fix(dub): auto-assign per-speaker cloned voices to segments (#486) (#576)

Multi-speaker dubbing diarizes the speakers and clones each one from the video
(the Voice dropdown shows "From Video → Speaker 1 / Speaker 2"), but every
segment was left on "Default" — the user had to set the voice on each row by
hand. The clone→segment binding simply never happened: the transcribe `final`
handler stored `speaker_clones` but set the segments without filling their
`profile_id`.

Bind them up front: new `applySpeakerCloneDefaults(segments, speakerClones)`
sets each segment's `profile_id` to its speaker's `auto:<safe>` clone id when a
clone exists and the user hasn't already chosen a voice. The id is computed by
`autoProfileId()`, which mirrors the backend clone-resolution key
(`speaker_id.lower().replace(" ","_")`) and the DubTab dropdown option value, so
all three agree. Only an *empty* profile_id is filled — an explicit per-speaker
or per-segment choice is never clobbered.

Pure helper + unit test (assign-when-cloned, never-clobber, no-clone-stays-
Default, no-op-without-clones).

Note: the issue's second symptom — different speakers' turns merged onto one
line — is a separate diarization/segment-grouping concern (speaker-turn
re-split) tracked as a follow-up; this fixes the per-speaker voice assignment.

Co-authored-by: mergetest <test@local>
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

frontend/src/hooks/useDubWorkflow.js

@@ -6,7 +6,7 @@ import {
   transcribeStreamUrl, dubImportSrt,
 } from '../api/dub';
 import { dialectMatchesLang } from '../api/dialects';
-import { segmentGenInputs } from '../utils/segments';
+import { segmentGenInputs, applySpeakerCloneDefaults } from '../utils/segments';
 import { apiPost } from '../api/client';
 import { API } from '../api/client';
 import { playPing, isTauri } from '../utils/media';
@@ -100,11 +100,14 @@ export default function useDubWorkflow({ loadProjects, loadProfiles, loadDubHist
       try {
         const m = JSON.parse(e.data);
         gotFinal = true;
-        setDubSegments((m.segments || []).map((s, i) => ({
+        const normalized = (m.segments || []).map((s, i) => ({
           ...s,
           id: s.id != null ? String(s.id) : String(i),
           text_original: s.text_original || s.text || '',
-        })));
+        }));
+        // #486: bind each segment to its detected speaker's clone up front, so
+        // a 2-speaker dub doesn't land every row on "Default".
+        setDubSegments(applySpeakerCloneDefaults(normalized, m.speaker_clones));
         setDubTranscript(m.full_transcript || '');
         if (m.speaker_clones && typeof m.speaker_clones === 'object') {
           setSpeakerClones(m.speaker_clones);

frontend/src/test/segments.speakerClone.test.js

@@ -0,0 +1,35 @@
+import { describe, it, expect } from 'vitest';
+import { applySpeakerCloneDefaults, autoProfileId } from '../utils/segments';
+
+// #486: multi-speaker dub must auto-bind each segment to its detected
+// speaker's cloned voice instead of leaving every row on "Default".
+describe('applySpeakerCloneDefaults (#486)', () => {
+  const clones = { 'Speaker 1': { duration: 4.2 }, 'Speaker 2': { duration: 3.1 } };
+
+  it('assigns the auto: clone id to segments whose speaker was cloned', () => {
+    const segs = [
+      { id: '0', speaker_id: 'Speaker 1', profile_id: '' },
+      { id: '1', speaker_id: 'Speaker 2', profile_id: '' },
+    ];
+    const out = applySpeakerCloneDefaults(segs, clones);
+    expect(out[0].profile_id).toBe('auto:speaker_1');
+    expect(out[1].profile_id).toBe('auto:speaker_2');
+    expect(autoProfileId('Speaker 2')).toBe('auto:speaker_2');
+  });
+
+  it('never clobbers a profile_id the user already chose', () => {
+    const segs = [{ id: '0', speaker_id: 'Speaker 1', profile_id: 'preset:narrator' }];
+    expect(applySpeakerCloneDefaults(segs, clones)[0].profile_id).toBe('preset:narrator');
+  });
+
+  it('leaves a segment on Default when its speaker has no clone', () => {
+    const segs = [{ id: '0', speaker_id: 'Speaker 9', profile_id: '' }];
+    expect(applySpeakerCloneDefaults(segs, clones)[0].profile_id).toBe('');
+  });
+
+  it('is a no-op when there are no clones', () => {
+    const segs = [{ id: '0', speaker_id: 'Speaker 1', profile_id: '' }];
+    expect(applySpeakerCloneDefaults(segs, {})[0].profile_id).toBe('');
+    expect(applySpeakerCloneDefaults(segs, null)).toEqual(segs);
+  });
+});

frontend/src/utils/segments.js

@@ -34,3 +34,27 @@ export function segmentGenInputs(s) {
     effect_preset: s.effect_preset || undefined,
   };
 }
+
+/** The `auto:<safe>` profile id for a diarized speaker. Mirrors the backend's
+ *  clone-resolution key (`speaker_id.lower().replace(" ", "_")`) and the
+ *  Voice-dropdown option value in DubTab, so the three always agree. */
+export function autoProfileId(speakerId) {
+  return `auto:${(speakerId || '').toLowerCase().replace(/\s+/g, '_')}`;
+}
+
+/**
+ * #486: auto-assign each diarized segment to its detected speaker's cloned
+ * voice instead of leaving it on "Default". When the backend cloned a speaker
+ * from the video (`speakerClones[speaker_id]` present) and the segment has no
+ * voice chosen yet, default its `profile_id` to that speaker's `auto:` clone.
+ * The user can still override per-speaker or per-segment afterwards — we only
+ * fill an *empty* profile_id, never clobber an explicit choice.
+ */
+export function applySpeakerCloneDefaults(segments, speakerClones) {
+  const clones = (speakerClones && typeof speakerClones === 'object') ? speakerClones : {};
+  if (!Array.isArray(segments) || !Object.keys(clones).length) return segments || [];
+  return segments.map((s) => {
+    if (!s || s.profile_id || !s.speaker_id || !clones[s.speaker_id]) return s;
+    return { ...s, profile_id: autoProfileId(s.speaker_id) };
+  });
+}