{
  "id": 18588141,
  "title": "How to remove silence from a recording in Python without cutting off words",
  "url": "https://urgent.news/2026/11/06/wie-man-in-python-stille-aus-einer-aufnahme-entfernt-ohne-worter",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-11-06T23:13:47.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/bestaitools-de/wie-man-in-python-stille-aus-einer-aufnahme-entfernt-ohne-worter-abzuschneiden-c3a"
  },
  "original_language": "de",
  "account": "An article compares the audio editing functions of the osFoundry desktop app with a custom-built numpy script and ffmpeg's silenceremove tool. The app's Sound Editor and the numpy script were used to remove silence from two test audio files, with the numpy script producing identical results to the app. The test files, recorded in English and Japanese, featured short text-to-speech sentences separated by 2.4-3.4 seconds of silence and background noise. The article examines the parameters used to determine silence, including threshold levels and minimum duration. The results show that a simple sample-by-sample comparison is sufficient to replicate the app's output.",
  "summary": "Removing pauses from an audio recording boils down to four numbers: a level below which audio is considered silent, how long a quiet section must be before it is removed, and how much to leave on either side of each remaining piece. The sound editor in the desktop app from osFoundry shows these four in the Strip Silence dialog, with defaults of −48…",
  "key_points": [
    "Sound Editor app displays these values with default settings of −48 dB, 50 ms, 5 ms, and 20 ms",
    "ffmpeg's silenceremove tool behaves differently at hard-to-notice spots, requiring careful handling"
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}