{
  "id": 578141,
  "title": "About torchaudio.transforms.MelSpectrogram() and librosa.feature.melspectrogram()",
  "url": "/competitions/birdclef-2025/discussion/578141",
  "author_name": "Dylan Liu",
  "post_date": "2025-05-09T03:31:05.303000",
  "votes": 1,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Which one are you using? <br>\nI got very different results from those 2 methods, but most important thing is, torchaudio.transforms.MelSpectrogram() is slower (Even with batchs, it takes ~40% more time to finish the same task).</p>",
  "messages": [
    {
      "id": 3198079,
      "postDate": "2025-05-09T03:31:05.303Z",
      "content": "<p>Which one are you using? <br>\nI got very different results from those 2 methods, but most important thing is, torchaudio.transforms.MelSpectrogram() is slower (Even with batchs, it takes ~40% more time to finish the same task).</p>",
      "rawMarkdown": "Which one are you using? \nI got very different results from those 2 methods, but most important thing is, torchaudio.transforms.MelSpectrogram() is slower (Even with batchs, it takes ~40% more time to finish the same task).",
      "votes": 1
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "3198079": "Which one are you using? \nI got very different results from those 2 methods, but most important thing is, torchaudio.transforms.MelSpectrogram() is slower (Even with batchs, it takes ~40% more time to finish the same task)."
  }
}