Number Five /

03 / PROJECT

Audio & Spotify

The Pi plays the music. The helm keeps the controls simple.

Pi playback demonstrated · development continuesUpdated: 1 November 2026
Audio by day
Current native audio view on the Lumina. Main and zone controls come from the running Pi DSP; no playback source was connected during capture. Boat telemetry in the header is labelled as simulated.HELM 206 · 0.8.134-sonar-start-depth-controls · Lumina K706 · Android 10 · 2000 × 1200

01 / STARTING POINT

What we wanted to solve.

The goal was not simply a Spotify window on a large display. Music should run independently of the screen, with large controls, physical knobs and one consistent volume state. Multiple speaker channels and useful sound processing were part of the task.

WHAT THIS COMPONENT MAKES POSSIBLE

The practical difference.

Music is not tied to one display. A control position can change while playback and sound processing remain in the independent Pi module.

Playback, equalizer and channel routing receive consistent controls. A requested state is only confirmed after a response. Current device captures show real Pi preset, equalizer, room and channel data. Track/playback state and sound controls are separate: track state was missing during capture while the DSP was running.

Read the decision in the story

02 / DEVELOPMENT

Our path to getting there.

We separated playback from the display: a Spotify Connect receiver feeds digital sound processing (DSP) through CamillaDSP on the Pi. The Android interface receives state and sends commands through the gateway.

State synchronisation, volume, reconnects and handover between Spotify and our controls required substantial work. One shared master avoids competing volume controls.

Sound presets, filters and channel routing are kept separate from playback. Development views may show example tracks; they do not establish a finished personal Spotify-library integration.

03 / RESULTS

What exists today.

  • Audible Spotify playback through the receiver and CamillaDSP is documented in the lab; Pi migration and stability have separate records.
  • A DSP bus for front, rear and subwoofer channels is implemented. Virtual channel checks and actual speaker installation remain separate.
  • On 20 October, sound presets were saved and reselected through HELM on the physical Pi; the catalog survived a gateway restart.

Images show the development revisions identified in their captions. Today’s source may be newer than an existing image.

NATIVE DEVICE CAPTURE32 s · 1×

Two devices. One confirmed volume setting.

32 seconds of native recording on receiving client Lumina 1: the main fader is moved from −39 to −40 dB and restored on Lumina 2. Both devices, the gateway and Pi DSP show the same state.

2026-11-01 · Receiving Lumina 1 K706 · Android10 · 2000 × 1200 (recording1280 × 768); controller Lumina2 K706 · HELM 206 · 0.8.134-sonar-start-depth-controls on both devices · Original speed, silent

Text description and context

32 seconds of native recording on receiving client Lumina 1: the main fader is moved from −39 to −40 dB and restored on Lumina 2. Both devices, the gateway and Pi DSP show the same state.

Confirms the shared main-fader state in this setup; no claim of frame-accurate synchronization or measured latency. No montage, music playback or acoustic measurement. Other boat values remain labelled as simulated.

Captions as text ↓
NATIVE DEVICE CAPTURE31 s · 1×

Sound, room and listening position on the device

31 seconds from the current native HELM app: existing sound presets, EQ, room view and speaker positions. Real Pi sound settings were independently checked against the running DSP.

2026-11-01 · Lumina K706 · Android 10 · 2000 × 1200 (recording: 1280 × 768) · HELM 206 · 0.8.134-sonar-start-depth-controls · Original speed, silent

Text description and context

31 seconds from the current native HELM app: existing sound presets, EQ, room view and speaker positions. Real Pi sound settings were independently checked against the running DSP.

No playback source connected; no music, artificial spectra or changed sound settings. The boat header is labelled as simulated. This demonstrates operation on one client, not two-client synchronization, acoustic acceptance or GPU performance.

Captions as text ↓
Earlier development revisions and designs +

These images explain the development journey. Historical test views and designs are separate from current device captures.

04 / REPRODUCE

What you will need.

  • Pi with a suitable audio output, amplifier and speakers
  • Spotify access subject to the terms of the services in use
  • CamillaDSP and channel routing matched to your installation
See sources and downloads

05 / KEEP BUILDING

What comes next.

Library, search and other media features have different implementation and test states across revisions.

The complete system still needs checking with its actual speakers and boat power supply.

Here is where you can help.

Help with practical audio hardware, reliable playback and calibration aboard.

Write to me

The foundations we build on

← All projectsDisplays & controls