Audio Inspector

Analyse audio files: waveform and spectrogram views, plus decoders that pull hidden content out of the signal.

See the whole file at once

Load a recording and the workspace draws it two ways at the same time: a waveform strip showing loudness over time and a spectrogram showing which frequencies are present at each moment. Both views share one time range, so zooming one moves the other, and the Analysis panel below turns what you can see into something you can read.

Synchronised waveform and spectrogram over one shared time range
Docked Analysis panel with Metadata, Decoders and Measurements tabs
Status bar reporting sample rate, channels, bit depth, codec and duration

Audio Inspector workspace with a Morse recording loaded, showing the toolbar, waveform strip, a spectrogram of repeating tone bursts and the Metadata tab open below

The Decoders tab listing DTMF, Morse, SSTV with a Martin M1 mode dropdown, and LSB extract with channel and bit-depth dropdowns, each with its own run button — Four decoders pull hidden content out of audio: DTMF digits, Morse text, SSTV images and LSB steganography payloads.

Decode what the audio is carrying

Some audio is not meant to be listened to. Select a region on the spectrogram, run a decoder and read the result: DTMF returns the dialled digits, Morse returns the keyed text, SSTV rebuilds a transmitted picture line by line, and LSB extract reads a payload hidden in the low bits of PCM WAV samples.

SSTV decoding across seven modes: Martin, Scottie, Robot and PD 120
LSB extract with selectable channel and one to four bits per sample
Decoders run over your selection, or the whole file with nothing selected

A decoded SSTV image in the results panel, showing a Robot 36 test pattern with a colour gradient and the text GraphDagger SSTV Robot36

Read the container before you trust the signal

A file that sounds ordinary can still be structurally wrong. The Metadata tab runs on its own the moment a file opens: it walks the container structure, lists every chunk with its offset and size, reports the claimed format and flags structural anomalies before you spend time on the signal. A chunk you do not recognise, or one far larger than its contents justify, is a good place to go looking for hidden bytes.

Anomalies surfaced first, only when there is something to report
Format claims you can compare against what the decoder produced
Embedded metadata tags left by recorders and editing tools

The Metadata tab reporting PCM format, one channel, 44100 Hz and 16-bit, with fmt and data chunks listed at their byte offsets

The Measurements tab showing peak -0.0 dBFS, RMS -3.0 dBFS and a sample count, above a strings list badged 100 of 2000 with RIFF and WAVE markers — Measurements answers two questions at once: peak and RMS in dBFS for the selected region, and every printable string sitting in the raw container bytes.

Tune the view until the signal stands out

The toolbar splits into analysis controls and display controls. FFT size from 512 to 8192 and a choice of Hann, Hamming, Blackman or Rectangular windows change how the audio is analysed; colour map, frequency scale, gain and range change how it is drawn. Narrow the range to isolate a strong signal, widen it to bring up something kept deliberately quiet, then export the spectrogram or spectrum straight from the toolbar. WAV, MP3, FLAC, OGG, AAC and M4A files are all accepted.

FFT size and window selection with per-channel analysis
Inferno, Viridis, Magma, Plasma and Greys colour maps
Linear, log and mel frequency scales with gain and range sliders
Opens WAV, MP3, FLAC, OGG, AAC and M4A files

The spectrum panel's decibel curve against a frequency axis running to 24 kHz, peaking near 2 kHz and falling to a noise floor around minus 100 dB — The spectrum panel plots the moment under your pointer as decibels against frequency: sharp peaks are tones, broad humps are noise, and the flat stretch on the right is the noise floor.