Spectrogram Viewer

A spectrogram is a visual representation of how the frequencies in an audio file change over time. Think of it as a heat map for sound. The horizontal axis represents time (left to right, from the start to the end of the file). The vertical axis represents frequency (bottom = low bass, top = high treble). The color at each point shows how loud that frequency is at that moment — bright colors mean loud, dark means quiet or silent.

Horizontal bright lines indicate sustained tones or notes at a specific frequency. Vertical bright columns indicate sudden bursts of energy across all frequencies (like a drum hit or a click). A bright band near the bottom that stays consistent is usually bass or vocals. A sharp cutoff where all color stops abruptly (e.g., nothing above 16 kHz) usually indicates a lossy format like MP3 — lossless files typically have content all the way up to the Nyquist frequency. Silence appears as dark/black areas. You can zoom in by clicking and dragging on the spectrogram to inspect specific sections more closely.

FFT stands for Fast Fourier Transform — it's the math that breaks audio into individual frequencies. The FFT Size controls the trade-off between frequency detail and time detail. A larger FFT size (like 16384) gives you very precise frequency information — you can see individual notes clearly — but the time resolution gets blurry. A smaller FFT size (like 1024) gives you sharp time resolution — you can see exactly when a sound starts and stops — but the frequency information is less precise. 4096 is a good default that balances both. Try switching between sizes to see the difference.

dB (decibels) measures how loud a sound is. The dB Range sliders control which loudness levels are visible in the spectrogram. The Min slider sets the "floor" — anything quieter than this value appears as black. The Max slider sets the "ceiling" — anything louder appears as the brightest color. Narrowing the range (e.g., -80 dB to 0 dB) increases contrast and makes quieter details more visible. Widening the range (e.g., -140 dB to 0 dB) shows more of the background noise floor. If your spectrogram looks too dark, try raising the Min slider. If it looks washed out, try lowering it.

SoX (default) uses a warm palette that goes from black through purple, red, orange, yellow, to white. It's inspired by the SoX audio tool and gives good perceptual contrast across the full range. Spectrum uses a classic rainbow palette — blue for quiet, through cyan, green, yellow, to red for loud. It's similar to what many scientific tools use. Mono is a simple grayscale — black for silence, white for loud. It's clean and easy to read, but shows less detail in the mid-range. Choose whichever makes the features you care about easiest to see.

A window function is applied to each chunk of audio before the FFT runs. It controls the trade-off between frequency precision and spectral leakage (unwanted blurring between adjacent frequencies). Hann (default) is the most common choice — it offers a good balance and works well for most audio. Hamming is similar to Hann but has slightly less leakage at the cost of wider main lobes — useful when you need cleaner separation between close frequencies. Blackman-Harris has the best side-lobe suppression of all three — it produces the cleanest frequency separation with minimal leakage, but each frequency band appears slightly wider. For most music and general use, Hann is fine. Try Blackman-Harris if you want the sharpest frequency isolation.

Linear (default) spaces all frequencies evenly — 1 kHz gets the same vertical space as 10 kHz. This is what Spek and most spectrum analyzers use. It gives a clear view of high-frequency content and makes it easy to spot lossy cutoffs. Logarithmic gives more space to lower frequencies, matching how we actually perceive pitch — the jump from 100 Hz to 200 Hz sounds the same as 1000 Hz to 2000 Hz (both are one octave). This makes the log scale much better for analyzing music, because most musical content (vocals, bass, guitars, drums) lives below 5 kHz. Use Linear to inspect the full spectrum or check for lossy cutoffs. Use Logarithmic when you want to see musical detail and note separation.

For stereo audio files, the Channel selector lets you view the spectrogram of the Left channel only, Right channel only, or a Mix of both (average). This is useful for spotting differences between channels — for example, some instruments may be panned to one side, or one channel may contain artifacts that the other doesn't. The selector only appears when you load a stereo file.

Click and drag on the spectrogram to select a rectangular area — the view will zoom into that region. You can zoom in multiple times to get closer. The time and frequency axes will update to show the zoomed range, and a zoom info bar will appear showing the exact range. To reset back to the full view, either double-click on the spectrogram or click the Reset Zoom button.

Yes. Lossy formats like MP3 and AAC cut off high frequencies to save space. A 128 kbps MP3 typically has no content above ~16 kHz. A 320 kbps MP3 cuts off around 20 kHz. You'll see a sharp horizontal line where all color abruptly stops — everything above is black. In contrast, a genuine lossless file (FLAC, WAV) typically shows content all the way up to the Nyquist frequency (half the sample rate). If someone claims a file is "hi-res" but the spectrogram shows a hard cutoff at 16 kHz, it was likely upsampled from a lossy source.

No. All processing — decoding, FFT analysis, and rendering — happens entirely in your browser using the Web Audio API and JavaScript. Your audio files never leave your device and no data is sent to any server. You can even use this tool offline once the page has loaded.