Scope#
WS08-007 exposes first-class audio pipeline controls and preview surfaces in
ISIS web UX so operators and creative users can configure audio conditioning /
voice consistency settings and directly preview audio artifacts during review.
Research and Existing Implementation Alignment#
Before implementation, existing ISIS web surfaces were audited:
- Job submission controls and request builder:
apps/isis/web/src/pages/JobSubmissionStudio.tsxapps/isis/web/src/pages/job-submission-utils.ts
- Job review and sequence diff surfaces:
apps/isis/web/src/pages/JobsPage.tsxapps/isis/web/src/pages/jobs-sequence-utils.ts
- API contracts for audio/voice controls and outputs:
libs/isis/client/src/types.ts
This ensured UI changes mapped onto already-implemented backend contracts without introducing parameter drift.
Deliverables#
- Audio pipeline controls in Job Submission Studio
- Added dedicated
Audio Conditioningcontrol section for audio generation types (text-to-audio,voice-synthesis,music-generation). - Added dedicated
Voice Consistency Controlssection for voice-centric flows (voice-synthesis,music-generation). - Added typed controls for:
- conditioning mode/provider mode/strength/prompt hints/reference audio URLs
- voice consistency profile/voice/speaker IDs, embedding url/vector, pronunciation assets, style tokens, consistency strength
- Request builder and form model expansion
- Extended
JobStudioFormValuesand defaults with audio and voice control fields. - Added parser/validator support for audio URLs, embedding vectors, and pronunciation dictionary locator payloads.
- Extended
buildStudioSubmissionRequestto emit normalizedaudioConditioning/audio_conditioningandvoiceConsistency/voice_consistencyparameter payloads.
- Job clone/prefill parity for audio controls
- Added extraction + mapping utilities so cloned job configs hydrate audio and voice control sections in Studio forms.
- Added alias stripping support to avoid duplicate audio payloads in extra parameter JSON.
- Audio preview players in job review
- Added
audioasset classification in sequence asset utilities. - Extended sequence asset collection to include audio package artifacts:
- mixdown
- stems
- multi-track files
- Added
<audio controls>preview player rendering inJobsPageside-by-side asset diff tool.
- Regression coverage
- Added utility tests for audio/voice payload build + alias extraction/mapping.
- Added Job Submission Studio test for audio control rendering and submission payload behavior.
- Added sequence asset tests for audio package asset collection.
- Added Jobs page test validating audio preview rendering in sequence review.
Verification Commands#
bash
pnpm --filter @isis/web exec vitest run src/pages/job-submission-utils.spec.ts src/pages/JobSubmissionStudio.spec.tsx src/pages/jobs-sequence-utils.spec.ts src/pages/JobsPage.spec.tsx --reporter=dot
node scripts/isis/generate-remediation-backlog-v2.mjs
node scripts/isis/validate-v2-compliance-gate.mjs
Result#
WS08-007 is now implemented with explicit audio pipeline control surfaces in
submission UX and first-class in-app audio preview playback for sequence/job
review flows.