Ikibazo gihuza ibibazo · Hagati Urwego

Text-Based Speech Editing Quiz

Test transcript-aligned edits, synthetic replacement spans, acoustic blending, speaker consent, meaning review, and provenance.

Inzira zijyanye no kuyoboraText Based Speech Editing
Ikibazo 1 Bya 8

Ni ubuhe buryo bwo guhindura imvugo ishingiye ku nyandiko ihinduka muri dosiye y'amajwi?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Ni ubuhe buryo bwo guhindura imvugo ishingiye ku nyandiko ihinduka muri dosiye y'amajwi?
    • A selected spoken span to match an approved text edit
    • Only the waveform file extension
    • The speaker’s past memory of the event
    • Every unedited word by definition
  2. Why is a replacement word hard to blend into recorded narration?
    • Written spelling alone determines the waveform
    • The room has no effect on speech
    • A replacement never touches neighboring sound
    • Timing, timbre, prosody and room sound must fit its neighbors
  3. Which use requires explicit provenance to avoid misleading listeners?
    • An authorized correction of a narrator’s own pickup
    • Changing a quoted interview sentence and publishing it as if untouched
    • A visibly labeled synthetic replacement
    • A separately recorded new voiceover
  4. What should an archive preserve alongside an edited interview file?
    • Only the final synthetic version
    • Only the model’s output confidence
    • The original recording and an edit history under proper access rules
    • Only a polished transcript with no timing
  5. A replacement sounds natural. What does that not prove?
    • That the speaker consented to saying those new words
    • That a replacement can be played back
    • That the edit may be acoustically smooth
    • That the generated span can match a voice
  6. Why listen to the full sentence after an edit?
    • Only an isolated word can reveal errors
    • Sentence context is irrelevant to prosody
    • A model score proves all transitions are good
    • Neighboring words, rhythm and meaning may have changed
  7. Which technical step helps place the edit at the correct point?
    • Classifying the interview’s subject category
    • Aligning source transcript tokens with the original audio
    • Measuring the speaker’s nationality
    • Ranking unrelated sound effects
  8. What must be distinguished when evaluating an edited file?
    • Only the bit rate and sample rate
    • Only whether the voice sounds familiar
    • Lexical/acoustic quality and whether the edit was authorized and disclosed
    • Only transcript length