amy.c: Chained oscs are evaluated backwards from the end of the chain. - #1200
Merged
Merged
Conversation
Contributor
🎛️ AMY HW CI (AMYboard bench)Flashed this PR's AMY (LoadTestChord: 6-voice Juno ✅ PASS — the bench ran the test to completion.
Full chord settled render μs: 2614 (was 2627, Δ -0.5%) (peak 2618, 39 samples) ⬇️ Artifacts: serial log · load trace · report Self-hosted bench (amyboardci). FAIL means only that the test could not run — the load values are informational, with no threshold and no audio compare. See |
Collaborator
⛓️ tulipcc integration PR openedThis merge was pinned into tulipcc for full-system CI: shorepine/tulipcc#1382 Test it there and merge that PR to move tulipcc onto this AMY. |
MinoruInachi
added a commit
to MinoruInachi/amy
that referenced
this pull request
Oct 4, 2026
Brings in amy up to 1.2.190 (e153f7d): the render lock (shorepine#1205), chained oscs evaluated tail-first (shorepine#1200), MIDI CC output and midi_cc naming an AMY parameter directly (shorepine#1175), a PCM render gain ramp (shorepine#1208), no osc allocation on the ingest path, and the docs that go with them. One conflict, in src/amy.c: upstream grew its own lock abstraction for ESP_PLATFORM (a FreeRTOS mutex behind lock_init/lock_take/lock_give) where we had a bare SemaphoreHandle_t plus esp_rom_sys.h. Took upstream's side, and dropped our a84e093 render-in-flight counter with it: upstream's render lock (f11a552, 2812743) closes the same FREE_OSC-under-render race more thoroughly -- the render thread holds it for a whole block and a patch load holds it across its flush -- and it was measured, where ours spun for up to 50 ms. Nothing else of ours changed; src/amy.c is now identical to upstream/main. f11a552 lists tulipcc's native render loops as needing the wrap themselves, but they do not: every tulipcc path reaches a render through an amy entry point that already takes the lock. The Tab5 sets platform.multithread = 0, so its audio task's amy_render_audio() takes the branch that grabs it around esp_render_on_cores() + amy_fill_buffer(); Tulip CC keeps the default multithread = 1 and renders in amy's own esp_fill_audio_buffer_task; AMYboard-in-VCV pulls blocks with amy_simple_fill_buffer(). All three are wrapped upstream. make ctest: all passed. make test: 135 tests pass, no failures, including our TestAllNotesOffNoteZero at err=-100.0 dB. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Claude informed me that, counter to my understanding, filter (and distortion) applied to the chain up until that point in the chain; I expected them to apply to everything after the filtering osc (so a filter osc at the head of a chain filters the entire chain). I'm not sure when this got turned around (since that's how the original Juno voices worked), but the fix (head-recursing on chained osc) simplified the code, since the only thing special about a SILENT osc now is that it applies its envelope to the cumulated sum.