{
  "id": 317975,
  "title": "Pre-computed mel-specs with PCEN",
  "url": "/competitions/birdclef-2022/discussion/317975",
  "author_name": "Bilzard",
  "post_date": "2022-04-10T00:27:25.465000",
  "votes": 17,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I released two datasets of pre-computed mel-frequency spectrogram with per-channel energy normalization (PCEN).<br>\nAs you know, dynamically computing STFT and PCEN takes time. So this dataset will save time to compute these pre-processing.<br>\nI created two of that with different time resolutions.<br>\nFor more detail, see description of the datasets.</p>\n<ul>\n<li>hop_len=500 <a href=\"https://www.kaggle.com/datasets/tatamikenn/birdclef-2022-precomputed-melspec-hop-size500\" target=\"_blank\">https://www.kaggle.com/datasets/tatamikenn/birdclef-2022-precomputed-melspec-hop-size500</a></li>\n<li>hop_len=320: <a href=\"https://www.kaggle.com/datasets/tatamikenn/birdclef-2022-precomputed-melspec-hop-size320\" target=\"_blank\">https://www.kaggle.com/datasets/tatamikenn/birdclef-2022-precomputed-melspec-hop-size320</a></li>\n</ul>",
  "messages": [
    {
      "id": 1750663,
      "postDate": "2022-04-10T00:27:25.467Z",
      "content": "<p>I released two datasets of pre-computed mel-frequency spectrogram with per-channel energy normalization (PCEN).<br>\nAs you know, dynamically computing STFT and PCEN takes time. So this dataset will save time to compute these pre-processing.<br>\nI created two of that with different time resolutions.<br>\nFor more detail, see description of the datasets.</p>\n<ul>\n<li>hop_len=500 <a href=\"https://www.kaggle.com/datasets/tatamikenn/birdclef-2022-precomputed-melspec-hop-size500\" target=\"_blank\">https://www.kaggle.com/datasets/tatamikenn/birdclef-2022-precomputed-melspec-hop-size500</a></li>\n<li>hop_len=320: <a href=\"https://www.kaggle.com/datasets/tatamikenn/birdclef-2022-precomputed-melspec-hop-size320\" target=\"_blank\">https://www.kaggle.com/datasets/tatamikenn/birdclef-2022-precomputed-melspec-hop-size320</a></li>\n</ul>",
      "rawMarkdown": "I released two datasets of pre-computed mel-frequency spectrogram with per-channel energy normalization (PCEN).\nAs you know, dynamically computing STFT and PCEN takes time. So this dataset will save time to compute these pre-processing.\nI created two of that with different time resolutions.\nFor more detail, see description of the datasets.\n\n- hop_len=500 https://www.kaggle.com/datasets/tatamikenn/birdclef-2022-precomputed-melspec-hop-size500\n- hop_len=320: https://www.kaggle.com/datasets/tatamikenn/birdclef-2022-precomputed-melspec-hop-size320",
      "votes": 17
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1750663": "I released two datasets of pre-computed mel-frequency spectrogram with per-channel energy normalization (PCEN).\nAs you know, dynamically computing STFT and PCEN takes time. So this dataset will save time to compute these pre-processing.\nI created two of that with different time resolutions.\nFor more detail, see description of the datasets.\n\n- hop_len=500 https://www.kaggle.com/datasets/tatamikenn/birdclef-2022-precomputed-melspec-hop-size500\n- hop_len=320: https://www.kaggle.com/datasets/tatamikenn/birdclef-2022-precomputed-melspec-hop-size320"
  }
}