Spectrogram
A picture of time and frequency: low sounds at the bottom, high at the top, loud in bright colors. It is computed in the browser with a dependency-free FFT, in slices, so the page never freezes.
Adding it
import { spectrogram } from 'sonicframe';
frame.use(spectrogram({ source: '/voice.mp3', map: 'magma' }));
// On the player, in the palette's colors:
createPlayer({ container, src, spectrogram: true });
Like the peaks, it is computed only once the frame comes
near the screen. The sound is decoded once; a function
given as source lets the app share one
decode between plugins.
Options
| Option | Default | What it does |
|---|---|---|
source |
An address or blob to decode, or a function that gives decoded audio. | |
samples |
Ready-made mono samples (skips decoding). | |
sampleRate |
44100 |
The sample rate of
samples, per second.
|
map |
'magma' |
viridis,
magma, gray,
ice, fire,
ocean, or your own
{ from, via, to }.
|
fftSize |
Window size, a power of two from 32 to 32,768 (1024 by default). | |
columns, rows
|
The picture's size in time and frequency: 512 × 256 by default, up to 16,384 × 4096. | |
minDb, maxDb
|
Decibels mapped to the quietest and the brightest color. | |
scale |
'linear' or
'log' (rows spread over
octaves).
|
|
detail |
false |
Zoomed in, a sharp picture of the part in view once the view rests. |
detailDelay,
detailColumns
|
200, the width |
How long the view rests first (ms), and the sharp picture's columns (at most 4096). |
Your own map colors take any CSS color or
var(--name, fallback), so a palette works;
after refresh(), the picture is repainted
in the new colors without computing again.