Skip to content

TODOS.md

Latest commit

 

History

History
49 lines (42 loc) · 2.7 KB

TODOS.md

File metadata and controls

49 lines (42 loc) · 2.7 KB
Raw

TODOS

P2: Restore from archive

  • What: Inverse of the archive step: browse the archive folder and copy a project back into the local working store.
  • Why: Completes the archive round trip; today restore is a manual folder copy.
  • Effort: S (CC). Depends on: archive dir configured and used for real.

P3: Conditioning profile UI

  • What: Per-project toggles for denoise (afftdn/RNNoise/off), trim, loudnorm and padding; today the defaults from server/src/pipeline/condition.ts apply to all. Dropping an .rnnn model into ~/LessonStudio/models/ already upgrades the denoiser.
  • Why: Dave should tune the chain once after hearing his mic through it.
  • Effort: S (CC). Depends on: first real recorded lesson.

Deferred scope from the gstack CEO review (2026-07-12). Context for each item assumes HANDOVER.md and PLAN.md have been read.

P2: Publish hook (Moodle/NAS course folders)

  • What: One-click copy of a produced MP4+SRT into the Moodle upload staging folder or the NAS course directory.
  • Why: Closes the last manual step between "video produced" and "video live in the course".
  • Pros: Completes the production line; pairs naturally with the M6 archive step.
  • Cons: Needs the NAS path decision and Moodle upload conventions settled first.
  • Effort: M (human) / S (CC). Depends on: M6 archive step, NAS path decision.

P3: TTS voice for scripts

  • What: Local or API TTS rendering of script blocks, as preview or as production voice (see CEO review finding F1; production TTS would bypass recording entirely).
  • Why: Instant re-produce after script edits without re-recording.
  • Pros: Removes recording time from the loop for lessons where a cloned voice is acceptable.
  • Cons: New infrastructure; voice-clone quality bar unproven for Dave's voice; challenges the settled audio-recording decision.
  • Effort: L (human) / M (CC). Depends on: F1 gate decision (spike first).

P3: Project templates

  • What: Per-course defaults: intro/outro slides, standard padding, naming pattern.
  • Why: Saves setup on lesson 2 through 34 of a course.
  • Cons: Only pays off at course scale; premature before several real projects.
  • Effort: M (human) / S (CC). Depends on: several projects produced for real.

P3: Waveform preview + take audition UI

  • What: Visual waveform per take with click-to-audition before picking the active take. Default disposition from the final gate: deferred here unless Dave opts in.
  • Why: Faster take selection than listening end to end.
  • Cons: Waveform renderer dependency; takes list with durations covers most need.
  • Effort: M (human) / S (CC). Depends on: M4 take management shipped.