BlueTune

Guided receive-audio checkups for AllStarLink 3. 0.1.2-preview.

Looking for a QSO? Visit On Air to find an AllStar conversation.

First time here? Follow Your first BlueTune checkup. It covers the prerequisite check, live access, SSH tunnel, first sample, and removal. See readiness and remaining limits for the tested scope.

BlueTune is a separate, Python-standard-library application in this folder. It does not depend on BlueNode. No installation, Asterisk restart, audio capture, transmit command, or tuning change is performed by BlueTune.

Try it

Requires Git and Python 3.11 or newer:

git clone https://github.com/BlueKF0OZX/BlueTune.git
cd BlueTune
python3 start.py

Open http://127.0.0.1:8091. On Windows use python start.py if Python is installed under that name (use python start.py). Stop with Ctrl+C.

The default workspace is explicitly demo mode, with synthetic examples for headroom, overload, and very little input. You can also paste real ASL3 RxAudioStats lines. This does not verify their collection time or device. Baseline and current results are held in page memory, and disappear on refresh. Export a checkup to keep them. No microphone permission or Internet service is needed by the app.

Experimental live SimpleUSB adapter

Run on the ASL3 machine as your normal account. The launcher first tries direct Asterisk access, then existing noninteractive sudo permission for three fixed read-only commands. It does not install or broaden permissions. The web app refuses to run as root. The fixed executable is /usr/sbin/asterisk.

1. In ASL3, select the intended SimpleUSB device. Close other tuning sessions.
2. Start BlueTune with its exact device/stanza name:

bash start.sh --check --device 1999
bash start.sh --live --device 1999

1999 is an example; replace it with your device name. Omit --device to use the active SimpleUSB device reported by ASL3. BlueTune checks that the active name matches. It does not change the selected device itself. Open the app, start a 10-second measurement, and speak into the node receiver under appropriate test conditions. Confirm the speech checkbox. A failed or canceled collection discards the partial sample.

BlueTune always listens on loopback. To view it from another computer, use an SSH forward (replace the example account/host):

ssh -N -L 8091:127.0.0.1:8091 operator@your-node

Then open http://127.0.0.1:8091 on that computer. Phone layouts are supported, but direct LAN hosting and remote authentication are not implemented yet.

The adapter runs only these literal commands:

• asterisk -rx 'susb active'
• asterisk -rx 'susb tune menu-support Y'
• asterisk -rx 'susb show settings' (only for a requested reference export)

When using existing sudo access, each command is prefixed with /usr/bin/sudo -n --. No browser-supplied command is executed. start.sh --check collects a read-only probe before reporting ready; quiet input is not a failed prerequisite check.

Unsupported output fails closed. USBRadio support is not implemented. Multiple simultaneous tuning tools are not supported: the CLI's global device selection cannot be made transactional by this adapter. Before/after checks detect many selection changes but cannot exclude a change away and back between checks. Statistics have no upstream timestamp, so a successful command does not prove fresh audio frames; identical observations are not labeled a frozen stream.

Interpreting results

• Unconfirmed speech is unknown, never a tuning pass.
• Any positive ClipCnt produces a clipping warning for a confirmed sample.
• Peak above −3 dBFS or any observed average power above −12 dBFS prompts a reduction in receive level, following ASL3's guidance.
• Peak below −40 dBFS is a conservative BlueTune heuristic for insufficient input, not an ASL calibration target. It never recommends raising gain.
• Otherwise the result is no overload detected in this sample, not a certification of loudness, intelligibility, RF deviation, or audio quality.
• ASL prints rolling-window statistics. BlueTune reports the maximum window ClipCnt, never sums overlapping windows. The average-power card shows the highest observed window average, not a session average.

Only compare the same source/device and speaking conditions. Demo and imported results cannot be compared with live readings. Imports have no verified device identity. Finish with a listening check using a consenting peer.

Keep a restorable backup

Use ASL3 Menu → Backup and Restore before manually adjusting levels. BlueTune's optional settings export is a human-readable reference only; it is not a configuration backup and cannot restore the node. Automated tuning, backup restoration, and TX measurement are outside this preview.

Validation

python3 -m unittest discover -s tests -v

Unit/integration tests cover interpretation, malformed/unsupported observations, device mismatch, command restrictions, timeouts, and HTTP access boundaries. Browser tests in tests/browser.cjs use Playwright; pass the package location with PLAYWRIGHT_MODULE if it is not on Node's normal module path. Start the demo server on 8091, then run node tests/browser.cjs. Set BROWSER_CHANNEL=msedge to use an installed Edge browser instead of Playwright's bundled Chromium.

The 17 backend tests and browser checks passed on Windows, including narrow phone layouts and simulated live success, failure, and cancellation.

On 2026-09-14, an operator-assisted receive session on one physical ASL3 node returned 15 SimpleUSB observations through SSH. The active device matched before and after collection. BlueTune's paste/import workflow accepted the output: highest peak −7.0 dBFS, highest window-average power −29 dBFS, maximum ClipCnt 0. This supports command/output compatibility and the import workflow on that node. It does not validate the deployed live web adapter, audio quality, or other hardware. The run included both voice and quiet observations; the highest peak was not independently attributed to speech versus a key/unkey transient.

Required next checks: a listening comparison with a consenting operator, deployment of the live adapter, actual device association, no-data/device-loss behavior, observation freshness, and comparison with the ASL tuning display.

First-user rehearsal (0.1.1 preview)

The candidate was copied into a fresh temporary folder on that existing ASL3 node, without installing a service or changing permissions/configuration. All 23 backend tests passed on Windows and on the node's Python 3.11.2. The startup check detected the active SimpleUSB device through existing sudo permission; a deliberately mismatched device was rejected.

The real live web adapter ran as the normal login account on loopback, through an SSH tunnel using a different local port. A browser measurement completed with nine actual quiet-input observations and remained inconclusive without speech confirmation. Cancellation discarded a partial run. This supplements the earlier voice/import check and verifies the basic live collection path on one installation. It is not a clean-OS installation test or a live speech quality test. Optional sudoers installation, other hardware/drivers/versions, device loss and stale-audio detection still need separate checks. New users should follow the first-checkup guide.

Sources inspected 2026-09-14

• ASL3 USB interface and voice-level guidance
• ASL3 menu and backup support
• SimpleUSB CLI implementation: susb_active, susb_tune, tune_menusupport, uppercase Y one-shot branch.
• Audio-statistics implementation: ast_radio_check_audio, ast_radio_print_audio_stats.

Local reference downloads under research/ are ignored development material; they are not part of the application. Threshold logic and UI are original.