Summary

This free tool scores how much AI audio restoration a recording needs, from a light denoise pass to professional-grade work. Pick your source (cassette, vinyl, phone memo, damaged tape), tick the problems you can hear, and set your target quality. The calculator returns a 0-100 difficulty score, a restoration tier, the recommended processing order (declick before denoise, de-reverb last), an estimated processing time, and a starting noise-reduction strength. Nothing uploads: the math runs in your browser.

How much AI audio restoration does your recording need?

Answer six questions about the source and the noise you hear. The score, the processing order, and a time estimate update as you type.

Score my audio

AI audio restoration score

Pick the source, tick the problems you can hear, and set a target quality. Everything runs in your browser: nothing is uploaded.

Problems you can hear
0 /100
Light cleanup

Recommended processing order

    Estimated processing time -
    Suggested noise-reduction strength -

    The score in numbers

    6
    noise types the score checks for, from hiss to dropouts
    4
    restoration tiers, from light cleanup to professional-grade
    0
    seconds of audio uploaded to run this calculator
    How this works

    What goes into the score

    A baseline per source

    Cassette, vinyl, phone memo, field recording, and water- or heat-damaged tape each start from a different severity baseline. A phone voice memo is not the same restoration job as a damaged reel.

    The real processing order

    Impulsive noise like clicks and pops gets treated before broadband hiss. Hum is handled as its own tonal pass. De-reverb runs last, since it is the step most likely to introduce artifacts.

    Target quality changes the plan

    Archival or broadcast targets get a more conservative noise-reduction strength to protect naturalness. Casual listening can take a heavier hand.

    Using the calculator

    Three inputs, one plan

    No account, no upload. The whole thing runs in the widget above.

    1. 1

      Pick the source

      Cassette, vinyl, a phone voice memo, a field recording, an old low-bitrate digital file, or a water- or heat-damaged tape. This sets the difficulty baseline before any problems are added.

    2. 2

      Tick what you actually hear

      Hiss, hum, clicks, clipping, dropouts, reverb: check only what is audible in the file. Guessing at problems that are not there inflates the score and the recommended processing chain.

    3. 3

      Set the target quality

      Casual listening, podcast-ready, or archival and broadcast. This changes the suggested noise-reduction strength, since a heavier hand is fine for a commute but risky for a preservation copy.

    4. 4

      Read the plan

      The score, the tier, the processing order, an estimated turnaround, and a starting noise-reduction percentage update instantly as the answers change.

    Questions people ask before running this

    Is this free, and does it upload my audio file?
    It's free and nothing uploads. The score, the processing order, and the time estimate are all computed in your browser from the answers you tick, not from the audio file itself.
    Where do the difficulty weights come from?
    From standard audio restoration practice: impulsive noise (clicks, pops) is treated before broadband hiss, hum gets its own tonal pass, and de-reverb runs last because it is the step most likely to add artifacts. Source baselines reflect how condition typically varies by format: a damaged tape starts worse than a low-bitrate digital file.
    Why does target quality change the noise-reduction percentage?
    A higher reduction strength removes more noise but risks a processed, unnatural sound. Archival and broadcast work needs that risk kept low, so the suggested strength drops. Casual listening tolerates a heavier pass.
    Can I use this before I buy restoration software or hire someone?
    Yes. That is the point: get a realistic sense of the tier (light, moderate, heavy, or professional-grade) before you commit budget or software licenses to the job.
    Does the score account for how long the recording is?
    Length does not change the difficulty score itself, since difficulty is about condition, not duration. It does drive the estimated processing time, which scales with both the tier and the minutes of audio.
    What if none of the listed problems match what I'm hearing?
    Leave every checkbox unticked and the tool falls back to a light normalize-and-denoise recommendation, which covers most clean-ish digital sources.
    Is 100 the worst possible score?
    Yes, the score is capped at 100. A water-damaged tape with hiss, hum, clicks, clipping, dropouts, and reverb, restored to archival quality, will hit that ceiling. That result is a signal to bring in a specialist rather than push AI tools past what they handle well.
    Does heartheweb do audio restoration itself?
    No. heartheweb turns written articles into a clean MP3 feed for listening while you commute or when screens are not an option. This calculator exists because narration fidelity and audio cleanup share the same underlying questions about noise and quality.

    Curious what heartheweb actually sounds like?

    heartheweb turns your reading list into a private MP3 feed, narrated cleanly, for the commute or the moments a screen is not available.

    Explore heartheweb
    VoicesPricingField notesComparisonsRatingsResourcesAboutContact Search
    Sign in Run demo