diff --git a/CHANGELOG.md b/CHANGELOG.md
index 923dd0d0..328aec8b 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -27,6 +27,7 @@ the frozen-backend fallback mirror it for their toolchains.
 
 ### Fixed
 
+- Dubbing repairs missing speech caches, rejects incomplete output, avoids oversized speaker references, and fits full speech without early clipping (#2129)
 - Workspace sidebars have a working right-edge resize handle, allow 40% more width, remember their size, and keep video controls inside the preview (#2129)
 - Pressing Play while a video is loading starts playback when it is ready instead of reporting playback unavailable (#2129)
 - Video previews show their thumbnail before playback, including the source video in Dub (#2129)
diff --git a/backend/api/routers/dub_generate.py b/backend/api/routers/dub_generate.py
index a6064666..647d5494 100644
--- a/backend/api/routers/dub_generate.py
+++ b/backend/api/routers/dub_generate.py
@@ -30,7 +30,7 @@ from services.ffmpeg_utils import (
 )
 from services.rvc import apply_rvc, is_enabled as rvc_is_enabled
 from services.incremental import segment_fingerprint, fit_fingerprint
-from services.fit_planner import UNDERRUN_TOLERANCE, FitParams, plan_fit
+from services.fit_planner import FitParams, plan_fit
 from services.watermark import mark_synthetic
 from services.speaker_clone import auto_profile_id
 from services.segment_bundle import extract_segment_wavs
@@ -39,16 +39,6 @@ from omnivoice.utils.voice_design import heal_design_instruct
 
 logger = logging.getLogger("omnivoice.dub")
 
-# Maximum compression ratio we'll attempt with pitch-preserving stretch
-# before declaring "no way to fit cleanly" and falling back. atempo
-# remains intelligible up to ~1.5× then introduces audible WSOLA
-# artefacts; above ~1.8× speech becomes a fast garbled stream that no
-# DSP can rescue. The contributing-factor pipeline (CPS-aware slot-fit
-# in services/speech_rate.py, gap absorption below) keeps us under this
-# in practice — this is only a guard rail.
-MAX_STRETCH_RATIO = 1.8
-
-
 class _RemoteDubBackend:
     """Sample-rate carrier while Dubbing runs without local TTS weights."""
 
@@ -378,6 +368,12 @@ def forget_missing_ref_warnings(job_id: str) -> None:
     _MISSING_REF_WARNED.pop(str(job_id), None)
 
 
+
+def _ref_within_limit(info) -> bool:
+    from services.speaker_clone import MAX_REF_DURATION_S
+    return bool(info) and float(info.get("duration") or 0.0) <= MAX_REF_DURATION_S
+
+
 def resolve_consistent_ref(job: dict, speaker_key: str, memo: dict | None = None):
     """ONE clone reference for every segment of `speaker_key`.
 
@@ -386,8 +382,9 @@ def resolve_consistent_ref(job: dict, speaker_key: str, memo: dict | None = None
          the per-line path uses as its fallback;
       2. no speaker clone (heuristic diarization skips extraction entirely —
          the key case): a deterministic pick among that speaker's per-segment
-         clips: longest clip ≥3 s, tie-break lowest segment id. Clips all
-         shorter than 3 s degrade to "longest overall", same tie-break.
+         clips: longest usable clip ≥3 s within the shared reference limit,
+         tie-break lowest segment id. Short clips fall back to longest usable.
+         Oversized references must never strand every short line for a speaker.
 
     Returns the clone info dict ({"ref_audio", "ref_text", ...}) or None.
     Pure function of the job dict; `memo` (keyed by speaker_key) just avoids
@@ -396,7 +393,11 @@ def resolve_consistent_ref(job: dict, speaker_key: str, memo: dict | None = None
     if memo is not None and speaker_key in memo:
         return memo[speaker_key]
 
+    from services.speaker_clone import MAX_REF_DURATION_S
+
     ref = _find_speaker_clone(job.get("speaker_clones") or {}, speaker_key)
+    if ref and float(ref.get("duration") or 0.0) > MAX_REF_DURATION_S:
+        ref = None
     if ref is None:
         seg_clones = job.get("segment_clones") or {}
         candidates = []
@@ -408,7 +409,8 @@ def resolve_consistent_ref(job: dict, speaker_key: str, memo: dict | None = None
                 continue
             sid = str(row.get("id", ""))
             info = seg_clones.get(sid)
-            if info and info.get("ref_audio"):
+            if (info and info.get("ref_audio")
+                    and float(info.get("duration") or 0.0) <= MAX_REF_DURATION_S):
                 candidates.append((sid, info))
         if candidates:
             usable = [
@@ -452,6 +454,9 @@ def _remote_voice(job: dict, profile_id: str | None, seg_id, voice_match: str,
             info = ((job.get("segment_clones") or {}).get(str(seg_id))
                     or _find_speaker_clone(job.get("speaker_clones") or {}, key))
             single_use = str(seg_id) in (job.get("segment_clones") or {})
+            if not _ref_within_limit(info):
+                info = resolve_consistent_ref(job, key, memo)
+                single_use = False
         if info:
             ref_audio, ref_text = info.get("ref_audio"), info.get("ref_text")
     elif profile_id:
@@ -701,8 +706,29 @@ async def dub_generate(job_id: str, req: DubRequest):
         _wav_kind = (
             _kind_map.get(lang_code) if isinstance(_kind_map, dict) else job.get("seg_wav_kind")
         )
-        if strategy != "strict_slot" and regen_only is not None and _wav_kind != "natural":
+        if regen_only is not None and _wav_kind != "natural":
             regen_only = None
+        # A partial rerun must repair absent or corrupt caches, including before
+        # remote batching decides which lines need synthesis.
+        if regen_only is not None:
+            for index, segment in enumerate(req.segments):
+                sid = seg_ids[index] if index < len(seg_ids) else f"seg_{index}"
+                if sid in regen_only or not segment.text.strip():
+                    continue
+                cache = _seg_lang_path(sid)
+                if not os.path.exists(cache) and _legacy_seg_cache_ok(job, lang_code):
+                    for key in (sid, index):
+                        legacy = dub_seg_path(job_id, key)
+                        if os.path.exists(legacy):
+                            cache = legacy
+                            break
+                try:
+                    info = torchaudio.info(cache)
+                    intact = info.num_frames > 0 and _cached_payload_intact(cache, info)
+                except Exception:
+                    intact = False
+                if not intact:
+                    regen_only.add(sid)
         # Manifest: stable segment id per current index. Per-segment WAVs are
         # named by stable id (dub_seg_path) so regen reuses the right audio after
         # reorder; index-keyed readers (preview/export) resolve via this manifest.
@@ -955,7 +981,7 @@ async def dub_generate(job_id: str, req: DubRequest):
                                 all_segment_wavs.append(
                                     (seg.start, seg.end, seg_wav_path, backend.sample_rate)
                                 )
-                                sync_scores.append(getattr(seg, 'sync_ratio', None) or 1.0)
+                                sync_scores.append(round(cached_info.num_frames / cached_info.sample_rate / max(seg_duration, 0.01), 3))
                                 _t_cache += time.perf_counter() - _t_cache_0
                                 continue
 
@@ -963,39 +989,20 @@ async def dub_generate(job_id: str, req: DubRequest):
                         if cached_sr != backend.sample_rate:
                             import torchaudio.functional as AF
                             cached_wav = AF.resample(cached_wav, cached_sr, backend.sample_rate)
-                        # strict_slot persists slot-sized buffers.  Every other
-                        # strategy consumes natural-rate audio and lets the mix
-                        # loop fit it to the current timeline.
-                        if strategy == "strict_slot":
-                            target_samples = int(seg_duration * backend.sample_rate)
-                            current_samples = cached_wav.shape[-1]
-                            if target_samples > current_samples:
-                                cached_wav = torch.nn.functional.pad(cached_wav, (0, target_samples - current_samples))
-                            elif current_samples > target_samples:
-                                cached_wav = cached_wav[..., :target_samples]
+                        cached_ratio = round(cached_wav.shape[-1] / backend.sample_rate / max(seg_duration, 0.01), 3)
                         all_segment_wavs.append(_store_mix_wav(seg.start, seg.end, cached_wav, backend.sample_rate, f"mix_{seg_id}"))
                         try:
                             del cached_wav
                         except Exception:
                             pass
                         _release_audio_tensors()
-                        sync_scores.append(getattr(seg, 'sync_ratio', None) or 1.0)
+                        sync_scores.append(cached_ratio)
                         _t_cache += time.perf_counter() - _t_cache_0
                         continue
-                    except Exception as e:
-                        # Fall through to a silent placeholder if the cached WAV
-                        # is broken — cleaner than aborting the whole mix.
-                        yield f"data: {json.dumps({'type': 'warning', 'segment': i, 'message': f'cached seg lost, padding silence: {str(e)[:120]}'})}\n\n"
-                sr = backend.sample_rate
-                silence = torch.zeros(1, max(0, int(seg_duration * sr)))
-                all_segment_wavs.append(_store_mix_wav(seg.start, seg.end, silence, sr, f"mix_{seg_id}"))
-                try:
-                    del silence
-                except Exception:
-                    pass
-                _release_audio_tensors()
-                sync_scores.append(1.0)
-                continue
+                    except Exception:
+                        logger.exception("Dub cached segment could not be read: %s", seg_id)
+                yield f"data: {json.dumps({'type': 'error', 'segment': i, 'segment_id': seg_id, 'error_code': 'dub_speech_missing', 'error': 'Cached speech could not be read. Regenerate this segment before exporting.'})}\n\n"
+                return
 
             def _gen(text, lang, instruct_str, dur_s, nstep, cfg, spd, profile_id, effect_preset,
                      *, execution_target="local", prepare_only=False, current_seg_id=None):
@@ -1092,7 +1099,7 @@ async def dub_generate(job_id: str, req: DubRequest):
                             if selected_is_segment_speaker
                             else None
                         )
-                        if seg_ref:
+                        if _ref_within_limit(seg_ref):
                             ref_audio = seg_ref.get("ref_audio")
                             ref_text = seg_ref.get("ref_text")
                             ref_single_use = True
@@ -1100,7 +1107,7 @@ async def dub_generate(job_id: str, req: DubRequest):
                             auto = _find_speaker_clone(
                                 job.get("speaker_clones") or {}, key
                             )
-                            if auto is None:
+                            if not _ref_within_limit(auto):
                                 # Short lines may have no line-specific clip.
                                 # Reuse this speaker's best source instead of
                                 # silently reverting to the engine default.
@@ -1458,27 +1465,17 @@ async def dub_generate(job_id: str, req: DubRequest):
                     yield f"data: {json.dumps({'type': 'cancelled', 'segments_processed': i + 1})}\n\n"
                     return
 
-                target_samples = int(seg_duration * backend.sample_rate)
                 current_samples = audio_tensor.shape[-1]
-                # Capture the real spoken duration before strict-slot padding
-                # or trimming. This is the evidence used by Agent timing and
+                # Capture the real spoken duration before assembly fitting.
+                # This is the evidence used by Agent timing and
                 # keeps sync badges truthful for every timing strategy.
                 natural_generated_dur = current_samples / backend.sample_rate
 
-                if strategy == "strict_slot":
-                    # Legacy: pad short audio + trim long audio so the mix
-                    # loop receives slot-sized buffers. The atempo squeeze
-                    # in the mix loop never fires here because we already
-                    # forced size = target_samples.
-                    if target_samples > current_samples:
-                        pad_amount = target_samples - current_samples
-                        audio_tensor = torch.nn.functional.pad(audio_tensor, (0, pad_amount))
-                    elif current_samples > target_samples:
-                        audio_tensor = audio_tensor[..., :target_samples]
-                # concise / stretch_video / smart_fit: keep audio at its
-                # natural length. The mix loop decides per-mode whether to
-                # trim, slip, stretch the video, or split audio/video
-                # retiming (smart_fit) to accommodate it.
+                # Keep the complete waveform in every cache. Fitting happens
+                # once during assembly; pre-trimming here destroyed words before
+                # the pitch-preserving stretcher could see them.
+                if current_samples == 0 or not torch.isfinite(audio_tensor).all() or not torch.any(audio_tensor.abs() > 1e-6):
+                    raise ValueError("The speech engine returned empty or silent audio")
 
                 generated_dur = natural_generated_dur
                 sync_ratio = round(generated_dur / max(seg_duration, 0.01), 3)
@@ -1488,7 +1485,7 @@ async def dub_generate(job_id: str, req: DubRequest):
                 # Duration-planner calibration sample: this text length spoke
                 # for this long at natural rate. Keyed by stable seg id and
                 # merged into the per-language job map after the loop.
-                if strategy != "strict_slot" and seg.text.strip() and generated_dur > 0:
+                if seg.text.strip() and generated_dur > 0:
                     _natural_dur_records[str(seg_id)] = {
                         "chars": len(seg.text.strip()),
                         "dur": round(generated_dur, 4),
@@ -1526,15 +1523,6 @@ async def dub_generate(job_id: str, req: DubRequest):
                         if rvc_sr == backend.sample_rate:
                             audio_tensor = rvc_wav
 
-                            if strategy == "strict_slot":
-                                target_samples = int(seg_duration * backend.sample_rate)
-                                current_samples = audio_tensor.shape[-1]
-                                if target_samples > current_samples:
-                                    audio_tensor = torch.nn.functional.pad(
-                                        audio_tensor, (0, target_samples - current_samples)
-                                    )
-                                elif current_samples > target_samples:
-                                    audio_tensor = audio_tensor[..., :target_samples]
                     except Exception as e:
                         yield f"data: {json.dumps({'type': 'warning', 'segment': i, 'message': f'RVC skipped: {str(e)[:120]}'})}\n\n"
 
@@ -1581,10 +1569,9 @@ async def dub_generate(job_id: str, req: DubRequest):
                 from core.public_errors import stream_generation_failure
 
                 error_detail = stream_generation_failure(e)["detail"]
-                yield f"data: {json.dumps({'type': 'error', 'segment': i, 'error': error_detail})}\n\n"
-                sr = backend.sample_rate
-                all_segment_wavs.append(_store_mix_wav(seg.start, seg.end, torch.zeros(1, max(0, int(seg_duration * sr))), sr, f"mix_{seg_id}"))
-                sync_scores.append(1.0)
+                logger.exception("Dub generation failed for segment %s", seg_id)
+                yield f"data: {json.dumps({'type': 'error', 'segment': i, 'segment_id': seg_id, 'error': error_detail})}\n\n"
+                return
 
         _t_loop_end = time.perf_counter()
 
@@ -1731,19 +1718,16 @@ async def dub_generate(job_id: str, req: DubRequest):
                 seg_gain = max(0.0, min(2.0, seg_gain))
                 try:
                     wav = _load_entry_wav((start, end, wav_path, sr), sr)
-                except Exception as e:
-                    # A WAV header can be readable while its payload is
-                    # truncated.  Direct cache reuse deliberately defers the
-                    # decode to assembly, so preserve the old recovery contract
-                    # here: warn and fill this slot with silence instead of
-                    # aborting the entire dub.
-                    warning = {
-                        "type": "warning",
-                        "segment": i,
-                        "message": f"cached seg lost, padding silence: {str(e)[:120]}",
-                    }
-                    yield f"data: {json.dumps(warning)}\n\n"
-                    wav = torch.zeros(1, max(0, int((end - start) * sr)))
+                except Exception:
+                    logger.exception("Dub assembly could not read segment %d", i)
+                    yield f"data: {json.dumps({'type': 'error', 'segment': i, 'error_code': 'dub_speech_missing', 'error': 'A speech segment could not be read. Regenerate it before exporting.'})}\n\n"
+                    return
+                from services.audio_dsp import trim_speech_padding
+                if seg_ref is not None and seg_ref.text.strip():
+                    if wav.numel() == 0 or not torch.isfinite(wav).all() or not torch.any(wav.abs() > 1e-6):
+                        yield f"data: {json.dumps({'type': 'error', 'segment': i, 'error_code': 'dub_speech_missing', 'error': 'A speech segment is empty or silent. Regenerate it before exporting.'})}\n\n"
+                        return
+                    wav = trim_speech_padding(wav, sr)
                 adjusted = wav * seg_gain
                 if adjusted.ndim == 2 and adjusted.shape[0] > 1:
                     adjusted = adjusted.mean(dim=0, keepdim=True)
@@ -1791,12 +1775,12 @@ async def dub_generate(job_id: str, req: DubRequest):
                                 align_corners=False,
                             ).squeeze(0)
                         wl = adjusted.shape[-1]
-                    # Residual overflow → hard-trim to the segment's new video
-                    # slot (fade below keeps the cut pop-free).
+                    # Never publish a complete track with speech discarded by
+                    # the fit caps. The user can shorten text or relax the caps.
                     new_slot_samples = int(max(0.0, sf.new_end - sf.new_start) * sr)
-                    if new_slot_samples > 0 and wl > new_slot_samples:
-                        adjusted = adjusted[..., :new_slot_samples]
-                        wl = adjusted.shape[-1]
+                    if new_slot_samples > 0 and wl > new_slot_samples + int(sr * 0.02):
+                        yield f"data: {json.dumps({'type': 'error', 'segment': i, 'error_code': 'dub_timing_overflow', 'error': 'Speech exceeds the fitting limits. Shorten the translation or choose Strict Slot or Stretch Video before exporting.'})}\n\n"
+                        return
                     # Truthful per-segment verdict for the UI badge.
                     entry = {"status": sf.status}
                     if abs(sf.audio_rate - 1.0) > 1e-6:
@@ -1816,8 +1800,8 @@ async def dub_generate(job_id: str, req: DubRequest):
                 elif strategy == "concise":
                     # Mode A: never compress. Allow the audio to extend into the
                     # silent gap before the next seg (existing heuristic) plus
-                    # any extra `overflow_budget_s`. Beyond that, hard-trim with
-                    # a short fade so we never overlap the next speaker.
+                    # any extra `overflow_budget_s`. Beyond that, require a
+                    # timing/text adjustment instead of discarding speech.
                     place_at = start
                     effective_end = end
                     if i + 1 < len(all_segment_wavs):
@@ -1831,47 +1815,25 @@ async def dub_generate(job_id: str, req: DubRequest):
                     slot_samples_eff = int(max(0.0, (effective_end - start)) * sr)
                     if slot_samples_eff > 0 and wl > slot_samples_eff:
                         overflow_s = (wl - slot_samples_eff) / sr
-                        adjusted = adjusted[..., :slot_samples_eff]
-                        wl = adjusted.shape[-1]
-                        fit_status.append({
-                            "status": "overflows",
-                            "overflow_s": round(overflow_s, 3),
-                        })
+                        yield f"data: {json.dumps({'type': 'error', 'segment': i, 'segment_id': job['seg_order'][i], 'error_code': 'dub_timing_overflow', 'error': 'Speech exceeds its time slot. Shorten the translation or choose Strict Slot or Stretch Video before exporting.', 'overflow_s': round(overflow_s, 3)})}\n\n"
+                        return
                     else:
                         fit_status.append({"status": "fits"})
 
                 else:
-                    # strict_slot (legacy): preserve the previous atempo / trim /
-                    # off semantics so existing callers and back-compat tests
-                    # keep passing.
+                    # Strict Slot fits the complete speech to the original
+                    # slot. Explicit legacy trim/off choices remain available.
                     place_at = start
                     effective_end = end
                     slowed_rate = None
-                    if i + 1 < len(all_segment_wavs):
-                        next_start = all_segment_wavs[i + 1][0]
-                        gap = next_start - end
-                        if gap > GAP_OVERFLOW_BUFFER_S:
-                            effective_end = end + min(
-                                gap - GAP_OVERFLOW_BUFFER_S, GAP_OVERFLOW_MAX_S,
-                            )
                     slot_samples = int(max(0.0, (effective_end - start)) * sr)
                     if slot_fit != "off" and slot_samples > 0 and wl > slot_samples:
                         if slot_fit == "time_stretch":
                             ratio = wl / slot_samples
-                            capped_ratio = min(ratio, MAX_STRETCH_RATIO)
-                            capped_target = int(wl / capped_ratio)
                             try:
                                 adjusted = await _pitch_preserving_stretch(
-                                    adjusted, capped_target, sr,
+                                    adjusted, slot_samples, sr,
                                 )
-                                if adjusted.shape[-1] > slot_samples:
-                                    adjusted = adjusted[..., :slot_samples]
-                                if ratio > MAX_STRETCH_RATIO:
-                                    logger.info(
-                                        "seg %d compression %.2f× exceeded cap; "
-                                        "stretched to %.2f×, tail trimmed",
-                                        i, ratio, capped_ratio,
-                                    )
                             except Exception as e:
                                 logger.warning(
                                     "atempo stretch failed for seg %d (%.2f×), "
@@ -1891,15 +1853,14 @@ async def dub_generate(job_id: str, req: DubRequest):
                         slot_fit == "time_stretch"
                         and slot_samples > 0
                         and wl > 0
-                        and wl < slot_samples * UNDERRUN_TOLERANCE
-                        and _underrun_min_rate() < 1.0 - 1e-6
+                        and wl < slot_samples
                     ):
                         # Underrun fill (mirror of the compression above): the
                         # dub finished early, leaving the on-screen mouth moving
                         # over the thin under-speech bed residue — perceived as
                         # dead air. Slow toward the slot, never below the floor.
-                        rate = max(wl / slot_samples, _underrun_min_rate())
-                        target = min(slot_samples, int(round(wl / rate)))
+                        rate = wl / slot_samples
+                        target = slot_samples
                         try:
                             adjusted = await _pitch_preserving_stretch(
                                 adjusted, target, sr,
@@ -2043,12 +2004,9 @@ async def dub_generate(job_id: str, req: DubRequest):
                 "fit_fp": fit_fp,
             }
             job["dubbed_tracks"][lang_code]["fit_fp"] = fit_fp
-        # Record what kind of per-segment WAVs are on disk so a later
-        # smart_fit run knows whether partial regen / fit-only re-mix can
-        # reuse them ("natural") or must regen once ("slotted"). Per-track
-        # (P1.3) — each language renders under its own strategy; the flat
-        # field stays in lock-step for older readers.
-        _kind = "slotted" if strategy == "strict_slot" else "natural"
+        # Every new cache preserves natural speech. Old slotted caches must
+        # be regenerated once because their missing tails cannot be recovered.
+        _kind = "natural"
         job.setdefault("seg_wav_kind_by_lang", {})[lang_code] = _kind
         job["seg_wav_kind"] = _kind
         _save_job(job_id, job)
diff --git a/backend/core/tasks.py b/backend/core/tasks.py
index 44c2451c..00d483f1 100644
--- a/backend/core/tasks.py
+++ b/backend/core/tasks.py
@@ -10,6 +10,27 @@ from core import run_sentinel
 logger = logging.getLogger("omnivoice.tasks")
 
 
+def _stream_failure(update):
+    """Recognize terminal SSE failures, including generators that do not raise."""
+    if isinstance(update, bytes):
+        update = update.decode("utf-8", errors="replace")
+    if not isinstance(update, str):
+        return None
+    lines = update.splitlines()
+    try:
+        payload = json.loads("\n".join(line[5:].strip() for line in lines if line.startswith("data:")))
+    except (ValueError, TypeError):
+        return None
+    if not isinstance(payload, dict):
+        return None
+    if payload.get("type") != "error" and not any(line.strip() == "event: error" for line in lines):
+        return None
+    detail = payload.get("reason") or payload.get("error") or payload.get("detail")
+    if isinstance(detail, dict):
+        detail = detail.get("message") or detail.get("reason")
+    return detail if isinstance(detail, str) and detail else "Task failed"
+
+
 class TaskManager:
     """In-memory task dispatcher with SQLite-backed metadata.
 
@@ -125,9 +146,19 @@ class TaskManager:
                             except Exception: logger.exception("job_store.mark_cancelled failed")
                             break
                         await self._push_event(task_id, update)
+                        stream_error = _stream_failure(update)
+                        if stream_error is not None:
+                            t["status"] = "failed"
+                            t["error"] = stream_error
+                            try:
+                                job_store.mark_failed(task_id, stream_error)
+                            except Exception:
+                                logger.exception("job_store.mark_failed failed")
+                            await res.aclose()
+                            break
                 elif inspect.iscoroutine(res):
                     await res
-                if t["status"] != "cancelled":
+                if t["status"] not in {"cancelled", "failed"}:
                     t["status"] = "done"
                     try: job_store.mark_done(task_id)
                     except Exception: logger.exception("job_store.mark_done failed")
diff --git a/backend/schemas/requests.py b/backend/schemas/requests.py
index 20d09bcd..ec222648 100644
--- a/backend/schemas/requests.py
+++ b/backend/schemas/requests.py
@@ -89,9 +89,9 @@ class DubRequest(BaseModel):
     # (Bengali, Hindi, Arabic…). Three modes:
     #   "concise"       — never compress TTS audio. Trim text up-front via
     #                     speech_rate so it fits naturally; if it still
-    #                     overflows, hard-trim at slot with a short fade and
-    #                     surface fit_status="overflows" so the UI can prompt
-    #                     the user to shorten the segment. DEFAULT.
+    #                     overflows, fail without replacing the current track;
+    #                     the user must shorten it or choose another fit mode.
+    #                     DEFAULT.
     #   "stretch_video" — never compress TTS audio. Re-lay the timeline so
     #                     each segment's video portion is stretched (via
     #                     ffmpeg setpts) to fit the natural-rate dub audio.
@@ -100,13 +100,13 @@ class DubRequest(BaseModel):
     #                     mild pitch-preserving audio speed-up (≤1.2× alone,
     #                     ≤1.5× in hybrid) and a mild per-segment video
     #                     slow-down (≤2.0×), per services/fit_planner.py.
-    #                     Residual overflow is trimmed and surfaced.
-    #   "strict_slot"   — legacy: keep `slot_fit` semantics (atempo squeeze
-    #                     when audio > slot). Kept for back-compat.
+    #                     Residual overflow fails without discarding words.
+    #   "strict_slot"   — pitch-preserving fit of the complete speech to
+    #                     the original start/end; may sound faster or slower.
     timing_strategy: Optional[Literal["concise", "stretch_video", "strict_slot", "smart_fit"]] = "concise"
 
-    # Per-job slip budget for "concise" mode. Hard-trim only kicks in once
-    # gap absorption + this much extra time has been consumed.
+    # Per-job slip budget for "concise" mode. Overflow fails once gap
+    # absorption + this much extra time has been consumed.
     overflow_budget_s: Optional[float] = 0.0
 
     # Knob overrides for `smart_fit` (ignored by other strategies). Omitted
diff --git a/backend/services/audio_dsp.py b/backend/services/audio_dsp.py
index c3a521dd..08758823 100644
--- a/backend/services/audio_dsp.py
+++ b/backend/services/audio_dsp.py
@@ -186,6 +186,26 @@ def trim_trailing_silence(
     return audio_tensor[..., :end]
 
 
+
+def trim_speech_padding(audio_tensor: torch.Tensor, sample_rate: int) -> torch.Tensor:
+    """Remove generated edge silence before timing, retaining 50 ms of context.
+
+    Never compress silence into the spoken slot or delete internal pauses.
+    Silent/invalid outputs remain intact for the generation integrity guard.
+    """
+    if audio_tensor.numel() == 0 or sample_rate <= 0:
+        return audio_tensor
+    envelope = audio_tensor.abs()
+    if envelope.ndim > 1:
+        envelope = envelope.amax(dim=tuple(range(envelope.ndim - 1)))
+    voiced = torch.nonzero(envelope > 10 ** (-50 / 20))
+    if voiced.numel() == 0:
+        return audio_tensor
+    margin = int(sample_rate * 0.05)
+    start = max(0, int(voiced[0].item()) - margin)
+    end = min(audio_tensor.shape[-1], int(voiced[-1].item()) + 1 + margin)
+    return audio_tensor[..., start:end]
+
 def apply_effects_chain(audio_tensor, sample_rate: int, chain: list[dict]) -> torch.Tensor:
     """Apply a chain of named effects to an audio tensor.
 
diff --git a/backend/services/segmentation.py b/backend/services/segmentation.py
index 4169ec02..302e5815 100644
--- a/backend/services/segmentation.py
+++ b/backend/services/segmentation.py
@@ -491,6 +491,35 @@ def _apply_scene_cuts(segments: List[Segment], scene_cuts: Iterable[float]) -> L
             dur_total = remaining.duration
             if dur_total <= 0:
                 break
+            words = remaining.extra.get("words")
+            if isinstance(words, list) and words:
+                # A camera cut is not a word boundary. Character-proportional
+                # splitting assigned words seconds away from their real speech.
+                candidates = [
+                    i for i in range(1, len(words))
+                    if float(words[i - 1]["end"]) <= float(words[i]["start"])
+                    and abs(float(words[i]["start"]) - cut) <= 0.25
+                ]
+                if not candidates:
+                    continue
+                split = min(candidates, key=lambda i: abs(float(words[i]["start"]) - cut))
+                left, right = words[:split], words[split:]
+                left_text = _clean(" ".join(str(w.get("text", w.get("word", ""))) for w in left))
+                right_text = _clean(" ".join(str(w.get("text", w.get("word", ""))) for w in right))
+                left_end, right_start = float(left[-1]["end"]), float(right[0]["start"])
+                if (len(left_text) < MIN_CHARS or len(right_text) < MIN_CHARS
+                        or left_end - remaining.start < MIN_DUR
+                        or remaining.end - right_start < MIN_DUR):
+                    continue
+                out.append(Segment(
+                    start=remaining.start, end=left_end, text=left_text,
+                    speaker_id=remaining.speaker_id, extra={**remaining.extra, "words": left},
+                ))
+                remaining = Segment(
+                    start=right_start, end=remaining.end, text=right_text,
+                    speaker_id=remaining.speaker_id, extra={**remaining.extra, "words": right},
+                )
+                continue
             ratio = (cut - remaining.start) / dur_total
             tentative_split = int(len(remaining.text) * ratio)
             pos = _best_boundary(remaining.text, tentative_split)
diff --git a/docs/electron-dubbing.md b/docs/electron-dubbing.md
index 8d56d65d..64855bbc 100644
--- a/docs/electron-dubbing.md
+++ b/docs/electron-dubbing.md
@@ -199,3 +199,27 @@ Below 40rem of workspace width, Dubbing stacks its controls above the editor wit
 Global speed/quality changes preserve explicit Dubbing production steps, including settings restored from projects. Unset steps continue to follow the backend's current preset.
 
 NLLB resolves explicit FLORES language/script codes, unambiguous ISO-639-3 codes and common short aliases, including Traditional Chinese. Unsupported or script-ambiguous source, target or per-segment languages are rejected before model loading instead of silently translating to English.
+
+### Speech integrity and timing
+
+A failed, empty or unreadable speech segment stops generation before a new track
+replaces the previous output. Partial regeneration repairs missing or corrupt
+segment caches, including missing clips outside the requested changed-line list;
+it does not substitute silence and report completion. Auto speaker references
+exclude oversized clips when selecting a shared fallback, so short lines reuse a
+usable reference from their own speaker.
+
+New segment caches retain the full generated speech in every timing mode. Strict
+Slot removes edge silence and fits the complete clip to its original start/end
+with pitch-preserving speed adjustment. Very long or short translations can still
+sound unnaturally fast or slow; shorten or expand the translation for natural
+pacing. Legacy clipped caches require regeneration once, because fitting cannot
+recover discarded words. Explicit legacy Trim and Off options retain their
+respective clipping and overlap behavior.
+
+Concise and Smart Fit stop on unresolved overflow rather than publishing cut-off
+words. Shorten the translation, choose Strict Slot, or allow Stretch Video before
+retrying. Camera-cut segmentation uses nearby timed word boundaries when available
+and skips cuts inside speech that cannot be assigned safely. This improves phrase
+timing; it does not promise phoneme-level lip sync or correct inaccurate source
+transcripts automatically.
diff --git a/electron/src/renderer/src/components/pipeline-failure.test.tsx b/electron/src/renderer/src/components/pipeline-failure.test.tsx
index bc38a03a..490caef5 100644
--- a/electron/src/renderer/src/components/pipeline-failure.test.tsx
+++ b/electron/src/renderer/src/components/pipeline-failure.test.tsx
@@ -1,3 +1,4 @@
+import i18next from 'i18next';
 import { cleanup, render, screen } from '@testing-library/react';
 import { afterEach, expect, it, vi } from 'vitest';
 import { publicFailureFromEvent } from '@/lib/api/failure';
@@ -39,3 +40,21 @@ it('rejects a non-web documentation URL and uses the shared safe fallback', () =
     'https://github.com/debpalash/VoiceStudio/blob/main/docs/install/troubleshooting.md',
   );
 });
+
+it.each([
+  ['dub_speech_missing', 'dubIntegrity.missingSpeech'],
+  ['dub_timing_overflow', 'dubIntegrity.timingOverflow'],
+])('localizes %s while retaining diagnostics', (errorCode, key) => {
+  const translate = vi.spyOn(i18next, 't').mockReturnValue('Localized recovery instructions');
+  try {
+    const failure = publicFailureFromEvent(
+      { error_code: errorCode, reason: 'Raw engine error', diagnostic: 'segment a' },
+      'Fallback',
+    );
+    expect(failure.reason).toBe('Localized recovery instructions');
+    expect(failure.diagnostic).toBe('segment a');
+    expect(translate).toHaveBeenCalledWith(key);
+  } finally {
+    translate.mockRestore();
+  }
+});
diff --git a/electron/src/renderer/src/i18n/locales/ar.json b/electron/src/renderer/src/i18n/locales/ar.json
index 191b525a..15853765 100644
--- a/electron/src/renderer/src/i18n/locales/ar.json
+++ b/electron/src/renderer/src/i18n/locales/ar.json
@@ -2578,5 +2578,9 @@
     "captured": "تم التقاط {{count}} من أسطر سجل المشكلات",
     "contextNotice": "يتلقى الوكيل المحدد هذا البلاغ والشاشة الحالية وسجلات التطبيق والخلفية الحديثة وتشخيصات النظام.",
     "complete": "مكتمل"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "الكلام مفقود أو غير قابل للقراءة. أعد توليد المقطع المتأثر قبل التصدير.",
+    "timingOverflow": "يتجاوز الكلام الوقت المخصص له. اختصر الترجمة أو اختر التوقيت الصارم أو تمديد الفيديو."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/de.json b/electron/src/renderer/src/i18n/locales/de.json
index 74823a77..ea635079 100644
--- a/electron/src/renderer/src/i18n/locales/de.json
+++ b/electron/src/renderer/src/i18n/locales/de.json
@@ -2570,5 +2570,9 @@
     "captured": "{{count}} problematische Protokollzeilen erfasst",
     "contextNotice": "Der ausgewählte Agent erhält diesen Bericht, die aktuelle Ansicht, aktuelle App- und Backend-Protokolle sowie Systemdiagnosen.",
     "complete": "Abgeschlossen"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Sprache fehlt oder ist nicht lesbar. Erzeuge das betroffene Segment vor dem Export erneut.",
+    "timingOverflow": "Die Sprache überschreitet ihr Zeitfenster. Kürze die Übersetzung oder wähle ein festes Zeitfenster oder die Videostreckung."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/en.json b/electron/src/renderer/src/i18n/locales/en.json
index a20973dc..06fe3f88 100644
--- a/electron/src/renderer/src/i18n/locales/en.json
+++ b/electron/src/renderer/src/i18n/locales/en.json
@@ -2570,5 +2570,9 @@
     "captured": "Captured {{count}} problem log lines",
     "contextNotice": "The selected agent receives this report, the current screen, recent app and backend logs, and system diagnostics.",
     "complete": "Complete"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Speech is missing or unreadable. Regenerate the affected segment before exporting.",
+    "timingOverflow": "Speech exceeds its time slot. Shorten the translation or choose Strict Slot or Stretch Video."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/es.json b/electron/src/renderer/src/i18n/locales/es.json
index e36734c8..bfe9d7f6 100644
--- a/electron/src/renderer/src/i18n/locales/es.json
+++ b/electron/src/renderer/src/i18n/locales/es.json
@@ -2572,5 +2572,9 @@
     "captured": "Se capturaron {{count}} líneas de registro con problemas",
     "contextNotice": "El agente seleccionado recibe este informe, la pantalla actual, los registros recientes de la aplicación y del backend, y los diagnósticos del sistema.",
     "complete": "Completado"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Falta audio de voz o no se puede leer. Regenera el segmento afectado antes de exportar.",
+    "timingOverflow": "La voz supera su intervalo de tiempo. Acorta la traducción o elige un intervalo estricto o alargar el vídeo."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/fr.json b/electron/src/renderer/src/i18n/locales/fr.json
index a198a50c..ec820aae 100644
--- a/electron/src/renderer/src/i18n/locales/fr.json
+++ b/electron/src/renderer/src/i18n/locales/fr.json
@@ -2572,5 +2572,9 @@
     "captured": "{{count}} lignes de journal problématiques capturées",
     "contextNotice": "L’agent sélectionné reçoit ce rapport, l’écran actuel, les journaux récents de l’application et du backend, ainsi que les diagnostics système.",
     "complete": "Complet"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "La parole est absente ou illisible. Régénérez le segment concerné avant l’exportation.",
+    "timingOverflow": "La parole dépasse son créneau. Raccourcissez la traduction ou choisissez un créneau strict ou l’étirement vidéo."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/hi.json b/electron/src/renderer/src/i18n/locales/hi.json
index 0fd93029..984c1b66 100644
--- a/electron/src/renderer/src/i18n/locales/hi.json
+++ b/electron/src/renderer/src/i18n/locales/hi.json
@@ -2570,5 +2570,9 @@
     "captured": "समस्या वाली {{count}} लॉग पंक्तियाँ कैप्चर की गईं",
     "contextNotice": "चुने गए एजेंट को यह रिपोर्ट, वर्तमान स्क्रीन, हाल के ऐप और बैकएंड लॉग तथा सिस्टम निदान मिलते हैं।",
     "complete": "पूरा"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "बोली गायब है या पढ़ी नहीं जा सकती। निर्यात से पहले प्रभावित खंड दोबारा बनाएँ।",
+    "timingOverflow": "बोली अपने समय खंड से लंबी है। अनुवाद छोटा करें या सख्त समय खंड अथवा वीडियो खिंचाव चुनें।"
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/id.json b/electron/src/renderer/src/i18n/locales/id.json
index 93b7283e..d5139f01 100644
--- a/electron/src/renderer/src/i18n/locales/id.json
+++ b/electron/src/renderer/src/i18n/locales/id.json
@@ -2570,5 +2570,9 @@
     "captured": "{{count}} baris log bermasalah ditangkap",
     "contextNotice": "Agen yang dipilih menerima laporan ini, layar saat ini, log aplikasi dan backend terbaru, serta diagnostik sistem.",
     "complete": "Selesai"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Audio ucapan hilang atau tidak dapat dibaca. Buat ulang segmen terkait sebelum mengekspor.",
+    "timingOverflow": "Ucapan melebihi jatah waktunya. Persingkat terjemahan atau pilih slot ketat atau rentangkan video."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/it.json b/electron/src/renderer/src/i18n/locales/it.json
index 1e587a2b..e9745abc 100644
--- a/electron/src/renderer/src/i18n/locales/it.json
+++ b/electron/src/renderer/src/i18n/locales/it.json
@@ -2572,5 +2572,9 @@
     "captured": "Acquisite {{count}} righe di log problematiche",
     "contextNotice": "L’agente selezionato riceve questa segnalazione, la schermata corrente, i log recenti dell’app e del backend e la diagnostica di sistema.",
     "complete": "Completato"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Il parlato manca o non è leggibile. Rigenera il segmento interessato prima di esportare.",
+    "timingOverflow": "Il parlato supera il tempo disponibile. Accorcia la traduzione oppure scegli uno slot rigoroso o estendi il video."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/ja.json b/electron/src/renderer/src/i18n/locales/ja.json
index 49807960..dcfd745f 100644
--- a/electron/src/renderer/src/i18n/locales/ja.json
+++ b/electron/src/renderer/src/i18n/locales/ja.json
@@ -2570,5 +2570,9 @@
     "captured": "問題のあるログを{{count}}行取得しました",
     "contextNotice": "選択したエージェントには、この報告、現在の画面、最近のアプリとバックエンドのログ、システム診断が渡されます。",
     "complete": "完了"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "音声がないか読み取れません。書き出す前に該当セグメントを再生成してください。",
+    "timingOverflow": "音声が割り当て時間を超えています。翻訳を短くするか、厳密な時間枠または動画の引き伸ばしを選んでください。"
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/ko.json b/electron/src/renderer/src/i18n/locales/ko.json
index ac3ea22c..04fedcbe 100644
--- a/electron/src/renderer/src/i18n/locales/ko.json
+++ b/electron/src/renderer/src/i18n/locales/ko.json
@@ -2570,5 +2570,9 @@
     "captured": "문제 로그 {{count}}줄을 수집했습니다",
     "contextNotice": "선택한 에이전트는 이 보고서, 현재 화면, 최근 앱 및 백엔드 로그와 시스템 진단 정보를 받습니다.",
     "complete": "완료"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "음성이 없거나 읽을 수 없습니다. 내보내기 전에 해당 구간을 다시 생성하세요.",
+    "timingOverflow": "음성이 할당된 시간을 초과합니다. 번역을 줄이거나 엄격한 시간 구간 또는 비디오 늘이기를 선택하세요."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/nl.json b/electron/src/renderer/src/i18n/locales/nl.json
index b6cf7b64..be2026d0 100644
--- a/electron/src/renderer/src/i18n/locales/nl.json
+++ b/electron/src/renderer/src/i18n/locales/nl.json
@@ -2570,5 +2570,9 @@
     "captured": "{{count}} problematische logregels vastgelegd",
     "contextNotice": "De geselecteerde agent ontvangt dit rapport, het huidige scherm, recente app- en backendlogboeken en systeemdiagnostiek.",
     "complete": "Voltooid"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Spraak ontbreekt of is onleesbaar. Genereer het betreffende segment opnieuw voordat je exporteert.",
+    "timingOverflow": "De spraak overschrijdt het tijdvak. Verkort de vertaling of kies een strikt tijdvak of het uitrekken van de video."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/pl.json b/electron/src/renderer/src/i18n/locales/pl.json
index 0eac4bb7..8481f616 100644
--- a/electron/src/renderer/src/i18n/locales/pl.json
+++ b/electron/src/renderer/src/i18n/locales/pl.json
@@ -2574,5 +2574,9 @@
     "captured": "Przechwycono {{count}} problematycznych wierszy dziennika",
     "contextNotice": "Wybrany agent otrzyma to zgłoszenie, bieżący ekran, ostatnie dzienniki aplikacji i backendu oraz diagnostykę systemu.",
     "complete": "Zakończono"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Brak mowy lub nie można jej odczytać. Przed eksportem wygeneruj ponownie dany segment.",
+    "timingOverflow": "Mowa przekracza przydzielony czas. Skróć tłumaczenie albo wybierz ścisły przedział czasu lub wydłużenie filmu."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/pt.json b/electron/src/renderer/src/i18n/locales/pt.json
index 022a283d..d86b614a 100644
--- a/electron/src/renderer/src/i18n/locales/pt.json
+++ b/electron/src/renderer/src/i18n/locales/pt.json
@@ -2572,5 +2572,9 @@
     "captured": "{{count}} linhas de log com problemas capturadas",
     "contextNotice": "O agente selecionado recebe este relatório, a tela atual, os logs recentes do aplicativo e do backend e os diagnósticos do sistema.",
     "complete": "Completo"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "A fala está ausente ou ilegível. Gere novamente o segmento afetado antes de exportar.",
+    "timingOverflow": "A fala excede o intervalo de tempo. Encurte a tradução ou escolha um intervalo estrito ou estique o vídeo."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/ru.json b/electron/src/renderer/src/i18n/locales/ru.json
index 228a90b9..622eb45c 100644
--- a/electron/src/renderer/src/i18n/locales/ru.json
+++ b/electron/src/renderer/src/i18n/locales/ru.json
@@ -2574,5 +2574,9 @@
     "captured": "Собрано строк журнала с проблемами: {{count}}",
     "contextNotice": "Выбранный агент получит этот отчёт, текущий экран, последние журналы приложения и бэкенда, а также диагностику системы.",
     "complete": "Завершено"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Речь отсутствует или не читается. Перед экспортом заново сгенерируйте этот сегмент.",
+    "timingOverflow": "Речь выходит за отведённое время. Сократите перевод или выберите строгий интервал либо растяжение видео."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/sv.json b/electron/src/renderer/src/i18n/locales/sv.json
index c87b1f9b..1185b09c 100644
--- a/electron/src/renderer/src/i18n/locales/sv.json
+++ b/electron/src/renderer/src/i18n/locales/sv.json
@@ -2570,5 +2570,9 @@
     "captured": "{{count}} problemrader i loggen samlades in",
     "contextNotice": "Den valda agenten får den här rapporten, den aktuella vyn, aktuella app- och backendloggar samt systemdiagnostik.",
     "complete": "Klart"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Tal saknas eller kan inte läsas. Generera det berörda segmentet på nytt före export.",
+    "timingOverflow": "Talet överskrider sin tidslucka. Korta översättningen eller välj strikt tidslucka eller sträck ut videon."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/th.json b/electron/src/renderer/src/i18n/locales/th.json
index 095063ed..d50048df 100644
--- a/electron/src/renderer/src/i18n/locales/th.json
+++ b/electron/src/renderer/src/i18n/locales/th.json
@@ -2570,5 +2570,9 @@
     "captured": "บันทึกบรรทัดปัญหาจากล็อกแล้ว {{count}} บรรทัด",
     "contextNotice": "เอเจนต์ที่เลือกจะได้รับรายงานนี้ หน้าจอปัจจุบัน บันทึกล่าสุดของแอปและแบ็กเอนด์ และข้อมูลวินิจฉัยระบบ",
     "complete": "เสร็จสิ้น"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "เสียงพูดหายไปหรืออ่านไม่ได้ สร้างช่วงที่มีปัญหาใหม่ก่อนส่งออก",
+    "timingOverflow": "เสียงพูดยาวเกินช่วงเวลาที่กำหนด ย่อคำแปลหรือเลือกช่วงเวลาแบบเคร่งครัดหรือยืดวิดีโอ"
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/tr.json b/electron/src/renderer/src/i18n/locales/tr.json
index 0b3dae3a..cd495915 100644
--- a/electron/src/renderer/src/i18n/locales/tr.json
+++ b/electron/src/renderer/src/i18n/locales/tr.json
@@ -2570,5 +2570,9 @@
     "captured": "{{count}} sorunlu günlük satırı yakalandı",
     "contextNotice": "Seçilen aracı bu raporu, mevcut ekranı, son uygulama ve arka uç günlüklerini ve sistem tanılamasını alır.",
     "complete": "Tamamlandı"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Konuşma eksik veya okunamıyor. Dışa aktarmadan önce ilgili bölümü yeniden oluşturun.",
+    "timingOverflow": "Konuşma ayrılan süreyi aşıyor. Çeviriyi kısaltın veya kesin zaman aralığını ya da videoyu uzatmayı seçin."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/uk.json b/electron/src/renderer/src/i18n/locales/uk.json
index 885e5adf..84ad5c0a 100644
--- a/electron/src/renderer/src/i18n/locales/uk.json
+++ b/electron/src/renderer/src/i18n/locales/uk.json
@@ -2574,5 +2574,9 @@
     "captured": "Зібрано рядків журналу з проблемами: {{count}}",
     "contextNotice": "Вибраний агент отримає цей звіт, поточний екран, останні журнали застосунку й бекенду та діагностику системи.",
     "complete": "Завершено"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Мовлення відсутнє або не читається. Перед експортом повторно згенеруйте цей сегмент.",
+    "timingOverflow": "Мовлення перевищує відведений час. Скоротіть переклад або виберіть строгий інтервал чи розтягнення відео."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/vi.json b/electron/src/renderer/src/i18n/locales/vi.json
index 7e81528e..01cd0234 100644
--- a/electron/src/renderer/src/i18n/locales/vi.json
+++ b/electron/src/renderer/src/i18n/locales/vi.json
@@ -2570,5 +2570,9 @@
     "captured": "Đã thu thập {{count}} dòng nhật ký có vấn đề",
     "contextNotice": "Tác nhân đã chọn nhận báo cáo này, màn hình hiện tại, nhật ký ứng dụng và backend gần đây cùng thông tin chẩn đoán hệ thống.",
     "complete": "Hoàn thành"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Lời nói bị thiếu hoặc không đọc được. Hãy tạo lại đoạn bị ảnh hưởng trước khi xuất.",
+    "timingOverflow": "Lời nói vượt quá thời gian được phân bổ. Hãy rút ngắn bản dịch hoặc chọn khung thời gian nghiêm ngặt hay kéo dài video."
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/zh-CN.json b/electron/src/renderer/src/i18n/locales/zh-CN.json
index 59c6d20a..ae36b0c3 100644
--- a/electron/src/renderer/src/i18n/locales/zh-CN.json
+++ b/electron/src/renderer/src/i18n/locales/zh-CN.json
@@ -2574,5 +2574,9 @@
     "captured": "已捕获 {{count}} 行问题日志",
     "contextNotice": "所选代理将收到此报告、当前界面、最近的应用与后端日志以及系统诊断信息。",
     "complete": "已完成"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "语音缺失或无法读取。请在导出前重新生成受影响的片段。",
+    "timingOverflow": "语音超出分配的时长。请缩短译文，或选择严格时间段或拉伸视频。"
   }
 }
diff --git a/electron/src/renderer/src/i18n/locales/zh-TW.json b/electron/src/renderer/src/i18n/locales/zh-TW.json
index 161475fc..bdfedad7 100644
--- a/electron/src/renderer/src/i18n/locales/zh-TW.json
+++ b/electron/src/renderer/src/i18n/locales/zh-TW.json
@@ -2570,5 +2570,9 @@
     "captured": "已擷取 {{count}} 行問題日誌",
     "contextNotice": "所選代理將收到此報告、目前畫面、最近的應用程式與後端記錄以及系統診斷資訊。",
     "complete": "已完成"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "語音缺失或無法讀取。請在匯出前重新產生受影響的片段。",
+    "timingOverflow": "語音超出分配的時長。請縮短譯文，或選擇嚴格時間區段或延展影片。"
   }
 }
diff --git a/electron/src/renderer/src/lib/api/failure.ts b/electron/src/renderer/src/lib/api/failure.ts
index 81c95717..fba3fb5b 100644
--- a/electron/src/renderer/src/lib/api/failure.ts
+++ b/electron/src/renderer/src/lib/api/failure.ts
@@ -1,3 +1,4 @@
+import i18next from 'i18next';
 export interface PublicFailure {
   reason: string;
   errorClass?: string;
@@ -16,6 +17,8 @@ export function publicFailureFromEvent(
 ): PublicFailure {
   return {
     reason:
+      (event.error_code === 'dub_speech_missing' ? i18next.t('dubIntegrity.missingSpeech') :
+        event.error_code === 'dub_timing_overflow' ? i18next.t('dubIntegrity.timingOverflow') : undefined) ||
       text(event.reason) ||
       text(event.detail) ||
       text(event.error) ||
diff --git a/frontend/src/hooks/useDubWorkflow.js b/frontend/src/hooks/useDubWorkflow.js
index ce3473dc..a5f9c051 100644
--- a/frontend/src/hooks/useDubWorkflow.js
+++ b/frontend/src/hooks/useDubWorkflow.js
@@ -1185,6 +1185,7 @@ export default function useDubWorkflow({
         const decoder = new TextDecoder();
         let buffer = '';
         let wasCancelled = false;
+        let generationError = null;
         let sawDone = false;
         while (true) {
           const { done, value } = await reader.read();
@@ -1261,8 +1262,14 @@ export default function useDubWorkflow({
                   setDubStep('editing');
                   setDubError(t('dub_workflow.generation_aborted'));
                   toast(t('dub_workflow.dubbing_aborted'), { icon: '⏹' });
-                } else if (evt.type === 'error')
-                  setDubError((p) => p + `\nSeg ${evt.segment}: ${evt.error}`);
+                } else if (evt.type === 'error') {
+                  generationError = evt.error_code === 'dub_speech_missing'
+                    ? t('dubIntegrity.missingSpeech')
+                    : evt.error_code === 'dub_timing_overflow'
+                      ? t('dubIntegrity.timingOverflow')
+                      : evt.reason || evt.error || t('dub_workflow.generation_stream_ended');
+                  setDubError(generationError);
+                }
               } catch (err) {
                 console.warn('Dub generate SSE handler failed:', err);
               }
@@ -1271,7 +1278,7 @@ export default function useDubWorkflow({
         }
         setDubTaskId(null);
         if (!wasCancelled) {
-          if (!sawDone) throw new Error(t('dub_workflow.generation_stream_ended'));
+          if (!sawDone || generationError) throw new Error(generationError || t('dub_workflow.generation_stream_ended'));
           if (dubStep !== 'done') setDubStep('done');
           loadDubHistory();
           loadProjects();
diff --git a/frontend/src/i18n/locales/ar.json b/frontend/src/i18n/locales/ar.json
index 2fc402dc..c9ae2dbb 100644
--- a/frontend/src/i18n/locales/ar.json
+++ b/frontend/src/i18n/locales/ar.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "إيقاف مؤقت",
     "play": "تشغيل"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "الكلام مفقود أو غير قابل للقراءة. أعد توليد المقطع المتأثر قبل التصدير.",
+    "timingOverflow": "يتجاوز الكلام الوقت المخصص له. اختصر الترجمة أو اختر التوقيت الصارم أو تمديد الفيديو."
   }
 }
diff --git a/frontend/src/i18n/locales/de.json b/frontend/src/i18n/locales/de.json
index 80ae1834..df1e06be 100644
--- a/frontend/src/i18n/locales/de.json
+++ b/frontend/src/i18n/locales/de.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "Pausieren",
     "play": "Abspielen"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Sprache fehlt oder ist nicht lesbar. Erzeuge das betroffene Segment vor dem Export erneut.",
+    "timingOverflow": "Die Sprache überschreitet ihr Zeitfenster. Kürze die Übersetzung oder wähle ein festes Zeitfenster oder die Videostreckung."
   }
 }
diff --git a/frontend/src/i18n/locales/en.json b/frontend/src/i18n/locales/en.json
index b56e6f92..5c237c6d 100644
--- a/frontend/src/i18n/locales/en.json
+++ b/frontend/src/i18n/locales/en.json
@@ -3313,5 +3313,9 @@
     "transcript": "What it heard",
     "result_kicker": "Converted take",
     "need_source_and_voice": "Add a source clip and pick a target voice first."
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Speech is missing or unreadable. Regenerate the affected segment before exporting.",
+    "timingOverflow": "Speech exceeds its time slot. Shorten the translation or choose Strict Slot or Stretch Video."
   }
 }
diff --git a/frontend/src/i18n/locales/es.json b/frontend/src/i18n/locales/es.json
index 26365bf6..d675c0f5 100644
--- a/frontend/src/i18n/locales/es.json
+++ b/frontend/src/i18n/locales/es.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "Pausar",
     "play": "Reproducir"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Falta audio de voz o no se puede leer. Regenera el segmento afectado antes de exportar.",
+    "timingOverflow": "La voz supera su intervalo de tiempo. Acorta la traducción o elige un intervalo estricto o alargar el vídeo."
   }
 }
diff --git a/frontend/src/i18n/locales/fr.json b/frontend/src/i18n/locales/fr.json
index 5e02f132..3c377ab9 100644
--- a/frontend/src/i18n/locales/fr.json
+++ b/frontend/src/i18n/locales/fr.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "Mettre en pause",
     "play": "Lire"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "La parole est absente ou illisible. Régénérez le segment concerné avant l’exportation.",
+    "timingOverflow": "La parole dépasse son créneau. Raccourcissez la traduction ou choisissez un créneau strict ou l’étirement vidéo."
   }
 }
diff --git a/frontend/src/i18n/locales/hi.json b/frontend/src/i18n/locales/hi.json
index 1c5f1d03..5205fda7 100644
--- a/frontend/src/i18n/locales/hi.json
+++ b/frontend/src/i18n/locales/hi.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "रोकें",
     "play": "चलाएँ"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "बोली गायब है या पढ़ी नहीं जा सकती। निर्यात से पहले प्रभावित खंड दोबारा बनाएँ।",
+    "timingOverflow": "बोली अपने समय खंड से लंबी है। अनुवाद छोटा करें या सख्त समय खंड अथवा वीडियो खिंचाव चुनें।"
   }
 }
diff --git a/frontend/src/i18n/locales/id.json b/frontend/src/i18n/locales/id.json
index d3f61558..1c714c5f 100644
--- a/frontend/src/i18n/locales/id.json
+++ b/frontend/src/i18n/locales/id.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "Jeda",
     "play": "Putar"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Audio ucapan hilang atau tidak dapat dibaca. Buat ulang segmen terkait sebelum mengekspor.",
+    "timingOverflow": "Ucapan melebihi jatah waktunya. Persingkat terjemahan atau pilih slot ketat atau rentangkan video."
   }
 }
diff --git a/frontend/src/i18n/locales/it.json b/frontend/src/i18n/locales/it.json
index d446fa39..91c72cba 100644
--- a/frontend/src/i18n/locales/it.json
+++ b/frontend/src/i18n/locales/it.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "Pausa",
     "play": "Riproduci"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Il parlato manca o non è leggibile. Rigenera il segmento interessato prima di esportare.",
+    "timingOverflow": "Il parlato supera il tempo disponibile. Accorcia la traduzione oppure scegli uno slot rigoroso o estendi il video."
   }
 }
diff --git a/frontend/src/i18n/locales/ja.json b/frontend/src/i18n/locales/ja.json
index ae990e24..61950cdc 100644
--- a/frontend/src/i18n/locales/ja.json
+++ b/frontend/src/i18n/locales/ja.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "一時停止",
     "play": "再生"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "音声がないか読み取れません。書き出す前に該当セグメントを再生成してください。",
+    "timingOverflow": "音声が割り当て時間を超えています。翻訳を短くするか、厳密な時間枠または動画の引き伸ばしを選んでください。"
   }
 }
diff --git a/frontend/src/i18n/locales/ko.json b/frontend/src/i18n/locales/ko.json
index c0195df2..38bdcb43 100644
--- a/frontend/src/i18n/locales/ko.json
+++ b/frontend/src/i18n/locales/ko.json
@@ -3313,5 +3313,9 @@
     "stop": "재생 중지",
     "streaming_preview": "스트리밍 미리듣기…",
     "untitled": "오디오"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "음성이 없거나 읽을 수 없습니다. 내보내기 전에 해당 구간을 다시 생성하세요.",
+    "timingOverflow": "음성이 할당된 시간을 초과합니다. 번역을 줄이거나 엄격한 시간 구간 또는 비디오 늘이기를 선택하세요."
   }
 }
diff --git a/frontend/src/i18n/locales/nl.json b/frontend/src/i18n/locales/nl.json
index b8acf385..1b75e33a 100644
--- a/frontend/src/i18n/locales/nl.json
+++ b/frontend/src/i18n/locales/nl.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "Pauzeren",
     "play": "Afspelen"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Spraak ontbreekt of is onleesbaar. Genereer het betreffende segment opnieuw voordat je exporteert.",
+    "timingOverflow": "De spraak overschrijdt het tijdvak. Verkort de vertaling of kies een strikt tijdvak of het uitrekken van de video."
   }
 }
diff --git a/frontend/src/i18n/locales/pl.json b/frontend/src/i18n/locales/pl.json
index c2288863..85073878 100644
--- a/frontend/src/i18n/locales/pl.json
+++ b/frontend/src/i18n/locales/pl.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "Wstrzymaj",
     "play": "Odtwórz"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Brak mowy lub nie można jej odczytać. Przed eksportem wygeneruj ponownie dany segment.",
+    "timingOverflow": "Mowa przekracza przydzielony czas. Skróć tłumaczenie albo wybierz ścisły przedział czasu lub wydłużenie filmu."
   }
 }
diff --git a/frontend/src/i18n/locales/pt.json b/frontend/src/i18n/locales/pt.json
index 1744e940..dc4f6069 100644
--- a/frontend/src/i18n/locales/pt.json
+++ b/frontend/src/i18n/locales/pt.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "Pausar",
     "play": "Reproduzir"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "A fala está ausente ou ilegível. Gere novamente o segmento afetado antes de exportar.",
+    "timingOverflow": "A fala excede o intervalo de tempo. Encurte a tradução ou escolha um intervalo estrito ou estique o vídeo."
   }
 }
diff --git a/frontend/src/i18n/locales/ru.json b/frontend/src/i18n/locales/ru.json
index c3495e14..c1393731 100644
--- a/frontend/src/i18n/locales/ru.json
+++ b/frontend/src/i18n/locales/ru.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "Пауза",
     "play": "Воспроизвести"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Речь отсутствует или не читается. Перед экспортом заново сгенерируйте этот сегмент.",
+    "timingOverflow": "Речь выходит за отведённое время. Сократите перевод или выберите строгий интервал либо растяжение видео."
   }
 }
diff --git a/frontend/src/i18n/locales/sv.json b/frontend/src/i18n/locales/sv.json
index f62dbfdb..9df60e98 100644
--- a/frontend/src/i18n/locales/sv.json
+++ b/frontend/src/i18n/locales/sv.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "Pausa",
     "play": "Spela upp"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Tal saknas eller kan inte läsas. Generera det berörda segmentet på nytt före export.",
+    "timingOverflow": "Talet överskrider sin tidslucka. Korta översättningen eller välj strikt tidslucka eller sträck ut videon."
   }
 }
diff --git a/frontend/src/i18n/locales/th.json b/frontend/src/i18n/locales/th.json
index 4a184cef..5f5d50b3 100644
--- a/frontend/src/i18n/locales/th.json
+++ b/frontend/src/i18n/locales/th.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "หยุดชั่วคราว",
     "play": "เล่น"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "เสียงพูดหายไปหรืออ่านไม่ได้ สร้างช่วงที่มีปัญหาใหม่ก่อนส่งออก",
+    "timingOverflow": "เสียงพูดยาวเกินช่วงเวลาที่กำหนด ย่อคำแปลหรือเลือกช่วงเวลาแบบเคร่งครัดหรือยืดวิดีโอ"
   }
 }
diff --git a/frontend/src/i18n/locales/tr.json b/frontend/src/i18n/locales/tr.json
index 7a236377..f5047677 100644
--- a/frontend/src/i18n/locales/tr.json
+++ b/frontend/src/i18n/locales/tr.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "Duraklat",
     "play": "Oynat"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Konuşma eksik veya okunamıyor. Dışa aktarmadan önce ilgili bölümü yeniden oluşturun.",
+    "timingOverflow": "Konuşma ayrılan süreyi aşıyor. Çeviriyi kısaltın veya kesin zaman aralığını ya da videoyu uzatmayı seçin."
   }
 }
diff --git a/frontend/src/i18n/locales/uk.json b/frontend/src/i18n/locales/uk.json
index 898a588e..4cddb118 100644
--- a/frontend/src/i18n/locales/uk.json
+++ b/frontend/src/i18n/locales/uk.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "Пауза",
     "play": "Відтворити"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Мовлення відсутнє або не читається. Перед експортом повторно згенеруйте цей сегмент.",
+    "timingOverflow": "Мовлення перевищує відведений час. Скоротіть переклад або виберіть строгий інтервал чи розтягнення відео."
   }
 }
diff --git a/frontend/src/i18n/locales/vi.json b/frontend/src/i18n/locales/vi.json
index 7aafe894..7ed58a93 100644
--- a/frontend/src/i18n/locales/vi.json
+++ b/frontend/src/i18n/locales/vi.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "Tạm dừng",
     "play": "Phát"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "Lời nói bị thiếu hoặc không đọc được. Hãy tạo lại đoạn bị ảnh hưởng trước khi xuất.",
+    "timingOverflow": "Lời nói vượt quá thời gian được phân bổ. Hãy rút ngắn bản dịch hoặc chọn khung thời gian nghiêm ngặt hay kéo dài video."
   }
 }
diff --git a/frontend/src/i18n/locales/zh-CN.json b/frontend/src/i18n/locales/zh-CN.json
index b08224bd..38d74265 100644
--- a/frontend/src/i18n/locales/zh-CN.json
+++ b/frontend/src/i18n/locales/zh-CN.json
@@ -3313,5 +3313,9 @@
     "transcript": "识别到的内容",
     "result_kicker": "转换后的音频",
     "need_source_and_voice": "请先添加源音频并选择目标声音。"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "语音缺失或无法读取。请在导出前重新生成受影响的片段。",
+    "timingOverflow": "语音超出分配的时长。请缩短译文，或选择严格时间段或拉伸视频。"
   }
 }
diff --git a/frontend/src/i18n/locales/zh-TW.json b/frontend/src/i18n/locales/zh-TW.json
index 4833a48a..70e03b57 100644
--- a/frontend/src/i18n/locales/zh-TW.json
+++ b/frontend/src/i18n/locales/zh-TW.json
@@ -2832,5 +2832,9 @@
   "player": {
     "pause": "暫停",
     "play": "播放"
+  },
+  "dubIntegrity": {
+    "missingSpeech": "語音缺失或無法讀取。請在匯出前重新產生受影響的片段。",
+    "timingOverflow": "語音超出分配的時長。請縮短譯文，或選擇嚴格時間區段或延展影片。"
   }
 }
diff --git a/tests/test_dub_complete_audio.py b/tests/test_dub_complete_audio.py
new file mode 100644
index 00000000..49ad1955
--- /dev/null
+++ b/tests/test_dub_complete_audio.py
@@ -0,0 +1,142 @@
+"""A published dub must contain every requested spoken segment, without early clipping."""
+import asyncio
+import json
+from types import SimpleNamespace
+
+import pytest
+import torch
+import soundfile as sf
+
+from schemas.requests import DubRequest
+
+
+@pytest.fixture
+def render_dub(monkeypatch, tmp_path):
+    import api.routers.dub_generate as dg
+    job = {'duration': 4.0, 'dubbed_tracks': {}, 'segments': [], 'seg_wav_kind_by_lang': {'en': 'natural'}}
+    path = tmp_path / 'job'
+    path.mkdir()
+    events = []
+    generated = []
+    async def resolve():
+        return 'omnivoice', SimpleNamespace(target='local', remote=False), backend
+    def generate(**kwargs):
+        generated.append(kwargs)
+        return output[0]()
+    backend = SimpleNamespace(sample_rate=24000, generate=generate, applies_own_mastering=True)
+    output = [lambda: torch.ones(1, 24000) * .1]
+    class Tasks:
+        def is_cancelled(self, *_): return False
+        async def add_task(self, tid, kind, func, *args):
+            async for event in func(*args):
+                if event.startswith('data: '): events.append(json.loads(event[6:]))
+    monkeypatch.setattr(dg, '_resolve_dub_execution', resolve)
+    monkeypatch.setattr(dg, '_get_job', lambda _: job)
+    monkeypatch.setattr(dg, '_save_job', lambda *_: None)
+    monkeypatch.setattr(dg, 'DUB_DIR', str(tmp_path))
+    monkeypatch.setattr(dg, 'dub_seg_path', lambda _, sid: str(path / f'seg_{sid}.wav'))
+    monkeypatch.setattr(dg, 'task_manager', Tasks())
+    monkeypatch.setattr(dg, 'rvc_is_enabled', lambda: False)
+    monkeypatch.setattr(dg, 'mark_synthetic', lambda a, *args, **kw: a)
+    monkeypatch.setattr(dg, 'get_effect_chain', lambda _: None)
+    monkeypatch.setattr(dg, 'normalize_audio', lambda a, **kw: a)
+    def run(**kwargs):
+        body = dict(segments=[dict(start=0, end=1, text='hello')], segment_ids=['a'], language_code='en', num_step=4)
+        body.update(kwargs)
+        asyncio.run(dg.dub_generate('job', DubRequest(**body)))
+        return events
+    return SimpleNamespace(run=run, output=output, job=job, path=path, generated=generated)
+
+
+def test_failed_segment_does_not_publish_complete_track(render_dub):
+    def fail(): raise RuntimeError('engine failed')
+    render_dub.output[0] = fail
+    events = render_dub.run()
+    assert any(e['type'] == 'error' for e in events)
+    assert not any(e['type'] == 'done' for e in events)
+    assert not render_dub.job['dubbed_tracks']
+
+
+def test_missing_partial_cache_is_regenerated(render_dub):
+    events = render_dub.run(regen_only=[])
+    assert render_dub.generated
+    assert (render_dub.path / 'seg_en_a.wav').exists()
+    assert any(e['type'] == 'done' for e in events)
+
+
+@pytest.mark.parametrize("rvc_enabled", [False, True])
+def test_strict_slot_keeps_full_cache_and_fits_the_tail(render_dub, monkeypatch, rvc_enabled):
+    import api.routers.dub_generate as dg
+    render_dub.output[0] = lambda: torch.cat((torch.ones(1, 24000)*.1, torch.ones(1, 24000)*.3), dim=-1)
+    monkeypatch.setattr(dg, "rvc_is_enabled", lambda: rvc_enabled)
+    monkeypatch.setattr(dg, "apply_rvc", lambda _: None)
+    calls=[]
+    async def stretch(wav, target, sr):
+        calls.append((wav.shape[-1], target, float(wav[0,-1])))
+        return torch.nn.functional.interpolate(wav.unsqueeze(0),size=target,mode='linear').squeeze(0)
+    monkeypatch.setattr(dg, '_pitch_preserving_stretch', stretch)
+    events=render_dub.run(timing_strategy='strict_slot')
+    assert any(e['type']=='done' for e in events)
+    assert sf.info(render_dub.path/'seg_en_a.wav').duration == 2
+    assert calls and calls[0][0] == 48000 and calls[0][1] == 24000
+    assert calls[0][2] == pytest.approx(.3, abs=1e-4)
+
+
+def test_concise_overrun_does_not_publish_truncated_track(render_dub):
+    render_dub.output[0] = lambda: torch.ones(1, 48000)*.1
+    events=render_dub.run()
+    assert any(e['type']=='error' for e in events)
+    assert not any(e['type']=='done' for e in events)
+    assert not render_dub.job['dubbed_tracks']
+
+
+def test_silent_engine_output_is_not_a_successful_segment(render_dub):
+    render_dub.output[0] = lambda: torch.zeros(1, 24000)
+    events = render_dub.run()
+    assert any(e['type'] == 'error' for e in events)
+    assert not render_dub.job['dubbed_tracks']
+
+
+def test_camera_cut_uses_word_times_instead_of_character_ratio():
+    from services.segmentation import Segment, _apply_scene_cuts
+    words = [
+        {'text': 'A very long early phrase', 'start': 0, 'end': 1.8},
+        {'text': 'Short later phrase', 'start': 4.0, 'end': 6.0},
+    ]
+    segment = Segment(start=0, end=6, text='A very long early phrase Short later phrase', extra={'words': words})
+    pieces = _apply_scene_cuts([segment], [4.0])
+    assert len(pieces) == 2
+    assert pieces[0].end == 1.8
+    assert pieces[1].start == 4
+    assert pieces[0].extra['words'] == words[:1]
+    assert pieces[1].text == 'Short later phrase'
+
+
+def test_camera_cut_inside_a_word_does_not_reassign_speech():
+    from services.segmentation import Segment, _apply_scene_cuts
+    words = [
+        {'text': 'A lengthy opening phrase', 'start': 0, 'end': 4},
+        {'text': 'And the closing phrase', 'start': 4.5, 'end': 7},
+    ]
+    segment = Segment(start=0, end=7, text='A lengthy opening phrase And the closing phrase', extra={'words': words})
+    assert _apply_scene_cuts([segment], [3]) == [segment]
+
+
+def test_failed_regeneration_preserves_previous_track(render_dub):
+    previous = render_dub.path / 'dubbed_en.wav'
+    previous.write_bytes(b'previous successful output')
+    render_dub.job['dubbed_tracks']['en'] = {'path': str(previous)}
+    def fail(): raise RuntimeError('failure')
+    render_dub.output[0] = fail
+    render_dub.run()
+    assert previous.read_bytes() == b'previous successful output'
+    assert render_dub.job['dubbed_tracks']['en']['path'] == str(previous)
+
+
+def test_timing_trims_edge_silence_but_keeps_internal_pauses():
+    from services.audio_dsp import trim_speech_padding
+    wav = torch.cat((torch.zeros(1, 1000), torch.ones(1, 200)*.1,
+                     torch.zeros(1, 200), torch.ones(1, 200)*.1, torch.zeros(1,1000)), dim=-1)
+    result = trim_speech_padding(wav, 1000)
+    assert result.shape[-1] == 700
+    assert torch.equal(result[..., 250:450], torch.zeros(1,200))
diff --git a/tests/test_dub_voice_match.py b/tests/test_dub_voice_match.py
index dfd4258f..60580f2a 100644
--- a/tests/test_dub_voice_match.py
+++ b/tests/test_dub_voice_match.py
@@ -449,3 +449,13 @@ def test_consistent_explicit_cross_auto_seg_binding_is_honoured(patched_generate
     model = patched_generate(_HEURISTIC_JOB, body)
     assert model.refs[0] == ("/v/seg3.wav", "seg3 ref", False)  # explicit wins
     assert [r[0] for r in model.refs[1:]] == ["/v/seg1.wav"] * 3  # rest unified
+
+
+def test_consistent_pick_rejects_oversized_fallback_reference():
+    from api.routers.dub_generate import resolve_consistent_ref
+    job = _job(
+        [{'id': 'too-long', 'speaker_id': 'Speaker 1'}, {'id': 'usable', 'speaker_id': 'Speaker 1'}],
+        {'too-long': {'ref_audio': '/v/long.wav', 'ref_text': 'long', 'duration': 20.92},
+         'usable': {'ref_audio': '/v/usable.wav', 'ref_text': 'usable', 'duration': 9.0}},
+    )
+    assert resolve_consistent_ref(job, 'speaker_1')['ref_audio'] == '/v/usable.wav'
diff --git a/tests/test_redub_incremental.py b/tests/test_redub_incremental.py
index 59a0f03f..e9ac7abe 100644
--- a/tests/test_redub_incremental.py
+++ b/tests/test_redub_incremental.py
@@ -405,9 +405,9 @@ def test_legacy_cache_ignored_once_job_has_another_language(patched_generate):
 
     model.calls.clear()
     run(_body(_ES_SEGS, regen_only=[]))
-    assert model.calls == []  # not in the regen list → still no TTS…
-    # …but the ambiguous legacy audio was NOT spliced in: the slot is silence.
-    assert abs(_track_sample(job_dir, "es", 1.25)) < 0.005
+    # Missing unambiguous speech must be regenerated, never padded with silence.
+    assert model.calls == ["Buenos dias", "Hasta luego"]
+    assert abs(_track_sample(job_dir, "es", 1.25) - _amp_for("Hasta luego")) < 0.01
 
 
 def test_seg_hashes_stored_per_language_with_flat_mirror(patched_generate):
diff --git a/tests/test_task_stream_failure.py b/tests/test_task_stream_failure.py
new file mode 100644
index 00000000..64158754
--- /dev/null
+++ b/tests/test_task_stream_failure.py
@@ -0,0 +1,41 @@
+import asyncio
+import pytest
+from core.tasks import TaskManager, _stream_failure
+
+@pytest.mark.parametrize('event', [
+    'data: {"type":"error","error":"segment failed"}\n\n',
+    'event: error\ndata: {"detail":"segment failed"}\n\n',
+])
+def test_error_stream_is_terminal(event, monkeypatch):
+    from core import job_store
+    from core import run_sentinel
+    states=[]
+    for name in ['create','mark_running','append_event']:
+        monkeypatch.setattr(job_store,name,lambda *a,**kw: None)
+    monkeypatch.setattr(job_store,'mark_failed',lambda *a: states.append('failed'))
+    monkeypatch.setattr(job_store,'mark_done',lambda *a: states.append('done'))
+    monkeypatch.setattr(run_sentinel,'touch_activity',lambda *a: None)
+    closed=[]
+    async def stream():
+        try:
+            yield event
+            yield 'data: {"type":"done"}\n\n'
+        finally: closed.append(True)
+    async def run():
+        manager=TaskManager()
+        await manager.add_task('test','dub_generate',stream)
+        worker=asyncio.create_task(manager.worker())
+        try:
+            await asyncio.wait_for(manager.queue.join(),2)
+            assert manager.active_tasks['test']['status']=='failed'
+            assert len(manager.active_tasks['test']['history'])==1
+        finally:
+            worker.cancel()
+            try: await worker
+            except asyncio.CancelledError: pass
+    asyncio.run(run())
+    assert states==['failed']
+    assert closed==[True]
+
+def test_warnings_remain_non_terminal():
+    assert _stream_failure('data: {"type":"warning","error":"retrying"}\n\n') is None