jimbozhang / speechocean762

A dataset for pronunciation scoring tasks.
95 stars 20 forks source link

What's the meaning of * in pronounced-phone when having mispronunciations? #11

Closed tangYang7 closed 3 weeks ago

tangYang7 commented 3 weeks ago

Hello, First of all, thank you for developing this corpus.

I noticed that there are new labels in score.json and have some question since there are no directions to answer it. I'm wondering what is after a phone in pronounced-phone, such as T, AH, or R below the following example.

Thank you very much for your time and assistance!

Best regards.

"000480033": {
    "accuracy": 5,
    "completeness": 10,
    "fluency": 7,
    "prosodic": 7,
    "text": "ZERO FIVE EIGHT THREE",
    "total": 5,
    "words": [
...
      {
        "accuracy": 3,
        "mispronunciations": [
          {
            "canonical-phone": "TH",
            "index": 0,
            "pronounced-phone": "S"
          },
          {
            "canonical-phone": "R",
            "index": 1,
            "pronounced-phone": "R*"
          }
        ],
        "phones": [
          "TH",
          "R",
          "IY0"
        ],
        "phones-accuracy": [
          0.2,
          0.4,
          1.6
        ],
        "stress": 10,
        "text": "THREE",
        "total": 4
      }
    ]
  },
jimbozhang commented 3 weeks ago

Thank you for pointing out the misinformation. I have just updated the README.md and provided a more detailed introduction: https://github.com/jimbozhang/speechocean762?tab=readme-ov-file#the-mispronunciations-block