01
Time moves from left to right
The start of the track is at the left edge. Choruses, drops, pauses, and individual hits appear in the same order as playback.
Free browser tool
Online spectrogram generator for WAV and MP3
Turn a local WAV or MP3 into a 1920 × 840 spectrogram online, right in this tab. Download the PNG, then use the guide below to read it without mistaking every bright patch for a problem.
How do I turn a WAV into a spectrogram?
Choose one WAV or MP3 file below. The generator reads it locally and creates a downloadable PNG. Your file must be 5 minutes or shorter and 100 MB or smaller.
Processed on your device.
Your audio never leaves this browser: it stays in this tab's memory. It is never uploaded, sent to an API, or saved in persistent browser storage. Refreshing or closing the tab clears the working data.
Local spectrogram lab
48 kHzFFT 4096PNG 1920 × 840
Choose an MP3 or WAV to begin
See what is inside the sound.
The browser will decode the track, measure its frequency content, and create a downloadable magma spectrogram.
This file could not be processed.
The app opens separately. Choose your audio again there if you want to enhance it.
Spotted unnatural haze or high-frequency cutoff? Learn how to clean it
Listen to that moment first. Haze may be cymbals or reverb; a cutoff may come from the source or MP3 encoding. Neither pattern proves an AI artifact. Cleanup cannot restore frequencies missing from your source.
How to remove Suno artifacts
Check the audible problem, try a small manual change, then compare a Sunofix cleanup pass with your original.
Fix harsh highs in AI music
If cymbals or consonants sound piercing, follow the listening checks and gentle treatment options. Stop if useful brightness disappears.
WAV to spectrogram vs MP3: why lossless is better for artifact inspection
Start with the original PCM WAV
Uncompressed PCM WAV avoids adding lossy encoding to the signal you want to inspect. Use the original export when available so you can check a vocal edge, noisy gap, or cymbal tail without a new compression layer.
Use MP3 when that is your source
MP3 encoding can discard quiet detail, add texture, or reduce the upper range. You can still use this free spectrogram maker to locate a suspicious section. Converting MP3 to WAV does not restore discarded information.
The image scale reaches 24 kHz, but the file may contain much less. A 44.1 kHz source cannot carry frequencies above 22.05 kHz; resampling does not create missing detail. A dark upper band is a clue to investigate, not a reason to brighten the track automatically.
For format limits, see Audacity’s MP3 export guidance and spectrogram frequency settings.
Read the image
How to read your audio spectrogram
Find the moment you heard first. The picture is most useful when it helps you return to a specific vocal, cymbal hit, reverb tail, or noisy gap.
02
Frequency rises from bottom to top
Bass sits near the bottom, mids occupy the center, and cymbals, air, and hiss live higher up. The scale reaches 24 kHz.
03
Color shows relative energy
Black and purple are quieter. Orange and yellow are stronger. Color is relative to this image, so it is not a loudness or quality score.
The important limit
A spectrogram is not an automatic diagnosis.
A bright line may be a healthy harmonic. A dark top can belong to a deliberate arrangement. A dense chorus is expected to look busier than a quiet verse. The image tells you where energy appears, not whether the music is good.
| Pattern | What it may show | What to listen for |
|---|---|---|
| Horizontal band | A sustained note, resonance, or harmonic | A tone that rings or whistles longer than expected |
| Vertical stripe | A drum hit, click, edit, or sudden attack | A clean transient versus a sharp tick or pop |
| Fuzzy upper haze | Cymbals, air, hiss, codec texture, or dense reverb | Splash, fizz, fatigue, or a texture that moves with the chorus |
| Sudden hard ceiling | A source or codec with little energy above one frequency | Whether the top end sounds naturally dark or abruptly closed |
A useful check
Compare sections before you compare files
One spectrogram can already answer a practical question: does the suspicious texture appear everywhere, or only when the arrangement gets crowded?
- 1
Start with your ears
Play the track and note the exact section that sounds metallic, harsh, smeared, noisy, or unexpectedly dull.
- 2
Find the same moment in the image
Use the left-to-right timeline and look for a change that begins and ends with what you heard.
- 3
Compare against a calmer section
A verse and chorus from the same file share the same rendering scale, which makes the contrast easier to interpret.
- 4
Confirm by listening again
The visual clue is only a lead. Make any cleanup decision with level-matched playback, then check headphones and speakers.
Keep diagnosing
Match the picture with what you hear
If the picture sent you looking for an answer, start with the sound that bothered you. These guides help you tell ordinary brightness from metallic shimmer, piercing highs, or brittle texture.
Metallic sound
Hear metallic shimmer before you try to spot it
Use the listening guide to separate moving, glassy texture from ordinary brightness.
Read the guideHarsh highs
Decide whether the top end is harsh or simply bright
Compare piercing hats, brittle upper mids, and sibilance across the busiest sections.
Read the guideArtifact guide
AI music artifacts: shimmer, hiss, wobble, and smearing
Match common visual clues with the sound they may accompany.
Read the guideSpectrogram guide
Read a spectrogram for AI music artifacts
Go deeper on axes, color, false positives, and listening confirmation.
Read the guideFAQ
Audio spectrogram FAQ
Does Sunofix upload my audio to generate the spectrogram?
No. This page reads, decodes, analyzes, and renders the file inside your browser. The audio is not sent to Sunofix or stored after you close or refresh the tab.
Should I use MP3 or WAV?
Use the original uncompressed PCM WAV when you have it. MP3 is lossy: encoding can discard quiet detail, add texture, or reduce high-frequency content. An MP3 spectrogram is still useful for locating a sound, but converting that MP3 to WAV will not restore missing detail. Compare the same section and confirm it by listening.
Can this tool decode spectrogram messages or hidden images?
It can reveal text or pictures deliberately drawn into the frequency content of an audio file. Choose the file, generate the PNG, and inspect it visually. This is a spectrogram viewer, not an automatic message decoder: it does not recognize text, decrypt ciphers, or recover every hidden message. A blank-looking image does not prove that no message exists; the fixed scale and resolution may miss fine detail.
Why is there a five-minute and 100 MB limit?
Decoded audio uses much more memory than the original MP3 or WAV. The limits reduce the chance of a mobile browser closing the tab while the spectrogram is being calculated.
Will this PNG exactly match every other spectrogram tool?
Not pixel for pixel. Windowing, FFT size, time pooling, color mapping, MP3 decoding, and resampling can differ. This generator uses the same core settings as the Sunofix before-and-after view.
Can a spectrogram tell me whether my track has AI artifacts?
No. It can help you locate a change, a persistent band, or unusual high-frequency texture, but it cannot decide whether that sound is unwanted. Start with the part that bothers you, use the image to inspect it, and confirm every conclusion by listening.
For an example of text carried by audio frequencies, see AudioCheck’s SpectroTyper explanation. Sunofix uses fixed rendering settings, so some messages need a viewer with adjustable frequency range or resolution.
Hear a problem too?
A spectrogram shows the track. Sunofix can help clean it.
The app opens separately and does not receive the file selected on this page.