V2#7
Open
joegiralt wants to merge 72 commits into
Open
Conversation
Marzhall/build instructions
replace overlapping structures with pointers for latest avr-gcc
suppress undefined behaviour warning
suppress undefined behaviour warning
formant_amplitude[3] is an alias of noise_modulation
comment apparent out-of-bounds loop
Replaces the stock Ambika oscillator set with a focused synthesis engine: saw, square, triangle, sine, pad, filtered noise, 4-op FM, Karplus-Strong plucked string, waveshaping, and west coast wavefolder. New synthesis engines with dedicated UI pages: - 4-op FM with 8 TX81Z algorithms, 8 waveforms per operator, feedback - Karplus-Strong with 4 excitation types, 3-read-head ensemble, stiffness - West coast wavefolder with variable fold stages, FM, self-sync, sub osc - Waveshaping through distortion table with variable drive Envelope system expanded with per-envelope curve selection (exponential, linear, looping, looping linear). Analog slop parameter adds per-note random pitch and envelope variation for vintage poly character. Smart patch randomizer with per-section depth control. Step sequencer removed to free controller flash. S1+encoder navigation for synthesis mode switching. Bug fixes: note stack uninitialized variables, transient generator off-by-one, SPI bulk write bounds check, page group init size.
- Add ExtraEnvelopeSettings struct (4 bytes: ADSR) and extra_env[4] array at offsets 112-127, expanding Patch from 112 to 128 bytes - Add MOD_SRC_ENV_4 through MOD_SRC_ENV_7 to ModulationSource enum - Remove MOD_SRC_SEQ_1 and MOD_SRC_SEQ_2 (sequencer sources) - Rename kNumEnvelopes to kNumEnvLfoSlots (value stays 3) - Add kNumTotalEnvelopes=7, kNumExtraEnvelopes=4, kNumLfos=3 - Add PRM_PATCH_EXTRA_ENV_* parameter enum entries Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Expand envelope array from 3 to kNumTotalEnvelopes (7), add rendering loop for all 7 mod sources (ENV_1-ENV_7), update ADSR parameters from both env_lfo[] (slots 0-2) and extra_env[] (slots 3-6), extend COMMAND_RETRIGGER_ENVELOPE cases, and append default ADSR values for the 4 extra envelopes in init_patch. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace removed MOD_SRC_SEQ_1/SEQ_2 enum values with MOD_SRC_ENV_4/ENV_5 in init_patch PROGMEM data, and add 16 bytes of extra_env[4] ADSR defaults to match the expanded 128-byte Patch struct. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace the ParameterEditor-based ENV_LFO page with a custom EnvPage that supports selecting all 7 envelopes. Envelopes 1-3 show LFO controls (rate, shape, curve) plus ADSR. Envelopes 4-7 show ADSR only with blank top-row cells where LFO controls would be. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…set engine state on mode change
Major FM4Op improvements based on analysis of TX81Z reference samples and ymfm/SC/ROM disassembly sources: - Exponential output level curve (0.75 dB/step, 128-byte PROGMEM table) Applied in voice.cc before envelope multiply, matching TX81Z signal chain: raw level → exp lookup → × envelope → amplitude - ScaleMod with no shift (was >> 4) for full 8-bit modulation depth, giving mod index up to ~3.1 radians at max level - ScaleCarrier on all 8 algorithms — per-op envelopes now actually control carrier output volume (was completely ignored before) - 16-bit feedback state (was int8_t) prevents the feedback loop from locking into a fixed repeating pattern, creating evolving oscillation closer to the TX81Z's 14-bit feedback dynamics - 3-carrier mix (ALG 6,7) changed from /3 division to >> 2 + DC correction, avoiding slow AVR integer division Flash: 32446 text + 82 data = 32528 (240 bytes free) RAM: +2 bytes for 16-bit feedback state Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replaced the 2-tap FIR damping filter with a one-pole IIR low-pass. The FIR barely changed the tone across its full range — now damping sweeps from bright metallic ring to dark muted thud. - Damping: IIR cutoff control (was FIR coefficient with uint8_t overflow bug at values > 111) - Decay: applied per delay-line cycle, not per sample (was killing the sound in <2ms at any non-zero value) - Feedback: reduces damping loss for longer sustain instead of adding gain (was self-oscillating at value 16+) - Stiffness: offset scales with delay length (was fixed 7 samples, wildly different effect at different pitches) - Body: halved range for stability (was unstable above 50%) +2 bytes RAM (int16_t lp_state_), fits in flash (198 bytes free). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Doubled read head deviation (>> 7 instead of >> 8) and spread (>> 1 instead of >> 2) for more dramatic chorus/detune effect at high wmix settings. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Wavefolder: - Quadratic gain curve (1x-33x) replaces linear 1x-4x. Now creates 16+ folds at max vs previous 2-3. Dramatically richer harmonic content matching real Buchla 259 character. - Removed separate fold stages parameter (folds determined automatically by gain, up to 16 iterations) - Drive and input gain properly combine with fold depth Color filter: - 2-pole IIR (12dB/oct) replaces single-pole (6dB/oct). The steep rolloff actually changes the timbre — sweeps from dark fundamental-only to bright full harmonics. - Min coefficient 16 ensures fundamental passes at darkest setting Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
FM: TX81Z-accurate modulation, 16-bit feedback, exponential levels KS: IIR damping, fixed parameter scaling WC: Buchla-style quadratic wavefolder, 2-pole color filter Flash optimization to fit within 32256-byte bootloader limit: - Removed ScaleCarrier from all algorithms (VCA handles carrier volume) - Shrunk exp level table from 128 to 64 entries (lower half is all 1s) - Total: 32228 bytes (28 bytes free) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Power-on and init now start in FM4OP mode with algorithm 1, all-sine operators at 1:1 ratio, moderate levels. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Trigger was only filling delay_line[0..delay_length_], but delay_length_ hadn't been updated for the new note yet (SetPitch runs later in RenderOscillators). Result: every note used the previous note's buffer length, making all notes sound the same pitch. Fix: fill the entire 192-sample buffer on trigger. The Render loop already wraps at the correct delay_length_ set by SetPitch. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Bug fixes: - Fix MIDI channel display: channels 1-16 show numbers instead of garbage strings (amnt, srce, oct...) - Fix engine switch: fully isolate shared patch bytes between all engines with per-engine state slots (save/restore on switch) - Fix voicecard frame drops on engine switch: only init the active engine, avoid clearing KS delay line unconditionally in audio ISR - Fix boot page: show engine-appropriate page after multi loads Improvements: - Per-engine mod destination names in mod matrix (algo/lvl1-4 for FM, damp/body/stif for KS, fold/sym/bias for WC) - EEPROM auto-reset on firmware upgrade via version marker - Octave range extended to -4/+4, default -1 - FM init levels reduced to gentle defaults (20/10/10/1) - Pitch slop gentler, envelope slop stronger - Voicecard init patch matches controller (FM4OP default) - Controller test harness for parameter display and engine switch logic Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <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
4 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.
No description provided.