{
  "id": 373150,
  "title": "Video Length Issues",
  "url": "/competitions/nfl-player-contact-detection/discussion/373150",
  "author_name": "",
  "post_date": "2022-12-19T20:50:46.806939400Z",
  "votes": 2,
  "comment_count": 4,
  "views": 0,
  "content": "<p>I noticed when I load the videos using pyav/pytorchvideo, trim the first 5 seconds before the snap, and split them into 0.1 second chunks, I get a number of chunks that is slightly off from the number of timesteps each video is supposed to contains. Futhermore, it is off by a variable amount (around +-10 steps). Has anyone else experienced anything like this?</p>",
  "messages": [
    {
      "id": "2070322",
      "postDate": "12/19/2022 20:50:46",
      "content": "<p>I noticed when I load the videos using pyav/pytorchvideo, trim the first 5 seconds before the snap, and split them into 0.1 second chunks, I get a number of chunks that is slightly off from the number of timesteps each video is supposed to contains. Futhermore, it is off by a variable amount (around +-10 steps). Has anyone else experienced anything like this?</p>",
      "rawMarkdown": "I noticed when I load the videos using pyav/pytorchvideo, trim the first 5 seconds before the snap, and split them into 0.1 second chunks, I get a number of chunks that is slightly off from the number of timesteps each video is supposed to contains. Futhermore, it is off by a variable amount (around +-10 steps). Has anyone else experienced anything like this?",
      "votes": null
    },
    {
      "id": "2070331",
      "postDate": "12/19/2022 21:08:38",
      "content": "<p>You might be facing an issue with the frame-rate?<br>\nThere could be also some bugs in the code, without looking at it its hard to tell.</p>\n<p>I havent experienced issues with the matching, you could check manually one video and post it here for review.</p>",
      "rawMarkdown": "You might be facing an issue with the frame-rate?\nThere could be also some bugs in the code, without looking at it its hard to tell.\n\nI havent experienced issues with the matching, you could check manually one video and post it here for review.",
      "votes": null
    },
    {
      "id": "2070364",
      "postDate": "12/19/2022 23:03:33",
      "content": "<p>Okay, so I manually checked one example. 58422_001140_Sideline.mp4 is 14.948 seconds long. Trim the first 5 seconds, that's 9.94800 sec. Yet train_labels.csv shows 92 steps for 58422_001140.</p>",
      "rawMarkdown": "Okay, so I manually checked one example. 58422_001140_Sideline.mp4 is 14.948 seconds long. Trim the first 5 seconds, that's 9.94800 sec. Yet train_labels.csv shows 92 steps for 58422_001140.",
      "votes": null
    },
    {
      "id": "2070463",
      "postDate": "12/20/2022 03:02:03",
      "content": "<p>Andrew, the labels should be for the shorter of the sideline or endzone views. Since they are not necessarily identical in length you might need to check both.</p>\n<p>Hope that helps.</p>",
      "rawMarkdown": "Andrew, the labels should be for the shorter of the sideline or endzone views. Since they are not necessarily identical in length you might need to check both.\n\nHope that helps.",
      "votes": null
    },
    {
      "id": "2155723",
      "postDate": "02/22/2023 20:46:09",
      "content": "<p>got it, thanks</p>",
      "rawMarkdown": "got it, thanks",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2070331,
      "author_name": "carloshuertas",
      "author_url": "",
      "post_date": "12/19/2022 21:08:38",
      "content": "<p>You might be facing an issue with the frame-rate?<br>\nThere could be also some bugs in the code, without looking at it its hard to tell.</p>\n<p>I havent experienced issues with the matching, you could check manually one video and post it here for review.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2070364,
          "author_name": "andrewaf1",
          "author_url": "",
          "post_date": "12/19/2022 23:03:33",
          "content": "<p>Okay, so I manually checked one example. 58422_001140_Sideline.mp4 is 14.948 seconds long. Trim the first 5 seconds, that's 9.94800 sec. Yet train_labels.csv shows 92 steps for 58422_001140.</p>",
          "votes": null,
          "replies": [
            {
              "id": 2070463,
              "author_name": "robikscube",
              "author_url": "",
              "post_date": "12/20/2022 03:02:03",
              "content": "<p>Andrew, the labels should be for the shorter of the sideline or endzone views. Since they are not necessarily identical in length you might need to check both.</p>\n<p>Hope that helps.</p>",
              "votes": null,
              "replies": [
                {
                  "id": 2155723,
                  "author_name": "",
                  "author_url": "",
                  "post_date": "02/22/2023 20:46:09",
                  "content": "<p>got it, thanks</p>",
                  "votes": null,
                  "replies": []
                }
              ]
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2070322": "I noticed when I load the videos using pyav/pytorchvideo, trim the first 5 seconds before the snap, and split them into 0.1 second chunks, I get a number of chunks that is slightly off from the number of timesteps each video is supposed to contains. Futhermore, it is off by a variable amount (around +-10 steps). Has anyone else experienced anything like this?",
    "2070331": "You might be facing an issue with the frame-rate?\nThere could be also some bugs in the code, without looking at it its hard to tell.\n\nI havent experienced issues with the matching, you could check manually one video and post it here for review.",
    "2070364": "Okay, so I manually checked one example. 58422_001140_Sideline.mp4 is 14.948 seconds long. Trim the first 5 seconds, that's 9.94800 sec. Yet train_labels.csv shows 92 steps for 58422_001140.",
    "2070463": "Andrew, the labels should be for the shorter of the sideline or endzone views. Since they are not necessarily identical in length you might need to check both.\n\nHope that helps.",
    "2155723": "got it, thanks"
  },
  "source": "meta"
}