音樂生成模型聽測

這個研究想了解:人耳能不能分辨兩個 AI 音樂生成模型的差別。每一題你會看到一段英文的音樂描述,並聽到依照這段描述生成的兩段 30 秒音樂(A 與 B),然後回答三個二選一的比較題。

  • 建議使用耳機,在安靜的環境作答。開始前會先請你調整音量。
  • 至少作答 10 題(約 15–20 分鐘),之後可隨時結束;想多做也歡迎。
  • 每題答完就會存檔,中途關閉頁面不會遺失已作答的題目。

Music generation listening test

We want to know whether people can hear the difference between two AI music generators. Each question shows an English description of a piece of music and two 30-second clips (A and B) generated from it; you then answer three A-or-B comparison questions.

  • Headphones in a quiet place are recommended. You will set your volume before starting.
  • Answer at least 10 questions (about 15–20 minutes); after that you can stop at any time, or keep going.
  • Each answer is saved as soon as you submit it.