LibriSpeech - Audio Transcription and Quality Rating

# LibriSpeech - Audio Transcription and Quality Rating
# Based on Panayotov et al., ICASSP 2015
# Paper: https://ieeexplore.ieee.org/document/7178964
# Dataset: https://www.openslr.org/12
#
# LibriSpeech is a corpus of approximately 1000 hours of English
# read speech derived from public domain audiobooks. This task
# involves two annotation components:
#
# 1. Audio quality rating (slider): Rate the recording quality on
#    a scale from 0 (very poor) to 10 (excellent)
#
# 2. Audio content classification: Label what type of audio content
#    is present in the segment
#
# Quality considerations:
# - Background noise level
# - Speaker clarity and articulation
# - Recording volume and consistency
# - Presence of artifacts or distortion

annotation_task_name: "LibriSpeech: Audio Transcription and Quality Rating"
task_dir: "."

data_files:
  - sample-data.json

item_properties:
  id_key: "id"
  text_key: "text"

output_annotation_dir: "annotation_output/"
output_annotation_format: "json"

port: 8000
server_name: localhost

annotation_schemes:
  - annotation_type: slider
    name: audio_quality
    description: "Rate the overall audio quality (0 = very poor, 10 = excellent)"
    min_value: 0
    max_value: 10
    starting_value: 5

  - annotation_type: audio_annotation
    name: content_type
    description: "Classify the type of audio content in this segment"
    mode: "label"
    labels:
      - "Clean Speech"
      - "Noisy Speech"
      - "Music"
      - "Silence"
      - "Other"

annotation_instructions: |
  You will evaluate audio segments from the LibriSpeech corpus:
  1. Listen to the audio clip and read the provided transcript.
  2. Use the slider to rate the audio quality from 0 (very poor) to 10 (excellent).
  3. Classify the audio content type (clean speech, noisy speech, music, silence, or other).
  4. Consider background noise, speaker clarity, and recording quality in your rating.

html_layout: |
  <div style="padding: 15px; max-width: 800px; margin: auto;">
    <div style="background: #fef3c7; border: 1px solid #fde68a; border-radius: 8px; padding: 16px; margin-bottom: 16px;">
      <strong style="color: #a16207;">Audio Sample:</strong>
      <p style="margin: 8px 0 0 0;"><audio controls src="{{audio_url}}" style="width: 100%;"></audio></p>
    </div>
    <div style="background: #f0f9ff; border: 1px solid #bae6fd; border-radius: 8px; padding: 16px; margin-bottom: 16px;">
      <strong style="color: #0369a1;">Reference Transcript:</strong>
      <p style="font-size: 16px; line-height: 1.7; margin: 8px 0 0 0; font-style: italic;">{{text}}</p>
    </div>
  </div>

allow_all_users: true
instances_per_annotator: 100
annotation_per_instance: 3
allow_skip: true
skip_reason_required: false