Feature ideas: lightweight DSP effects and new sequencer generators

[gospodindilly-ops]

I’d like to suggest two additions that could expand Felucca’s sound design and generative workflow without requiring heavy DSP.

1. Lightweight DSP effects

Potential effects worth exploring:

• Comb Resonator — a short feedback delay with adjustable tuning and damping. Example: turn a drum hit into a metallic, pitched resonance.
• Wavefolder with ADAA — adds complex harmonics to simple waveforms. Antiderivative anti-aliasing can reduce aliasing without the CPU cost of conventional oversampling. Example: turn a sine bass into an aggressive West Coast-style lead.
• Three-band Isolator — independent LOW / MID / HIGH controls for fast performance cuts. Example: remove the bass and bring it back on the next beat.
• Tilt EQ — one control shifts the balance between low and high frequencies. Example: quickly darken a pattern before a transition.

Comb Resonator and Tilt EQ should be especially inexpensive. Wavefolder + ADAA is a more experimental option and should be benchmarked on the target hardware. References: [ADAA research (DAFx 2025)](https://hdl.handle.net/11566/352938), [lightweight DSP examples](https://github.com/im9/ideath).

2. New sequencer generators

Rather than introducing incompatible sequencer formats, generators could create patterns that remain editable in the existing 64-step sequencer.

• Euclidean — distribute a chosen number of hits across a cycle. Example: 5 hits over 16 steps, with rotation.
• Memory Probability — probability changes based on recent events. Example: repeated hi-hat hits gradually become less likely, creating more natural variation.
• Markov Melody — the next note depends on the previous note, constrained by the selected scale. Example: generate evolving melodies using repeat, step, jump and rest weights.
• Conditional Steps — trigger notes only on certain pattern cycles or during fills. Example: a percussion hit plays every fourth loop.

A useful workflow would be GENERATE → LOCK → EDIT: generate a pattern, convert it into normal steps, then edit notes, parameter locks and automation manually.

I’d prioritize Comb Resonator and Euclidean generation first, then benchmark the wavefolder and explore stateful probability. CPU usage and audio quality should be measured on the target device before merging.

[kurogedelic]

1.6 is out: https://github.com/hugelton/Felucca/releases/tag/v1.6

Thanks for the detailed list. I added a Euclidean generator to SEQ TOOLS: hold SEQ and press F4. KNOB 1–3 set hits, steps and rotation, and the pattern is written right away so you can hear it while it plays. OCT+ keeps it, OCT− cancels it. On a DRUM track it writes the selected lane; on a melodic track the hits keep the step's own note (or the root), and the result is ordinary steps you can edit afterwards, much like the GENERATE → EDIT flow you described. Conditional steps (1:2, 1:4, FILL …) are in 1.6 too (see #209), and PHASE got a wavefolder (FOLD, see #104). It's a sine fold that folds less on high notes, not ADAA. The comb resonator, isolator, tilt EQ, memory probability and Markov ideas are noted for later.

[gospodindilly-ops]

what a nice day!