{
  "conventions": {
    "ctc": "merge adjacent repeats first, then delete <blank>; path product; target probability is path sum; loss is unnormalized natural logarithm in nat",
    "dft": "X[k]=sum_n x[n] exp(-j*2*pi*k*n/N); forward transform unnormalized, inverse has 1/N",
    "floating_point": "IEEE-754 binary64; near-zero DFT residue retained",
    "framing": "left-closed/right-open [mH,mH+F); complete frames only; no tail padding",
    "samples": "x[n]=sin(2*pi*f0*n/fs), n=0..15; seconds and hertz",
    "wer": "minimum token edit distance divided by six reference tokens; tie order diagonal, deletion, insertion"
  },
  "creator": "speech-technology-editor",
  "ctc": {
    "alphabet": [
      "A"
    ],
    "blank": "<blank>",
    "forward_rows": [
      [
        0.6,
        0.4,
        0.0
      ],
      [
        0.18,
        0.7,
        0.12
      ],
      [
        0.09,
        0.43999999999999995,
        0.41
      ]
    ],
    "length_normalized": false,
    "loss_negative_natural_log_nat": 0.16251892949777494,
    "target_probability_by_enumeration": 0.85,
    "target_probability_by_forward": 0.8499999999999999,
    "time_steps": [
      {
        "<blank>": 0.6,
        "A": 0.4
      },
      {
        "<blank>": 0.3,
        "A": 0.7
      },
      {
        "<blank>": 0.5,
        "A": 0.5
      }
    ],
    "valid_paths": [
      {
        "path": [
          "A",
          "<blank>",
          "<blank>"
        ],
        "probability": 0.06
      },
      {
        "path": [
          "<blank>",
          "A",
          "<blank>"
        ],
        "probability": 0.21
      },
      {
        "path": [
          "A",
          "A",
          "<blank>"
        ],
        "probability": 0.13999999999999999
      },
      {
        "path": [
          "<blank>",
          "<blank>",
          "A"
        ],
        "probability": 0.09
      },
      {
        "path": [
          "<blank>",
          "A",
          "A"
        ],
        "probability": 0.21
      },
      {
        "path": [
          "A",
          "A",
          "A"
        ],
        "probability": 0.13999999999999999
      }
    ]
  },
  "framing": {
    "complete_frame_count": 98,
    "frame_length_ms": 25,
    "frame_length_samples": 400,
    "hop_ms": 10,
    "hop_samples": 160,
    "last_frame_half_open": [
      15520,
      15920
    ],
    "padding": false,
    "sample_rate_hz": 16000,
    "tail_samples": 80,
    "total_samples": 16000
  },
  "license": "CC BY 4.0",
  "license_url": "https://creativecommons.org/licenses/by/4.0/",
  "provenance": "Original synthetic sine and fictional token sequences; no people, recording, external corpus or deployment data.",
  "sampling": {
    "dft": [
      {
        "frequency_hz": 0.0,
        "imag": 0.0,
        "k": 0,
        "magnitude": 3.3306690738754696e-16,
        "real": 3.3306690738754696e-16
      },
      {
        "frequency_hz": 500.0,
        "imag": 1.6653345369377348e-16,
        "k": 1,
        "magnitude": 2.0014830212433605e-16,
        "real": -1.1102230246251565e-16
      },
      {
        "frequency_hz": 1000.0,
        "imag": -8.000000000000002,
        "k": 2,
        "magnitude": 8.000000000000002,
        "real": -1.27675647831893e-15
      },
      {
        "frequency_hz": 1500.0,
        "imag": -1.5543122344752192e-15,
        "k": 3,
        "magnitude": 1.602148379482783e-15,
        "real": -3.885780586188048e-16
      },
      {
        "frequency_hz": 2000.0,
        "imag": -2.220446049250313e-16,
        "k": 4,
        "magnitude": 1.4782236529619926e-15,
        "real": 1.46145179927358e-15
      },
      {
        "frequency_hz": 2500.0,
        "imag": -1.9984014443252818e-15,
        "k": 5,
        "magnitude": 2.401779625492033e-15,
        "real": -1.3322676295501878e-15
      },
      {
        "frequency_hz": 3000.0,
        "imag": -1.887379141862766e-15,
        "k": 6,
        "magnitude": 2.8721321588324255e-15,
        "real": 2.1649348980190553e-15
      },
      {
        "frequency_hz": 3500.0,
        "imag": -5.551115123125783e-15,
        "k": 7,
        "magnitude": 5.978733960281817e-15,
        "real": -2.220446049250313e-15
      },
      {
        "frequency_hz": 4000.0,
        "imag": -2.106335683235129e-15,
        "k": 8,
        "magnitude": 2.1092595767504945e-15,
        "real": 1.1102230246251565e-16
      },
      {
        "frequency_hz": 4500.0,
        "imag": 5.218048215738236e-15,
        "k": 9,
        "magnitude": 6.9244459630691484e-15,
        "real": -4.551914400963142e-15
      },
      {
        "frequency_hz": 5000.0,
        "imag": -1.1102230246251565e-16,
        "k": 10,
        "magnitude": 3.554447978966673e-15,
        "real": 3.552713678800501e-15
      },
      {
        "frequency_hz": 5500.0,
        "imag": 4.9960036108132044e-15,
        "k": 11,
        "magnitude": 4.998778397814585e-15,
        "real": -1.6653345369377348e-16
      },
      {
        "frequency_hz": 6000.0,
        "imag": -4.9960036108132044e-15,
        "k": 12,
        "magnitude": 5.7568970043401595e-15,
        "real": 2.8603865192176604e-15
      },
      {
        "frequency_hz": 6500.0,
        "imag": -8.215650382226158e-15,
        "k": 13,
        "magnitude": 1.268038008882232e-14,
        "real": 9.658940314238862e-15
      },
      {
        "frequency_hz": 7000.0,
        "imag": 8.0,
        "k": 14,
        "magnitude": 8.0,
        "real": -2.0650148258027912e-14
      },
      {
        "frequency_hz": 7500.0,
        "imag": -1.0380585280245214e-14,
        "k": 15,
        "magnitude": 1.0401936599481387e-14,
        "real": -6.661338147750939e-16
      }
    ],
    "dft_bin_spacing_hz": 500.0,
    "nyquist_frequency_hz": 4000.0,
    "sample_count": 16,
    "sample_period_s": 0.000125,
    "sample_rate_hz": 8000,
    "samples": [
      0.0,
      0.7071067811865475,
      1.0,
      0.7071067811865476,
      1.2246467991473532e-16,
      -0.7071067811865471,
      -1.0,
      -0.7071067811865477,
      -2.4492935982947064e-16,
      0.7071067811865467,
      1.0,
      0.7071067811865483,
      3.6739403974420594e-16,
      -0.7071067811865466,
      -1.0,
      -0.7071067811865485
    ],
    "tone_frequency_hz": 1000
  },
  "schema": 1,
  "source": "AI Encyclopedia CN, Chapter 15",
  "wer": {
    "deletions": 1,
    "hypothesis_tokens": [
      "请",
      "请",
      "把",
      "灯",
      "打开",
      "离去"
    ],
    "insertions": 1,
    "reference_length": 6,
    "reference_tokens": [
      "请",
      "把",
      "灯",
      "打开",
      "然后",
      "离开"
    ],
    "substitutions": 1,
    "wer": 0.5
  }
}