Cuestionario vinculado a una guía · Medio Nivel

Prueba de UX de audio conversacional

Prueba estados de interacción de audio, interrupción, discurso conciso, recuperación del usuario, detección de turnos, alternativas y evaluación realista.

Rutas de guía relacionadasExperiencia de usuario de audio conversacional
Pregunta 1 de 8

Why should an audio interface show whether it is listening, processing, or speaking?

Vista previa de cada pregunta8 preguntas

Las respuestas y explicaciones aparecen a medida que realiza el cuestionario.

  1. Why should an audio interface show whether it is listening, processing, or speaking?
    • State indicators improve speech recognition accuracy automatically
    • Users need to understand the current interaction state
    • Every system must speak continuously
    • It removes the need for error recovery
  2. A user starts correcting the assistant while it speaks. What should a good interaction support?
    • Continue playback and ignore the new input
    • Execute the earlier interpretation immediately
    • Restart the full conversation without context
    • Stop playback, preserve the correction, and avoid acting on the outdated interpretation
  3. Why put key information early in a spoken answer?
    • Speech has no timing constraints
    • Long answers guarantee comprehension
    • Listeners may have difficulty following long spoken lists or remembering detail
    • It improves the model’s training labels
  4. Which recovery options make an audio task more controllable?
    • Repeat, edit, cancel, or switch input modes
    • Hide the recognized text
    • Require users to restart after every error
    • Make the assistant choose whether to continue
  5. Before a voice system performs a consequential action, what design helps?
    • Use a pleasant voice to imply correctness
    • Skip confirmation when the input was spoken
    • Rely on the audio quality score
    • Show an editable interpretation and provide appropriate review
  6. What should an end-of-turn detector be tested against?
    • Only perfectly continuous speech
    • Natural pauses and the languages used by the audience
    • Only studio silence
    • Voice similarity across synthetic speakers
  7. Which test set better reflects real conversational audio use?
    • Cases with realistic noise, accents, accessibility needs, and network conditions
    • Only one speaker in a quiet room
    • Only scripted commands with no interruptions
    • Only recordings from the development team
  8. Why can silence without explanation be a UX problem?
    • It proves the system is processing accurately
    • Silence always improves accessibility
    • A user may interpret it as a system failure
    • It guarantees a shorter response