Speaker Anti-Spoofing and ASVspoof Quiz

Answers and explanations appear as you take the quiz.

1. What is the goal of an anti-spoofing (countermeasure) system?
   • To decide whether audio is genuine or a spoof
   • To transcribe speech into text
   • To compress voice recordings
   • To translate speech between languages
2. Which of these is a 'logical access' spoofing attack in ASVspoof terminology?
   • Playing a recording through a loudspeaker
   • Text-to-speech synthesis of the target's voice
   • Holding a phone close to a microphone
   • Speaking in a noisy room
3. What does the tandem detection cost function (t-DCF) measure?
   • Only the spoofing detector in isolation
   • Audio file size after compression
   • The combined performance of the spoofing detector and the speaker-verification system
   • The number of speakers in a recording
4. What kind of clue do many anti-spoofing models rely on to catch synthetic speech?
   • The grammatical correctness of the words
   • The speaker's vocabulary size
   • The length of pauses between sentences
   • Subtle artifacts like unnatural phase or missing high-frequency detail
5. AASIST, a strong anti-spoofing model, is best described as which kind of system?
   • An end-to-end graph attention network over spectral and temporal sub-bands
   • A rule-based filter on transcripts
   • A simple energy threshold detector
   • A text translation model
6. Why is generalization to unseen attacks a growing concern in this field?
   • Because microphones are getting cheaper
   • Because new voice-cloning methods leave fewer detectable artifacts and differ from training data
   • Because users speak more slowly over time
   • Because audio files are getting smaller
7. In ASVspoof terminology, what is a physical-access attack?
   • Generating speech from text without a recording
   • Changing the font of a transcript
   • Replaying a recording of the target speaker into a microphone
   • Compressing a clean audio file for storage
8. What new type of evaluation track did ASVspoof 2021 add?
   • A video-captioning track
   • A music transcription track
   • A language identification track
   • A speech deepfake detection track