Radio Music Mk2 Discovery Test Script (Combination Bug Hunt)

Radio Music Mk2 Discovery Test Script (Combination Bug Hunt)

This script is intentionally different from checklist/unit-style verification. It is designed to find serious real-world bugs caused by interactions, unusual workflows, and unexpected content.

1. Core idea

Instead of only asking “does feature X work?”, this approach asks:

• what happens when many valid features are used together,
• what happens with unfamiliar files and card states,
• what happens during performance-style use where state changes are fast and non-linear.

2. Session format

Run in 3 parts:

1. Guided charters (10 missions, 15-25 min each).
2. Combination sweeps (targeted mode intersections).
3. Free exploration (creative abuse with full logging).

Target total runtime: one long day (6-10 hours) or split across 2-3 shorter sessions.

3. Prep before starting

3.1 Media and SD prep

• Use the generated pack: test-media/rm2_generated_v1/.
• Add at least 20-40 extra files exported by tools not used in firmware bring-up, for example:
  • Logic Pro
  • Ableton Live
  • Reaper
  • Audacity
  • Zoom recorder / phone audio app
• Keep one SD card formatted on macOS and one on Linux/Windows if possible.

3.2 Logging setup

For every anomaly, capture:

• continuous phone video of panel + patch,
• exact bank/folder and station position,
• exact settings.txt used,
• whether 8mu was connected and active,
• whether issue survives reboot,
• copy of SD card contents (zip or image) used when issue occurred.

4. Ten discovery charters

D-01 Conventional sampler performance

Goal: test normal musical play with rapid station changes.

• Use loopMode=0, tunerMode=0, speedMode=0.
• Play it like a one-shot sample player for 15 minutes.
• Alternate between reset-button hits and reset-jack triggers. Watch for:
• stuck voices, missed retriggers, unintended fades.

D-02 Drum machine behavior

Goal: high retrigger pressure + station churn.

• Use short percussive files.
• Drive RESET from an external clock and change STATION during clocking.
• Repeat with crossfadeTime short vs long. Watch for:
• timing drift, transient smearing, freezes.

D-03 Virtual radio behavior

Goal: tuner interaction realism and station transitions.

• Use tunerMode=2 (radio).
• Sweep STATION continuously with and without CV.
• Vary minRadioStationStrength, noiseVol, whistleVol, ssbEffect. Watch for:
• abrupt discontinuities, unstable gain behavior, mode lockups.

D-04 Clock source behavior

Goal: pulse output stability while playback state changes.

• Set pulseMode to output mode.
• Patch RESET jack output to an external clock destination.
• Change loop mode, start point, bank, and 8mu loop region during run. Watch for:
• missing/double pulses, pulse stalls, output mode confusion.

D-05 8mu live-control behavior

Goal: remote-control edge interactions.

• Connect 8mu and use all sliders/buttons in performance style.
• Hold motion control while changing speed/tuner/loop modes.
• Return sliders to zero and verify clean return to local control. Watch for:
• stale remote state, control jumps, unexpected mode persistence.

D-06 Adjacent-bank interaction stress

Goal: catch bank-to-bank edge bugs.

• Put one bank with identical-length files next to one with mixed long files.
• Switch between those banks while keeping file active for very short times.
• Repeat with crossfade times longer than active dwell. Watch for:
• wrong seek position, wrong file state carry-over, crash/lockup.

D-07 Station-density stress (pot feel)

Goal: detect selection instability with dense banks.

• Compare 8_pot_density_3, 9_pot_density_30, 10_pot_density_60.
• Sweep STATION slowly and pause at boundary areas.
• Add small station CV offsets. Watch for:
• unreachable files, oscillation near boundaries, jumpy selection.

D-08 Settings edit workflow stress

Goal: validate real-world settings iteration.

• Enter USB mass-storage mode.
• Make small settings.txt edits, reboot, test immediately.
• Include case/spacing/comment variants and per-bank overrides. Watch for:
• settings not applied, partial apply, parser oddities.

D-09 Format roulette

Goal: parser and playback robustness with mixed provenance files.

• Build one bank with mixed WAV/AIFF/RAW from many programs.
• Include odd headers, unusual metadata chunks, and near-edge valid files.
• Rapidly move through stations and retrigger. Watch for:
• scan crashes, mis-detected files, silent-but-counted files.

D-10 Creative freeform abuse

Goal: find unknown unknowns.

• Spend 30-60 minutes using RM2 in any unusual way you can think of.
• Switch roles repeatedly: sampler -> radio -> clock -> pitched voice.
• Use 8mu, CV, panel, bank changes, and settings edits in one flow. Watch for:
• anything surprising, inconsistent, or hard to explain.

5. Combination sweep matrix

Run at least 12 combinations from this matrix (more is better):

Combo ID	loopMode	tunerMode	speedMode	Bank type	crossfadeTime	8mu
C01	0	0	0	identical	5	off
C02	0	1	1	nonidentical	25	off
C03	0	2	2	mixed	200	on
C04	1	0	1	identical	5	on
C05	1	1	2	nonidentical	25	on
C06	1	2	0	mixed	200	off
C07	2	0	2	identical	5	off
C08	2	1	0	nonidentical	25	on
C09	2	2	1	mixed	200	on
C10	1	2	2	adjacent-bank stress	100	on
C11	2	2	0	dense station bank (60)	25	off
C12	0	1	2	dense station bank (60)	150	on

For each combo:

1. Run for 5-10 minutes.
2. Include station sweeps, reset triggers, and at least one bank switch.
3. Record anomalies immediately with timestamp.

6. What counts as a serious bug here

Prioritize reporting when you see:

• hard lockups, reboots, or audio engine stalls,
• wrong file/position/state after mode or bank transitions,
• behavior that depends on hidden historical state and is hard to predict,
• controls that become unreliable only in high-density or mixed-mode usage,
• data-dependent failures (specific file export chain/card format/settings combos).

7. Bug report payload for discovery bugs

When filing:

• include short video from before failure to after failure,
• include exact SD content snapshot (zip or image),
• include exact settings files (root + bank),
• include firmware build identifier,
• include an ordered action list that another tester can replay.

Use template:

• .github/ISSUE_TEMPLATE/radio-music-bug-report.md

8. How this complements existing scripts

• Core/Standard/Full scripts remain the structured baseline.
• This discovery script is for high-value combination bugs and workflow surprises.
• Run this after baseline scripts, or in parallel when baseline confidence is already high.