'Video Indexer API returning no transcript instances

We use Azure Video-Indexer via a service to upload a dozen or so videos for indexing each week. Starting 2022-04-29 we noticed that some videos were returning an empty array in the JSON at

videos[n].insights.transcript[n].instances[]

(where "n" is any of the array objects)

{
   "videos": [{
      "insights": {
         "transcript": [{
            "instances": []
         ]}
      }
   }]
}

The array of instances under a transcript until that date was always populated. This has now happened for 8 videos of different lengths and formats.

Does anyone have any ideas what has happened?

video-indexer


Solution 1:[1]

thank you for reaching out. We are looking into this.

Azure Video Indexer team

Solution 2:[2]

Transcript insights should looks like:

"transcript": [ { "id": 1, "text": "some text", "confidence": 0.8875, "speakerId": 1, "language": "en-US", "instances": [ { "adjustedStart": "0:00:35.85", "adjustedEnd": "0:00:42.5", "start": "0:00:35.85", "end": "0:00:42.5" } ] }, { "id": 2, "text": "some text 2", "confidence": 0.8875, "speakerId": 2, "language": "en-US", "instances": [ { "adjustedStart": "0:00:42.51", "adjustedEnd": "0:00:44.46", "start": "0:00:42.51", "end": "0:00:44.46" } ] }, ... ]

  1. Do you mean that just "instances" is empty, so you do have "id", "text", "confidence", etc. but nothing under "instances"?
  2. Can you please share Account Id and some Video Ids which this occured to?
  3. Is the same media/video worked differently before April 29?

Thanks.

Sources

This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.

Source: Stack Overflow

Solution Source
Solution 1 Ika
Solution 2 Almog

Related Questions