Generate an audio spectrogram online from MP3, WAV, FLAC or M4A files. Linear, log and mel frequency scales, adjustable FFT size and PNG export. No upload.
Also known as: audio spectrogram · spectrogram online · spectrogram generator · audio spectrum analyzer · mp3 spectrogram · wav spectrogram · mel spectrogram · frequency analyzer online · spectrogram viewer · audio frequency visualizer
Generate an audio spectrogram online from MP3, WAV, FLAC or M4A files. Linear, log and mel frequency scales, adjustable FFT size and PNG export. No upload.
Is Audio Spectrogram free to use?
Yes. Audio Spectrogram is completely free with no usage limits, and there is no paid tier required to use it.
Do I need an account to use Audio Spectrogram?
No. Audio Spectrogram works immediately in your browser with no signup, login, or email address required.
Share code snippets, logs, and text with a shareable link. Syntax highlighting for 30+ languages, expiration controls, and password protection for Pro users.
WAV, AIFF, MP3, FLAC, OGG, Opus, M4A or WebM, or click to browse. The file is decoded in your browser and never uploaded.
Spectrogram Output
Frequency scale
Hover over the spectrogram to read time, frequency and level.
Open an audio file, load the example or turn on the microphone to see a spectrogram here.
About the Audio Spectrogram
An audio spectrogram shows how the frequency content of a sound changes over time, so you can see hum, noise, clipping, pitch and edits that are hard to hear. This tool decodes WAV, AIFF, MP3, FLAC, OGG, Opus, M4A and WebM files in your browser, runs a short-time Fourier transform with the window and FFT size you pick, and draws a labeled spectrogram on a linear, log or mel axis that you can save as a PNG.
Key Features
Linear, logarithmic and mel frequency scales, with 64 to 256 mel bands built on the same Slaney mel scale librosa uses
FFT sizes from 256 to 16,384 points, 50 to 90 percent overlap, and Hann, Hamming, Blackman-Harris or rectangular windows
Levels in dBFS, calibrated so a full-scale sine wave reads 0 dBFS, with an adjustable dynamic range and four colormaps
Reads WAV and AIFF sample values exactly at their native rate: 8 to 32-bit PCM, 32 and 64-bit float, A-law, mu-law and WAVE_FORMAT_EXTENSIBLE
Hover readout of time, frequency and level, a waveform strip, the strongest spectral peaks, and PNG export with axes and a color bar
Live microphone mode that scrolls a real-time spectrogram with the same window and scaling
Use Cases
Find the 60 Hz mains hum and its harmonics in a podcast, Zoom call or guitar recording before you reach for a noise filter
Check whether a FLAC or WAV from a download store is really a transcoded MP3 by looking for a hard cutoff near 16 kHz
See the mel spectrogram a speech recognition or audio classification model will be fed, on the librosa mel scale
Verify DTMF tones, sine sweeps and test signals while debugging telephony, IVR or audio pipeline code
Spot clipping, dropouts and edit points in field recordings, bird calls or voice memos
Compare the frequency balance of a mix before and after mastering
Supported Formats
.wav
.aiff
.mp3
.flac
.ogg
.opus
.m4a
.aac
.webm
Related Keywords
audio spectrogram
spectrogram online
spectrogram generator
audio spectrum analyzer
mp3 spectrogram
wav spectrogram
mel spectrogram
frequency analyzer online
spectrogram viewer
audio frequency visualizer
How do you read a spectrogram, and what do the dBFS numbers mean?
Time runs left to right, frequency runs bottom to top, and color shows how much energy sits at that time and frequency. Levels are in dBFS, decibels relative to digital full scale: a sine wave that just reaches the maximum sample value reads 0 dBFS, and quieter content reads negative. The FFT size sets the tradeoff between frequency and time detail. At 44.1 kHz, an FFT of 2,048 points gives 21.5 Hz frequency bins from 46 ms frames, which suits speech and music; 8,192 points resolves notes and hum harmonics to 5.4 Hz but smears fast transients. The log scale gives each octave the same height, as your ear does, and the mel scale goes further toward how pitch is perceived, which is why speech and audio models are trained on mel spectrograms. The peak list in the report comes from the long-term average spectrum, so a tone that plays for a fifth of the file shows about 7 dB below its momentary level.
Is my audio uploaded anywhere?
No. WAV and AIFF files are parsed by JavaScript in this tab, and compressed formats are decoded by your browser's own Web Audio decoder. The microphone stream, if you turn it on, is analyzed live and never recorded or sent anywhere.