio and imperfect location power.
Shellac and damaged discs
Start with Vinyl, increase click sensitivity carefully and review every transient. Wide groove noise often needs a specialist spectral tool.
Valuable archival masters
Use the Studio as an evidence map and receipt. Mechanical alignment, azimuth, cleaning and specialist transfer decisions belong before DSP.
Where browser restoration stops
No lost sample recovery
Interpolation estimates a plausible bridge. It cannot know the waveform removed by clipping, dropout, scratches or damaged tape.
No universal wow correction
Reliable speed repair needs a stable tonal reference and strong coverage. Use the focused Wow & Flutter Corrector when the recording provides suitable evidence.
No automatic artistic judgment
Surface noise can be part of the document, and aggressive noise reduction can erase ambience, consonants and cymbal detail. Difference listening is mandatory.
Restoration workflow notes
Keep an untouched raw transfer before processing. AudioWrench keeps repair decisions visible and provides Difference audition so you can hear what changed. For difficult spectral damage, unstable pitch, long dropouts, mechanical transfer problems or valuable masters, use a specialist restoration service.
Old Audio Restoration Studio: quick answer and technical limits
Quick answer: A complete local restoration desk that maps common vinyl, cassette and tape-transfer defects, keeps every repair decision visible and exports an aligned result with an open archival receipt.
- Best for
- Reviewing a complete old recording when clicks, clipping, short dropouts, DC offset, mains hum, rumble and high-frequency noise need one coordinated workflow rather than unrelated passes.
- How it works
- The studio measures global and time-coded evidence, applies only accepted short-region interpolation, then runs bounded DC, hum, rumble, hiss and output-protection stages while retaining Source, Restored and Difference audition.
- What you get
- An exact-frame lossless WAV restoration master, optional MP3 listening copy, Difference WAV and JSON, CSV and standalone HTML evidence packaged locally.
Know before you use it: Interpolation cannot recover unknowable missing samples, aggressive reduction can erase wanted detail and the workflow does not equal specialist spectral restoration, mechanical alignment or archival engineering. Keep the raw transfer and listen to every Difference.
Privacy: Selected audio is processed in this browser and is not uploaded to AudioWrench. Normal page assets can still be requested as described in the privacy notice. Privacy details →
FAQ
Does AudioWrench upload my old recording?
No. Audio is decoded, analyzed, restored, auditioned and encoded locally in the current browser tab. Normal page assets can still be requested.
Can software recover the exact original recording?
No. Click, clipping and dropout repairs interpolate missing or damaged samples; hum and noise stages attenuate measured energy. Difference audition shows what changed.
When should I use a specialist restoration service?
Use specialist help for difficult spectral damage, unstable pitch, long dropouts, mechanical transfer problems or valuable masters that need archival handling.
What does the degradation map measure?
It maps time-coded impulse, flat-top clipping and short dropout candidates, plus full-file DC, hum-family and high-frequency noise evidence.
Why review clicks manually?
Fast musical transients can resemble clicks. The studio protects the source, exposes confidence and lets you ignore any candidate before rendering.
What should I export for an archive?
Keep the untouched transfer and export a lossless WAV restoration master plus the JSON, CSV or HTML receipt. MP3 is best treated as a listening copy.