{
  "id": 45987,
  "title": "blending noise",
  "url": "/competitions/tensorflow-speech-recognition-challenge/discussion/45987",
  "author_name": "",
  "post_date": "2017-12-19T03:39:11.569913200Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>how to blending noise to the train sample ? </p>",
  "messages": [
    {
      "id": "259841",
      "postDate": "12/19/2017 03:39:11",
      "content": "<p>how to blending noise to the train sample ? </p>",
      "rawMarkdown": "how to blending noise to the train sample ?",
      "votes": null
    },
    {
      "id": "260038",
      "postDate": "12/19/2017 13:19:09",
      "content": "<p>Take a look at those argumentation kernel</p>",
      "rawMarkdown": "Take a look at those argumentation kernel",
      "votes": null
    },
    {
      "id": "260125",
      "postDate": "12/19/2017 16:21:34",
      "content": "<p>You can blend in noise by simply adding a portion of the noise sample to the train sample.  For example (0.5 * noise_sample + 0.5 *train_sample).  The resulting sample is therefore an average of the two samples, or you can adjust the ratios for a different mix.</p>",
      "rawMarkdown": "You can blend in noise by simply adding a portion of the noise sample to the train sample.  For example (0.5 * noise_sample + 0.5 *train_sample).  The resulting sample is therefore an average of the two samples, or you can adjust the ratios for a different mix.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 260038,
      "author_name": "ryanzhang",
      "author_url": "",
      "post_date": "12/19/2017 13:19:09",
      "content": "<p>Take a look at those argumentation kernel</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 260125,
      "author_name": "robertkag",
      "author_url": "",
      "post_date": "12/19/2017 16:21:34",
      "content": "<p>You can blend in noise by simply adding a portion of the noise sample to the train sample.  For example (0.5 * noise_sample + 0.5 *train_sample).  The resulting sample is therefore an average of the two samples, or you can adjust the ratios for a different mix.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "259841": "how to blending noise to the train sample ?",
    "260038": "Take a look at those argumentation kernel",
    "260125": "You can blend in noise by simply adding a portion of the noise sample to the train sample.  For example (0.5 * noise_sample + 0.5 *train_sample).  The resulting sample is therefore an average of the two samples, or you can adjust the ratios for a different mix."
  },
  "source": "meta"
}