P3: Deep Sample and Loop Intelligence

[thunderdew-dawn]

Source: ROADMAP.md P3.

Scope:

• Research deeper sample and loop intelligence within exposed FL/audio-analysis limits.
• Known API limitations around deep sample parameters, Stretch Mode, and Normalize must remain explicit.

Acceptance criteria:

• Capability evidence documented before any user-facing promise.
• Output separates measured audio analysis from FL API-limited sample settings.

[thunderdew-dawn]

I think this issue should be implemented primarily as an offline Python audio-analysis workflow, not as an FL Studio sampler-parameter feature.

Most useful sample/loop intelligence can be derived safely from the audio file itself: duration, BPM estimate, beat count, onset/transient density, loop length in bars, key/chroma estimate, spectral balance, low-end ratio, stereo width/correlation, peak/RMS/LUFS, silence/tail detection, and possible loop-boundary clicks.

This also fits the direction of the existing soundwalker component, which is already being developed and tested in pre-alpha. soundwalker is intended to provide deterministic offline audio analysis outside the live FL Studio API path, and should become a natural backend/component for this kind of sample and loop intelligence once it is added to fls-pilot.

FL-specific sampler state such as Stretch Mode, Normalize, internal time-stretch settings, or channel sampler parameters should remain explicitly API-limited unless proven by live FL API evidence. The output should clearly separate “measured from audio / soundwalker analysis” from “not available from FL API”.