Issue #720: Rework zero-amp termination - now only in release. Skip …

[dpwe]

…oscs where amp[0] == 0. Change last_amp.

Issue #720 responded to a couple of instances where voices with very slow attacks were never being started because of some logic that rapidly terminates new oscs that don't appear to be emitting sound, in order to avoid wasted computation on Juno patches. Instead, we now skip rendering oscs whose amp[CONST]==0 (we already skipped them if their msynth.amp and msynth.last_amp were zero, but this is more aggressive and avoids running hold_and_modify).

In testing this, I noticed that some AM envelopes were very noisy because we had msynth.last_amp logic that failed to linearly interpolate rising AM envelopes. I forget why I put this in, I think it was to do with voice-stealing clicks, but anyway I simplified it and most tests were unchanged. I listened to all the changed tests (with the new make playfailed) and they all sounded OK.

[bwhitman]

🎛️ AMYboard HW CI

Dispatched this PR's amy SHA to the tulipcc bench — it builds the AMYboard firmware with your change and runs it on the physical board (audio spectral-compared to a committed reference). Results post here in a few minutes.

[dpwe]

I tested this build against the previous version for the number of simultaneous Juno 1 voices that can be rendered on an RP2350 (4), and it's surely no worse than it was before.

[bwhitman]

🎛️ HW CI (physical bench)

Built from this AMY PR, pinned into the tulipcc amy submodule.

AMYboard (USB-MIDI + AMY zP → audio): ✅ PASS — flashed this PR’s firmware; all checks matched the references.

Tulip (TULIP4_R11; serial-REPL audio + WiFi screenshot): ✅ PASS — flashed this PR’s firmware; all checks matched the references.

⬇️ Artifacts: recordings · screenshot · serial logs · run logs

_{Self-hosted bench. Audio spectral-compared to ref/hwci_basic.wav + ref/tulip_basic.wav; Tulip screenshot pixel-compared to ref/tulip_screenshot.png. Both analog outs share one capture card, so the tests run sequentially.}