HD Vertical AI-Generated Video Dataset for AI Training
Access 147,007 AI-generated clips delivered exclusively in vertical Full HD 1080 × 1920. The approximately 220.3-hour corpus is built for mobile video generation, portrait framing, short-form creative outputs, vertical retrieval, and evaluation of models that must preserve subjects and action inside a 9:16 composition.
Metadata can distinguish portrait-native generation from adapted or reframed output where known, while separately recording the generation workflow, provenance, synthetic-content flag, generation date, model or production source where permitted, prompt availability, reference-input status, human review, and known limitations.
Dataset Preview
Representative 9:16 previews showing portrait framing, mobile composition, short-form pacing, subject placement, edge behavior, and synthetic visual artifacts.
Preview assets shown here represent content available within the dataset.
Key Highlights
- • 147,007 synthetic clips delivered only in vertical 1080 × 1920
- • Approximately 220.3 hours with an average clip length of approximately 5.4 seconds
- • Portrait 9:16 geometry for mobile, social, and short-form model workflows
- • Portrait-native versus adapted or reframed status can be recorded where known
- • Explicit synthetic-content flag and vertical-format validation
- • Generation workflow, date, provenance, and permitted source information can be included
- • Prompt and reference-input availability recorded as asset-level statuses
- • Human-review and limitation fields can capture framing, edge, text, anatomy, and temporal issues
Metadata Fields
Example Metadata Record
Illustrative delivery record. Portrait-origin and generation fields are populated only when supported by retained records.
clip_id: aivert_000001
title: AI-generated vertical lifestyle sequence
description: Portrait-format synthetic scene with a centered subject and upward camera movement
keywords: AI-generated video, synthetic video, vertical, 9:16, portrait framing, mobile, short-form
resolution: 1080x1920
duration: 00:00:05
fps: Available in source metadata
orientation: Vertical
vertical_source_status: portrait-native / reframed / cropped / extended / unknown
generation_workflow: Recorded where retained
provenance: Source and portrait-processing history supplied to the available level
synthetic_content: true
generation_date: Available where retained
model_or_production_source: Disclosed where known and permitted
prompt_availability: full / partial / summary / unavailable / restricted
reference_input_status: none / image / video / unavailable / restricted
human_review: Vertical composition and visual QA status where recorded
known_limitations: Asset-level portrait-framing and temporal notes where identified
Technical Specifications
Dataset Type
Portrait Synthetic Video: HD Vertical
Primary Technical Intent
Mobile generation, 9:16 composition, short-form outputs, and vertical-model evaluation
Clip Count
147,007 clips
Total Duration
220.3 hours
Total Duration in Minutes
Approximately 13,217.4 minutes
Average Clip Length
Approximately 5.4 seconds
Resolution
1080 × 1920
Format
Full HD
Codec
H.264
Bitrate
Approximately 22 Mbps
Container
MP4
Orientation
Vertical
Aspect Ratio
9:16
Portrait-Origin Status
Native vertical, reframed, cropped, extended, or unknown where this can be established
Synthetic-Content Label
Explicit asset-level flag
Generation Documentation
Workflow, date, provenance, and source fields supplied where retained and permitted
Prompt & Reference Inputs
Availability and restriction status recorded separately
Vertical QA
Human-review state and known framing, edge, text, anatomy, or temporal limitations supplied where recorded
Licensing, Documentation, and Delivery
Wavebreak Media can provide applicable licensing, generation provenance, and portrait-source disclosure status information for the vertical AI-generated video content, with dataset delivery organized around agreed file, metadata, and technical requirements. See our Dataset Licensing & Compliance page for more detail.
Need the Full Dataset?
Request the vertical synthetic video corpus with the required 9:16 selection, portrait-origin status, generation metadata, prompt and reference availability, human-review state, and limitation reporting.
