Why only the first thirty seconds?

The image gives one column of pixels per analysed frame, so a longer file would either be blurred down or need a scrolling view. Capping the analysis at about thirty seconds keeps every frame represented at full resolution, and long material can be cut down first if a later section matters.

What do the colours mean?

Brighter is louder. The ramp runs from black through purple and red to yellow and white across a range of −90 to 0 dBFS, so a solid white feature is at or near full scale and a faint purple one is deep in the noise floor.

How do I read a hum or a hiss on this view?

Hum appears as one or more bright horizontal bands at a fixed low frequency, usually with harmonics above it. Hiss appears as a fairly even wash across the upper part of the image for the file's whole length, which is what distinguishes it from musical content that comes and goes.

Why is the frequency axis not logarithmic?

A linear axis is simpler to read for reference work and keeps each hertz at a constant distance, but it squeezes the low octaves together and stretches the top. That is why a bass line looks compressed near the bottom while cymbal energy occupies a wide bright region above it.

Does the file get uploaded to build the image?

No. The file is decoded in the browser and the transform runs in the page, so nothing is transmitted. The cost is that the analysis holds the frame array in memory, which is another reason the duration is capped rather than unlimited.

Spectrogram Viewer

You Might Also Need

Spectrum Analyzer

Analyze frequency spectrum

Waveform Visualizer

Visualize audio waveforms in real-time

Frequency Analyzer

Detailed frequency analysis

Noise Reducer

Reduce background noise in audio

Spectrogram Viewer turns a file into a time-frequency heatmap, so you can see not only where the energy sits but when it arrives. Hovering any point reads out the exact time, frequency and level underneath the pointer.

What is Spectrogram Viewer?

Time runs left to right and frequency runs bottom to top, with brightness standing in for level on a dark background that makes quiet detail legible against the loud events.

The view and what it reports:

A heatmap of roughly thirty seconds of audio, taken from the start of the file, drawn as a bitmap rather than as a live scrolling trace.
A fire-style colour ramp from black through purple and red to yellow and white, mapped across a range of −90 to 0 dBFS.
A hover readout giving the time in seconds, the frequency in hertz and the level in dBFS at the pointer's position.
A caption recording the channel count, sample rate, analysed duration, frame count and the top bin frequency.

The underlying analysis is a short-time Fourier transform with overlapping frames, and the frames are converted to decibels before being mapped to colour, which is why a quiet room tone still shows as faint structure instead of vanishing to black. Mapping to decibels rather than to linear amplitude matters here: a linear map would show only the loudest moments and leave everything else indistinguishable. Only the first channel is analysed, and the run is capped at about thirty seconds because the bitmap hands one pixel per column to the frame array; longer files would either blur or need scrolling that the reference view does not offer. Since each frame records level per bin, a sustained tone appears as a horizontal line, a sweep as a diagonal, and broadband noise such as hiss as a uniform wash — the three patterns worth learning first. Naming it precisely: this is a spectrogram, a picture whose axes are time and frequency, and it is the standard view for anything that changes as the recording plays. The complement is the time-frequency trade-off itself, because a short analysis window resolves rapid events while a long one resolves fine frequency detail, and no single setting serves both purposes at once. That is why a spectrogram is the right view for locating a problem in time while the averaged spectrum remains the right view for judging balance.

What the view does not attempt:

Only the first thirty seconds are analysed, so a problem later in a file will not appear.
Frequency is mapped linearly, not logarithmically, so the low octaves are compressed into the bottom of the image.
Only one channel is used, so a stereo difference is invisible here.
There is no axis labelling beyond the hover readout and the caption.
The bitmap is fixed at 900 by 320, so fine detail below one pixel is not represented.
No image export or saved analysis is offered.

Use it when you need to see structure in time: a mains hum as a bright band at fifty or sixty hertz, a hiss that runs the whole length, a click as a single vertical stripe, or a whistle that drifts. Once the pattern tells you what the problem is, the cleanup tools in the same category act on it, and the average spectrum is the right view for judging tonal balance without the time axis.

How to use Spectrogram Viewer

Choose an audio file with the file picker; the first channel of the first thirty seconds is analysed.
Look for the three signature patterns: horizontal lines for steady tones, diagonals for sweeps, and even washes for noise.
Hover any part of the image to read the exact time, frequency and level at that point.
Check the caption for frame count and top frequency to understand the resolution behind the picture.

When to use Spectrogram Viewer vs related tools

Spectrogram Viewer is the right tool when a problem or a feature has a distinctive time-frequency signature, such as hum, hiss, chirps or clicks. Spectrum Analyzer shows the same energy averaged over the file, which is better for tonal balance, and Frequency Analyzer follows a single dominant pitch through time. To clean up what the picture reveals, Noise Reducer and the notch-based hum tools address the patterns this view makes obvious.

Privacy & Security

This tool runs entirely in your browser — no data ever leaves your device. There is no server round-trip, no upload, no logging, and no account required. Your input is processed locally using client-side JavaScript and is never stored, transmitted, or accessible to anyone else. When you close the tab, everything disappears.