Spaces:
Sleeping
Sleeping
Upload 2 files
Browse files- config_mos.yaml +37 -0
- config_original.yaml +93 -0
config_mos.yaml
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Configuration for a standard Mean Opinion Score (MOS) test.
|
| 2 |
+
title: "MOS Test - Audio Quality Evaluation"
|
| 3 |
+
header_markdown: "Listen to the audio sample and rate its overall quality on a scale of 1 to 5."
|
| 4 |
+
|
| 5 |
+
instructions_markdown: |
|
| 6 |
+
**Welcome, Annotator!**
|
| 7 |
+
|
| 8 |
+
Instructions for MOS test:
|
| 9 |
+
|
| 10 |
+
Please follow these steps carefully:
|
| 11 |
+
1. Enter your unique **Annotator ID** before you begin.
|
| 12 |
+
2. Listen to each audio clip from start to finish.
|
| 13 |
+
3. Rate the clip using the sliders provided based on the scoring guide.
|
| 14 |
+
4. Provide any extra details in the comments box.
|
| 15 |
+
5. Click 'Save & Next' to submit your rating and load the next clip.
|
| 16 |
+
|
| 17 |
+
# The directory where your audio files are stored.
|
| 18 |
+
samples_directory: "sample-audios"
|
| 19 |
+
|
| 20 |
+
# Set to 'true' to shuffle the audio files, 'false' for alphabetical order.
|
| 21 |
+
randomize_samples: true
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
# MOS tests typically use a single criterion for overall quality.
|
| 25 |
+
criteria:
|
| 26 |
+
- label: "Overall Quality"
|
| 27 |
+
min: 1
|
| 28 |
+
max: 5
|
| 29 |
+
step: 1
|
| 30 |
+
default: 3
|
| 31 |
+
# These are standard definitions for the 5-point Absolute Category Rating (ACR) scale.
|
| 32 |
+
explanations:
|
| 33 |
+
1: "Bad - The quality is very distracting and unpleasant."
|
| 34 |
+
2: "Poor - The quality is distracting and annoying."
|
| 35 |
+
3: "Fair - The quality is slightly distracting, but acceptable."
|
| 36 |
+
4: "Good - The quality is not distracting, it is fine."
|
| 37 |
+
5: "Excellent - The quality is flawless and natural."
|
config_original.yaml
ADDED
|
@@ -0,0 +1,93 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# General UI Configuration
|
| 2 |
+
title: "TTS Rubric — Dynamic Evaluation"
|
| 3 |
+
|
| 4 |
+
instructions_markdown: |
|
| 5 |
+
**Welcome annotator!**
|
| 6 |
+
|
| 7 |
+
Instructions for multiple aspect test
|
| 8 |
+
|
| 9 |
+
Please follow these steps carefully:
|
| 10 |
+
1. Enter your unique **Annotator ID** before you begin.
|
| 11 |
+
2. Listen to each audio clip from start to finish.
|
| 12 |
+
3. Rate the clip using the sliders provided based on the scoring guide.
|
| 13 |
+
4. Provide any extra details in the comments box.
|
| 14 |
+
5. Click 'Save & Next' to submit your rating and load the next clip.
|
| 15 |
+
|
| 16 |
+
# The directory where your audio files are stored.
|
| 17 |
+
samples_directory: "sample-audios"
|
| 18 |
+
|
| 19 |
+
# Set to 'true' to shuffle the audio files, 'false' for alphabetical order.
|
| 20 |
+
randomize_samples: true
|
| 21 |
+
# Define the evaluation criteria. The UI will be built from this list.
|
| 22 |
+
criteria:
|
| 23 |
+
- label: "Clarity & Intelligibility"
|
| 24 |
+
min: 1
|
| 25 |
+
max: 5
|
| 26 |
+
step: 1
|
| 27 |
+
default: 3
|
| 28 |
+
explanations:
|
| 29 |
+
1: "Unacceptable."
|
| 30 |
+
2: "Often unclear or distorted; difficult to follow."
|
| 31 |
+
3: "Understandable but requires effort; some words unclear."
|
| 32 |
+
4: "Mostly clear, minor issues (with fast/slow playback)."
|
| 33 |
+
5: "Speech is clear, easy to understand (at all speeds)."
|
| 34 |
+
|
| 35 |
+
- label: "Accent & Pronunciation"
|
| 36 |
+
min: 1
|
| 37 |
+
max: 5
|
| 38 |
+
step: 1
|
| 39 |
+
default: 3
|
| 40 |
+
explanations:
|
| 41 |
+
1: "Severe pronunciation problems; largely unintelligible."
|
| 42 |
+
2: "Frequent pronunciation issues that impede understanding."
|
| 43 |
+
3: "Some mispronunciations that require effort to interpret."
|
| 44 |
+
4: "Minor pronunciation quirks but overall fine."
|
| 45 |
+
5: "Pronunciation is natural and appropriate for the target dialect."
|
| 46 |
+
|
| 47 |
+
- label: "Tone & Suitability"
|
| 48 |
+
min: 1
|
| 49 |
+
max: 5
|
| 50 |
+
step: 1
|
| 51 |
+
default: 3
|
| 52 |
+
explanations:
|
| 53 |
+
1: "Tone is inappropriate or harmful for the content."
|
| 54 |
+
2: "Tone often feels off or distracting from the content."
|
| 55 |
+
3: "Tone is acceptable but occasionally inappropriate."
|
| 56 |
+
4: "Generally appropriate tone with small mismatches."
|
| 57 |
+
5: "Tone fits the content and use-case perfectly."
|
| 58 |
+
|
| 59 |
+
- label: "Voice quality"
|
| 60 |
+
min: 1
|
| 61 |
+
max: 5
|
| 62 |
+
step: 1
|
| 63 |
+
default: 3
|
| 64 |
+
explanations:
|
| 65 |
+
1: "Unusable voice quality."
|
| 66 |
+
2: "Poor quality with frequent artifacts."
|
| 67 |
+
3: "Noticeable quality issues but still usable."
|
| 68 |
+
4: "Minor artifacts but overall high quality."
|
| 69 |
+
5: "Natural, pleasant voice with no artifacts."
|
| 70 |
+
|
| 71 |
+
- label: "Customization & Flexibility"
|
| 72 |
+
min: 1
|
| 73 |
+
max: 5
|
| 74 |
+
step: 1
|
| 75 |
+
default: 3
|
| 76 |
+
explanations:
|
| 77 |
+
1: "No useful customization; inflexible."
|
| 78 |
+
2: "Very limited or brittle customization options."
|
| 79 |
+
3: "Limited customization; acceptable for simple use-cases."
|
| 80 |
+
4: "Some customization available; works well for most cases."
|
| 81 |
+
5: "Highly flexible and customizable for different styles."
|
| 82 |
+
|
| 83 |
+
- label: "Listening comfort"
|
| 84 |
+
min: 1
|
| 85 |
+
max: 5
|
| 86 |
+
step: 1
|
| 87 |
+
default: 3
|
| 88 |
+
explanations:
|
| 89 |
+
1: "Uncomfortable or painful to listen to."
|
| 90 |
+
2: "Often fatiguing or distracting to listen to."
|
| 91 |
+
3: "Some listening fatigue; tolerable for short durations."
|
| 92 |
+
4: "Mostly comfortable with occasional sharpness or fatigue."
|
| 93 |
+
5: "Comfortable to listen to for extended periods."
|