Bitzu

AI-901 ยท Information extraction with Content Understanding

How do audio and video analyzers differ from image analyzers?

Transcripts with speakers, summaries of calls, and video segmentation with per-segment descriptions and fields, and how audio and video analyzers differ from image ones.

Key points

  • Audio and video analyzers first transcribe speech (WebVTT, with speakers told apart through diarization), then use a generative model to fill fields such as a summary or sentiment.
  • Video also detects shots and key frames and can split the file into segments, each with its own description and fields (enableSegment).
  • Image analyzers work on a single picture and have no transcript or timeline.

Source: https://learn.microsoft.com/en-us/azure/ai-services/content-understanding/video/overview

Practice questions

Single choice

In what format do Content Understanding audio analyzers return transcripts?

  1. WebVTT, separated by speaker
  2. A WAV file of the cleaned audio
  3. A single unlabeled paragraph
  4. A CSV file of word frequencies
Show answer

Answer: A. WebVTT, separated by speaker

Transcripts are produced in WebVTT, separated by speaker. SRT-like or plain-paragraph output is not what the service documents.

Select all that apply

Which outputs can a video analyzer provide that an image analyzer cannot?

  1. A transcript of what is spoken
  2. Key frames and shot boundaries
  3. Descriptions and fields per segment of the timeline
  4. A field schema with named fields
  5. Markdown or JSON output
Show answer

Answer: A. A transcript of what is spoken, B. Key frames and shot boundaries, C. Descriptions and fields per segment of the timeline

Video analysis yields a transcript, key frames and shot boundaries, and per-segment descriptions across time. A single image has none of these.

True or false

True or false: the video analyzer transcribes music and sound effects as well as speech.

Show answer

Answer: False

False. Only spoken words are transcribed; music, sound effects and ambient noise are ignored.

Practice 8 questions on this topic in Bitzu

Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.

Download on the App StoreGet it on Google Play

More AI-901 topics in this module