Sample
Prompt example (English translated from Japanese)
Compare the first and second audio clips across the following nine aspects.
Use the first audio as the reference point, and evaluate the impression difference of the second audio on a continuous 1–7 scale (decimals allowed).
If there is no difference, assign a score of 4.0.
For each aspect, output a single score based on your overall average impression.
After scoring all nine aspects, list the nine scores separated by commas as the “Final Score.”
Finally, describe the most notable difference or its reason in one or two concise sentences.
Output Format (must follow exactly):
[Audio1] vs [Audio2]
Pitch: x.x
Clarity: x.x
Calmness: x.x
Power: x.x
Age impression: x.x
Thickness: x.x
Tension: x.x
Brightness: x.x
Warmth: x.x
Final Score: x.xx, x.xx, x.xx, x.xx, x.xx, x.xx, x.xx, x.xx, x.xx
Reason: (Briefly describe the most noticeable difference or the reason for it in 1–2 sentences.)
Evaluation Scale (1 = left label, 7 = right label):
Pitch: High (1) – Low (7)
Clarity: Hoarse (1) – Clear (7)
Calmness: Calm (1) – Restless (7)
Power: Powerful (1) – Weak (7)
Age impression: Young (1) – Aged (7)
Thickness: Thick (1) – Thin (7)
Tension: Resonant (1) – Flat (7)
Brightness: Dark (1) – Bright (7)
Warmth: Cold (1) – Warm (7)
Example:
1st: 140.wav, 2nd: 140-1.wav
Pitch: 2.7
Clarity: 4.9
Calmness: 4.9
Power: 3.7
Age impression: 3.3
Thickness: 5.2
Tension: 2.9
Brightness: 5.8
Warmth: 3.5
Final Score: 2.7, 4.9, 4.9, 3.7, 3.3, 5.2, 2.9, 5.8, 3.5
Reason:
You will now be given pairs of audio files. Please evaluate each pair following the above instructions.