51黑料不打烊

DocumentationGenStudioUser Guide

Video features

Last update: Thu Feb 27 2025 00:00:00 GMT+0000 (Coordinated Universal Time)

Video features represent distinct and informative elements, sounds, or patterns within a video for analysis with Insights. These features help in categorizing and understanding the video content, enabling more accurate and detailed insights. By identifying various attributes such as audio mood, music genre, and objects, the AI can provide a comprehensive analysis of the video, aiding in better decision-making and strategy formulation.

Audio detection

Audio detection in GenStudio for Performance Marketing involves analyzing the audio track of a video to identify various attributes. This detection process determines the overall mood of the audio by identifying the types of audio present, tagging specific genres or categories of music, and extracting keywords from speech. By understanding these audio elements, the AI can provide deeper insights into the video鈥檚 content and context, enhancing the overall analysis and categorization process.

Search video features

To preview a video and hear a sample of the audio:

  1. In Insights, select the Attributes view.

  2. Change the table view by selecting Video.

  3. Select a feature from the Attribute category list. For an audio example, select Music genre.

  4. Select an attribute for a detailed view of videos that share that category.

    As an example, the Music genre category may have electronic as an attribute.

  5. The Attribute details page lists all the videos with this attribute. Hold your mouse pointer over any video in the list to start a video preview.

  6. Toggle the unmute (located in the lower-left corner of a video preview) button and listen to the audio of the video preview.

The following table lists the video feature categories recognized by the GenStudio for Performance Marketing AI. The detected attributes list for an asset is not exhaustive. Assets that contain a rich set of features may be limited to the three most dominant features identified by the AI.

Category
Description
Example
Audio Mood
Determines the overall emotional tone or atmosphere of the audio track, such as calm, upbeat, or tense.
Energetic, Happy, Emotional Ambient/atmospheric, Relaxing, Dramatic, Expressive/characterful, Intense, Slow, Neutral
Audio Types
Tags the video with one or more types of audio present, such as music or speech.
Music, Speech, Silence, Special effects, Ambience
Categories
Classifies the video into one or more broad content categories.
Entertainment, Sports, Music, Gaming, Howto tutorials, Fashion and style, Film and animation, Science and technology, Autos and vehicles, Pets and animals, People and blogs, News and politics, Social causes and activism, Travel and events, Education, Sales and offers
Music Category
Broad classification of music genre when music is present in the video. This helps in identifying the general type of music, such as contemporary or traditional styles.
Contemporary/pop music, Traditional/folk-based music, Instrumental/orchestral music, Rock music, Acoustic/unplugged music, Specialised/occasional music, Experimental/unique music
Music Genre
Specific classification of music style when music is present in the video, which provides a more detailed identification of the music, such as electronic or jazz.
electronic, hip-hop, dance, novelty, rock, world, reggae, pop, film, jazz, background, latin
Objects
Identifies one or more items, entities, and elements that appear in the video.
Values are too numerous, but some examples include: backpack, book, hawk, glasses, fish, pencil, mountain bike, soap
Orientation
The alignment of the video in relation to its width and height. Detects whether it is wider than it is tall (landscape), taller than it is wide (portrait), or equal in width and height (square).
landscape, portrait, square
People
When at least one person is present, one or more attributes may describe the person, or persons present in the video.
person, woman, man, girl, boy, social group, kid, crowd, people
Scenes
Identifies the setting or environment within a video, providing context about where the video was made or the type of location depicted.
Values are too numerous, but some examples include: lake, underwater, highway, hill, log cabin, island, beach, lounge
Styles
Detects visual treatments applied to video elements, such as those used in After Effects or Lightroom.
design, illustration, logo, square, cartoon, art, matte, neon
Tags
Detects other video characteristics that do not fall under a specific classification. Tags provide additional context and metadata about the video.
Values are too numerous, but some examples include: construction, gothic, healing, military, selfie, football, typing, dancer, dancing
recommendation-more-help
28475a92-7296-4969-af5f-82e50c3f1519