Film Sound Design 3D

Synthesise a movie sound effect on a dubbing-stage console: faders, level LEDs, speakers, screen and live spectrum.

Presets

Playback

Waveform

Synthesis parameters

Output

Camera

Layers

Quality

Model check

The science

Building a sound effect

Each effect is synthesised sample by sample at 22,050 Hz (identical to the 2D page): an oscillator (sine, triangle, square or saw) whose frequency glides linearly from the start to the end value over the whole sound, mixed with white noise by the noise mix, passed through a one-pole low-pass filter y[n] = y[n-1] + α(x[n] - y[n-1]) with α = 1 - e-2π fc/fs, and shaped by an attack-decay-sustain-release envelope with a fixed 150 ms sustain hold.

Spectrum and what it tells you

The spectrum is the magnitude of a 2048-point radix-2 FFT, |Xk|/N × 2, with bin width fs/N = 10.8 Hz. The page shows the loudest 93 ms window of the whole sound, and during playback also a Hann-windowed live window around the playhead. The spectral centroid Σ fk|Xk| / Σ |Xk| is a rough brightness measure: a laser zap is bright, a door thud dark.

What you see

The eight faders on the console are the eight synthesis parameters and follow the sliders; the lit button is the waveform; the LEDs are the live level (RMS over 20 ms in dB, with a peak marker). The speaker cones vibrate in proportion to the level, the film screen and the ribbon in the air show the waveform (blue) with its envelope (gold) and the playhead, and the 64 spectrum bars use a logarithmic frequency axis from 40 Hz to 10 kHz and a 60 dB range. Playback uses the Web Audio API with the same samples; the silent auto-preview drives the same visuals without sound.

Independent checks

(1) The FFT of a 440 Hz sine peaks within one bin of 440 Hz. (2) Parseval's theorem Σ|x|2 = (1/N)Σ|X|2 holds, and the envelope equals 1 at the end of the attack and the sustain level during the hold. (3) A pure sine rendered at your current start frequency peaks within one bin of that frequency. (4) A flat-envelope tone in the chosen waveform has its strongest component at an integer multiple of f0 (a saw, square or triangle has harmonics).

Honest simplifications

This is a teaching synthesiser, not a sound-design workstation: no reverb, no stereo, no anti-aliasing of the saw and square (they alias above the Nyquist frequency of 11 kHz), a linear rather than musical pitch glide, and a one-pole low-pass instead of a resonant filter. The room, speakers and console are an illustration and not an acoustic simulation; cone motion is a visual cue scaled from the level, not the physical excursion. Educational model.

Loading model...
Loading Blender model...
Amplitude envelope (ADSR) with playhead
Waveform with playhead
Spectrum: best window and live window
Made with 🟠 Blender · Three.js