Sound Event Detection Quiz

Answers and explanations appear as you take the quiz.

1. What extra output does temporal sound event detection estimate beyond a clip-level tag?
   • Only the audio file’s compression ratio
   • Only its filename’s alphabetical order
   • Event onset and offset boundaries within the clip
   • Only the display color of a waveform
2. Why is sound event detection often described as a polyphonic problem?
   • Because it only handles musical notes
   • Because multiple sounds can overlap and occur simultaneously
   • Because it requires multiple microphones
   • Because audio is always stereo
3. What does 'weak labeling' mean in SED training data?
   • The labels are encrypted
   • The audio quality is poor
   • Only quiet sounds are labeled
   • Clips are labeled that an event occurred but not precisely when
4. Which classifier structure is identified in the cited DCASE 2024 baseline?
   • A convolutional recurrent neural network with added pretrained frame features
   • Only a static spreadsheet sort
   • Only an unchanged relational database table
   • Only a mandatory image-file compressor
5. Which time-frequency representation is used in the documented sound-detection baseline context described here?
   • Only a speech transcript
   • A log-mel representation
   • Only a JPEG photograph
   • Only an unchanged output filename
6. Which cited research challenge supplies the 2018 and 2024 sound-event task settings discussed here?
   • The Olympic Games
   • A retail shopping holiday
   • A mandatory social-network ranking
   • DCASE
7. Considering only the DCASE 2018 onset collar of 200 ms, does a 1.15 s prediction satisfy it for a 1.00 s reference onset?
   • Yes; the onset error is 150 ms
   • No; the error is 1.15 s
   • No; every predicted onset must be numerically identical
   • Yes; onset errors are never evaluated
8. In the heterogeneous 2024 task, how does the baseline handle outputs for classes whose annotations are missing in a training dataset?
   • Treats every unannotated class as certainly absent
   • Copies all missing labels from the final test answers
   • Masks those unavailable outputs in the relevant loss computation
   • Eliminates every labeled class from the model