{
  "absUrl": "https://arxiv.org/abs/2609.24163",
  "arxivId": "2609.24163",
  "arxivVersion": null,
  "arxivVersionedId": null,
  "authors": [
    {
      "affiliations": [
        "机构信息未在 arXiv HTML 中可靠披露"
      ],
      "name": "Chee-En Yu"
    },
    {
      "affiliations": [
        "机构信息未在 arXiv HTML 中可靠披露"
      ],
      "name": "Yi-Cheng Lin"
    },
    {
      "affiliations": [
        "机构信息未在 arXiv HTML 中可靠披露"
      ],
      "name": "Sung-Feng Huang"
    },
    {
      "affiliations": [
        "机构信息未在 arXiv HTML 中可靠披露"
      ],
      "name": "Yun-Shao Tsai"
    },
    {
      "affiliations": [
        "机构信息未在 arXiv HTML 中可靠披露"
      ],
      "name": "Ho-Lam Chung"
    },
    {
      "affiliations": [
        "机构信息未在 arXiv HTML 中可靠披露"
      ],
      "name": "Xuanjun Chen"
    },
    {
      "affiliations": [
        "机构信息未在 arXiv HTML 中可靠披露"
      ],
      "name": "Hung-yi Lee"
    }
  ],
  "contract": "researcher-sidecars-v1",
  "id": "2609.24163",
  "originalTitle": "Listen, Critique, and Refine: RL-Based Self-Refinement for Instruction-Following Speech Synthesis",
  "pdfUrl": "https://arxiv.org/pdf/2609.24163.pdf",
  "schemaVersion": 1,
  "type": "preprint"
}