# Desktop ↔ Web Comparison: comm-in-thread-forum__43_exploring_tb303

- **Timestamp:** 2026-06-18T15:11:30.924Z
- **Capture window:** 30000 ms

## Code
```ruby
# Source: https://in-thread.sonic-pi.net/t/exploring-tb303/9451
# Author: Relaxnow
# Title: Exploring tb303
# Captured: 2026-04-24

set_volume! 2

use_bpm 77

with_fx :reverb, room: 0.75 do
 with_fx :krush do |krush|

 live_loop :a1 do
 tick

 control krush, mix: line(0,0.8,steps: 64).look

 sample :bd_fat, amp: 4 if spread(1,9).look
 sample :bd_fat, rpitch: 24, amp: 6 if spread(1,5).look

 sample :sn_generic, finish: 0.01, amp: 2 if spread([1,1,1, 1,1,1, 5,1,1, 1,1,1].look, 9).look

 k = 64

 use_synth :tb303

 n = :f2+knit(0,k,-2,k).look

 use_synth_defaults release: [0.1,0.2,0.4,0.8].choose, cutoff: [70,90,110].look+knit(-20,k,-10,k).look
 play n+[0,0,0,0,12].look, pan: rdist(0.25) if spread(9,12).look

 use_synth_defaults release: [0.1,0.2,0.3,0.4,0.8].choose, cutoff: [70,90,110].look+knit(-20,k,-10,k).look
 play n+[0,0,0,0,12].look+24, amp: 0.5, pan: rdist(0.5) if spread(5,12).rotate(5).look


 use_synth [:noise,:gnoise].choose
 density [1,2,1,1,4,1,1,1,1,1,1,1,1,].choose do
 use_synth_defaults release: 0.001, cutoff: rrand_i(90,130)
 play :c2, pan: rdist(1) if spread(rrand_i(14,15),16).look
 sleep 0.25
 end

 end
 end

 live_loop :a2 do
 tick
 use_synth :beep
 play :a5+knit(0,32,1,32,1,32,0,32).look, amp: 0.6 if spread(1,32).rotate(5).look
 sleep 0.25
 end

end
```

## Stats (Level 3 — observation, not inference)

| Metric | Desktop SP | SonicPi.js (web) | Δ (desk − web) |
|---------------|------------------------|-------------------------|----------------|
| Duration (s) | 31.371 | 29.696 | 1.675 |
| Peak | 1.0000 | 0.9716 | 0.0284 |
| RMS | 0.3060 | 0.1931 | 0.1129 |
| Clipping (%) | 1.99 | 0.00 | 1.99 |
| Sample rate (Hz) | 48000 | 48000 | 0 |
| Channels | 2 | 2 | — |

## Verdict
### ✅ EVENT-MATCH — per-synthdef `/s_new` onset-sequence + note parity STRUCTURE-MATCHES desktop : the engine emitted the right notes at the right times. The audio Tier-1 said _"pitch-class histogram cos=0.958 (≥0.92), tempo match, density ratio 1.00 (≥0.5), PRNG token in source; same composition, different random walk (cross-engine seed parity is not a v1 goal — #358/#364/#367)"_, but that measures stage-7+ scsynth DSP (WASM-vs-native rendering / pitch-tracker noise), not the engine. Authoritative cross-engine check passes. · ⚠ Tier-0 SOFT: level/count aggregates unreliable (Tier 1 pitch unaffected)

### Tier 0 — Validity gates
- ✓ 0.1 Sample rate consistent (48000 Hz)
- ⚠ 0.2 Capture-window misaligned (Δ 1.67s > 0.5s) — note-count / level aggregates unreliable **(SOFT — Tier 3 + 1.3 unreliable; Tier 1 pitch still valid)**
- ◦ 0.3 equal preconditions / 0.4 lossless capture / 0.5 routing sanity — not auto-checked; ensure SP.app reset + raw-float32 + FX-bus wired 

### Tier 1 — Musical correctness (THE verdict — energy/MFCC may never override)
- **1.1 Note progression:** ≈ pitch-class histogram cos=0.958 (≥0.92), tempo match, density ratio 1.00 (≥0.5), PRNG token in source; same composition, different random walk (cross-engine seed parity is not a v1 goal — #358/#364/#367)
 - method: desktop `onset` (conf 1) · web `onset` (conf 1)
 - desktop: `53,41,58,53,53,89,41,65,41,53,53,65,53,41,53,41,82,51,59,63,63,63,51,63`
 - web   : `53,41,58,53,53,89,65,41,53,53,53,53,41,53,69,82,52,59,40,63,63,57,63,75`
- **1.2 Tempo (inter-onset):** ✓ desktop 0.770s · web 0.725s/note
- **1.3 Onset count:** desktop 33 · web 33 (Δ explained by Tier-0 window misalignment)
- ◦ 1.4 note duration / 1.5 polyphony / 1.6 determinism — not auto-tracked here (unit tests cover determinism; see /)

### Tier 1.0 — Event parity (`/s_new` onset-sequence vs desktop — authoritative tiebreaker)
- **Structure (synthdef multiset):** STRUCTURE-MATCH — desktop 467 voice `/s_new`, web 453
- **Onset sequence (ε=15ms, prefix-compared, + per-tick NOTE multiset):** ✓ MATCH
 - ✓ `tb303` — 184 onsets within ε (max Δ 1ms), notes match
 - ✓ `noise` — 109 onsets within ε (max Δ 1ms), notes match
 - ✓ `gnoise` — 82 onsets within ε (max Δ 1ms), notes match
 - ✓ `basic_mono_player` — 50 onsets within ε (max Δ 1ms), notes match
 - ✓ `stereo_player` — 23 onsets within ε (max Δ 1ms), notes match
- ✅ **EVENT-MATCH:** structure + onset sequences both match — the engine's deterministic output (stages 1-6, terminating at the OSC emission) is correct. The audio Tier-1 divergence is stage-7+ scsynth rendering or tracker noise, not an engine bug (; #453/#379/#378 class).

### Tier 3 — Level / gain (reported; NOT a musical-correctness blocker — known ~0.5× web gain-staging)
> ⚠ Tier-0 SOFT failed — these ratios span misaligned windows; treat as indicative only.
- 3.1 RMS ratio web/desktop = 0.63× (within 0.5–2× band)
- 3.2 Peak ratio web/desktop = 0.97×
- 3.3 Clipping: desktop 1.99% · web 0% ⚠

### Tier 2 — Spectral / timbral (supporting only) · Tier 4 — FX/routing · Tier 5 — lifecycle
- Tier 2: see **Spectrogram comparison** section below (MFCC carries its mandatory caveat there).
- Tier 4 (FX accumulation/suppression 200ms scan, per-FX-scope energy): **not analysed** by this tool — use the FX-sweep / boundary-scan tools when FX is in scope.
- Tier 5 (Run/Stop/hot-swap, cold-start, long-run drift): **not analysed** — single capture; use `tools/test-run-stop-cycle.ts` for lifecycle.

## Source WAVs
- **Desktop:** /Users/mrityunjaybhardwaj/Documents/projects/sonicPiWeb/.captures/desktop-recordings/desktop_2026-06-18T15-10-15-453Z_comm-in-thread-forum__43_exploring_tb303.wav
- **Web:** /Users/mrityunjaybhardwaj/Documents/projects/sonicPiWeb/.captures/2026-06-18T15-10-16-734Z_inline_audio.wav

## Spectrogram comparison
![spectrogram comparison](/Users/mrityunjaybhardwaj/Documents/projects/sonicPiWeb/.captures/compare_2026-06-18T15-10-52-854Z_comm-in-thread-forum__43_exploring_tb303_spectrogram.png)

| Metric | Value | Reading |
|---|---|---|
| L2 distance (mel-dB) | 12.61 | < 10 = very close · 10–25 = similar shape · > 25 = divergent |
| MFCC distance (timbre) | 115.95 | < 30 = similar · 30–80 = noticeably different · > 80 = unrelated |
| ↳ MFCC caveat | — | **Tier-2 supporting only.** Confounded by the known ~0.5× web gain ratio + desktop reverb-tail length; **never overrides Tier 1** . A high MFCC with a Tier-1 PITCH-MATCH means timbre/gain, not wrong notes. |
| Frames compared | 2785 | overlapping window after length-aligning |
| Peak freq desktop | 174.5 Hz | dominant frequency |
| Peak freq web | 173.8 Hz | dominant frequency |
⚠ MFCC distance 115.95 is high — **check Tier 1 first**: if pitch-track matched, this is timbre/gain (the known 0.5× + reverb tail), NOT wrong notes. Only treat as "different synth/sample chain" when Tier 1 also diverges.

## Tool stdout (debug)
### Desktop
```
▶ Desktop capture (30000ms): comm-in-thread-forum__43_exploring_tb303
✓ Report: /Users/mrityunjaybhardwaj/Documents/projects/sonicPiWeb/.captures/desktop_2026-06-18T15-10-15-453Z_comm-in-thread-forum__43_exploring_tb303.md
✓ WAV: /Users/mrityunjaybhardwaj/Documents/projects/sonicPiWeb/.captures/desktop-recordings/desktop_2026-06-18T15-10-15-453Z_comm-in-thread-forum__43_exploring_tb303.wav
 31.37s · peak 1 · RMS 0.306 · clip 1.99%
```
### Web
```
Launching Chromium (headed, audio capture)...
 Running: inline (30000ms)...

Capture saved: /Users/mrityunjaybhardwaj/Documents/projects/sonicPiWeb/.captures/inline.md
No errors detected.
```