NoiseColour Lab Experimental

Inside the project

Prepared with the project’s development agent from source and retained evidence. This is a project-owner assessment, not an independent customer review. Dated checks were not rerun for this website.

NoiseColour Lab Experimental is a native macOS sound laboratory for making, shaping, performing and examining sound. Its strongest idea is to keep musical exploration and technical inspection close together. You can start with coloured noise, an oscillator or an imported recording, connect processors, play the result, and inspect the settings and measurements behind it. It is an experimental application with working musical capabilities and an explicit research backlog.

The main Soundscapes workspace offers Play, Shape and Inspect views around the same recipe and instrument. That gives the project a clearer centre than a collection of unrelated research panels: choose a source, hear how it behaves, then reveal the exact controls when needed. Sources can join through Mix stages, while shaping and space processors follow the signal. Some operators have a continuous live path; others use prepared rendering or finite previews. That distinction matters when presenting the app as a playable instrument.

Stellar brings a visible oscillator editor into this workspace. Sine, triangle, saw and pulse waveforms sit alongside tuning, pulse width and selectable 1×, 2× or 4× quality. Frequency, phase and amplitude modulation connect a modulator and carrier, with explicit feedback behaviour. The keyboard, MIDI input, score and existing sound routing remain the performance surface. The visual waveform is a parameter guide rather than a measured spectrum.

External AU instruments are another source option. The installed OP-Xa V has been tested through the shared background producer, convolution and graph routing. Notes, score playback, pitch bend and pressure follow the existing input path. This integration does not turn Stellar-specific oscillator controls into undocumented controls for a third-party plugin. OP-Xa V is a tested example of hosting, not an included instrument or an endorsement.

Convolution now combines an immediate early response with the selected impulse's full tail, without adding an extra wet block buffer. The wet blend stays in the existing keyboard controls. Recent work reduced the steady render-ahead queue and added a temporary reserve around structural routing changes. A producer-specific Swift task executor supplies a separate interactive scheduling path while the audio callback remains a ring-buffer reader. A failed earlier candidate was retained rather than reported as a success.

The accepted 17 October build records 139 passing release tests. Its six sustained device trials covered OP-Xa V and Stellar at 1× and 4×, with Fast and Balanced response settings. Across 469 seconds of playback, 450 tracked events and 24 deliberate UI stalls, the final trials recorded zero device underruns, AU render failures or mailbox overflows. Application-event-to-callback p95 latency fell from about 67–69 ms to about 42–45 ms on the same MacBook Pro speakers at 44.1 kHz and 512-frame callbacks. These are bounded application-pipeline measurements, not physical key-to-speaker latency or a guarantee for every machine and patch.

Beyond performance, the lab offers imported audio, grains, voice/carrier processing, digital analysis, deterministic recipes and research models. The practical next steps are a retained end-to-end save/reopen/export session, portable assets, broader undo coverage and a small collection of verified patches. Some physical and advanced-noise experiments remain finite previews, and the presence of a panel does not establish scientific, acoustic or clinical validation. The current package targets macOS 26. The website now includes two real native Stellar screenshots and a 22.6-second silent interface video captured on 18 October 2026. A separate two-second generated noise reference demonstrates a rendered noise source. The video has no audio; an AU-specific walkthrough remains to be captured.

Features & evidence

Workflow walkthrough

A guide to the native workflow. These steps describe the app; only the recordings and images embedded on this page are available as website demonstrations.

1. Open Soundscapes and choose Stellar through the existing instrument selector.
2. Select Pulse, drag its width guide and compare exact values in the editor.
3. Choose a modulation mode and adjust depth; use the existing keyboard or score to hear changes.
4. Route the instrument through a Mix with coloured noise and a Monster Filter.
5. Select an impulse and compare dry/partial/full wet; show queue and underrun diagnostics.
6. Return to Inspect to explain the recipe and evidence. Describe export only for the documented supported built-in route; do not imply live AU/convolution capture is complete.