Introduction:

  • Controllable singing voice enables the prosody modification of existing singing voice based on a given modified musical score.
  • We provide several samples for each singing type, and each singing type is transformed into other types by changing the conditions of the score.
  • Each item contains Original Audio and Original Score. +x pitches denotes that the Original Score is increased or decreased by x as the conditional score, which is input into the system together with the Original Audio, and finally the modified audio is obtained.
  • In this experiment we use 6 pitches as the transition condition for adjacent types (pitch can be freely modified). At the same time, we provide +0 pitches for each item as a comparison for other transformations.


    • Audio Samples

    Alto

    1. 暖nuan 暖nuan 就jiou 在zai 胸xiong 膛tang 我uo 想xiang 说shuo 其qi 实shi 你ni 很hen 好hao

      2. Original Audio (voc.) Original Score To Alto (+0 pitches)
      wav
      To Soprano (+6 pitches) To Tenor (-6 pitches) To Bass (-12 pitches)
      wav

    2. 在zai 贝bei 加jia 尔er 湖hu 畔pan 多duo 少shao 年nian 以i 后hou

      3. Original Audio (voc.) Original Score To Alto (+0 pitches)
      wav
      To Soprano (+6 pitches) To Tenor (-6 pitches) To Bass (-12 pitches)
      wav

    Soprano

    1. 一i 座zuo 座zuo 山shan 川chuan

      2. Original Audio (voc.) Original Score To Soprano (+0 pitches)
      wav
      To Alto (-6 pitches) To Tenor (-12 pitches) To Bass (-18 pitches)
      wav

    2. 花hua 落luo 知zhi 多duo 少shao

      3. Original Audio (voc.) Original Score To Soprano (+0 pitches)
      wav
      To Alto (-6 pitches) To Tenor (-12 pitches) To Bass (-18 pitches)
      wav

    Tenor

    1. 只zhi 是shi 无u 处chu 停ting 摆bai 就jiou 歌ge 唱chang 吧ba

      2. Original Audio (voc.) Original Score To Tenor (+0 pitches)
      wav
      To Alto (+6 pitches) To Bass (-6 pitches) To Soprano (+12 pitches)
      wav

    2. 只zhi 是shi 那na 种zhong 温uen 柔rou 再zai 也ie 找zhao 不bu 到dao 拥iong 抱bao 的de 理li 由iou

      3. Original Audio (voc.) Original Score To Tenor (+0 pitches)
      wav
      To Alto (+6 pitches) To Bass (-6 pitches) To Soprano (+12 pitches)
      wav

    Bass

    1. 听ting 海hai 哭ku 的de 声sheng 音in

      2. Original Audio (voc.) Original Score To Bass (+0 pitches)
      wav
      To Tenor (+6 pitches) To Alto (+12 pitches) To Soprano (+18 pitches)
      wav

    2. 就jiou 当dang 最zuei 后hou 约ve 定ding

      3. Original Audio (voc.) Original Score To Bass (+0 pitches)
      wav
      To Tenor (+6 pitches) To Alto (+12 pitches) To Soprano (+18 pitches)
      wav