Set the display controls#

The toolbar holds two kinds of control, and it helps to know which is which. FFT, Window, and Channel change how the app analyses the audio, so a change re-runs the analysis. Colors, Wave, Freq, Gain, and Range only change how the app draws the result. None of them change the file.

Set FFT, Window, and Channel#

Three controls decide how the app analyses the audio:

• FFT: sets how many samples go into each analysis window. Options 512, 1024, 2048, 4096, 8192 (default 2048).
• Window: sets the taper the app applies to each analysis frame. Options Hann (default), Hamming, Blackman, Rectangular.
• Channel: picks what the app analyses on a multi-channel file. Mix averages the channels together, or you can analyse Left or Right on its own.

FFT size is the classic trade. A larger window separates frequencies that sit close together, but it smears sharp events in time. A smaller window keeps clicks and edges crisp, but it blurs neighbouring tones. If two tones you care about merge into one line, go up. If a transient spreads into a wide band, go down.

Hann suits almost everything. The other tapers are there for the case where a strong tone leaks across the spectrum and buries a weak one nearby.

On a mono file, the Channel control reads Mono and stays disabled. This choice also decides which channel the decoders and measurements read.

Set the palette and frequency axis#

Three controls decide how the app draws the result:

• Colors: chooses the palette that maps energy to pixel colour. Options Inferno (default), Viridis, Magma, Plasma, Greys.
• Wave: sets the colour of the waveform strip, separately from the spectrogram palette. It opens on a set of swatches, and you can enter a hex value instead. The default is a neutral grey.
• Freq: chooses how the app lays out the vertical axis. Options Linear, Log, Mel.

Inferno runs from near-black through purple and orange to bright yellow at the loudest points. Greys suits printing, and screenshots that must survive a photocopier. Change the Wave colour when the waveform competes with the palette in a screenshot, or when you want a marked-up copy to read clearly in a report.

The three frequency layouts differ in what they make easy to read:

• Linear: gives equal pixels per hertz. Use it when absolute frequency matters, for example a carrier tone or ultrasonic content near the top of the band.
• Log: gives more room to the low end, and matches how you hear pitch.
• Mel: a perceptual scale weighted towards the frequencies that carry speech.

The axis labels and the cursor readout always report true frequency in hertz, whichever layout is active.

Set gain and range#

Two sliders decide how the palette maps onto what you see:

• Gain: shifts the whole picture brighter or darker before the app applies the palette. Runs from -20 dB to 40 dB (default 20 dB).
• Range: sets how many decibels the palette spans, from 20 dB to 120 dB (default 80 dB).

Together they decide what you can see. A narrow range pushes everything but the strongest features into the background, which isolates a signal from a noisy recording. A wide range brings up the noise floor and faint harmonics, which finds something that someone kept quiet on purpose. Raise gain when the whole picture is too dark to read.

Reset on the toolbar does not change either slider. It only clears the zoom.