Avatar Studio

Avatar Video Upload Guidelines

How to record and upload source video for a high-quality Voxforce AI avatar — framing, lighting, audio, consent, and quality checks for customers and partners.

Section 1

Purpose

To create a high-quality AI avatar, the uploaded video should clearly capture the person’s face, voice, expressions, and speaking style. A clean recording helps generate a more realistic, professional, and human-like avatar.

Section 2

Video duration

2–5 min
Recommended length
90 sec
Minimum

The person should speak continuously and naturally during the recording.

Recording flow (during capture)

Use this in-app flow when prompted; overall upload should still meet the recommended duration above.

Section 3

Camera setup

Required

  • Record in landscape (horizontal) mode only — rotate your phone sideways; portrait/vertical video is not supported.
  • Use a 16:9 frame; head and shoulders centered in the wide shot.
  • Camera at eye level.
  • Face clearly visible; look directly into the camera.
  • Head and shoulders in frame — do not crop forehead, chin, or shoulders.
Landscape recording required: horizontal phone with 16:9 head-and-shoulders framing is correct; vertical portrait mode is not supported. Rotate phone sideways with camera at eye level.

Example framing

Good and poor framing before you record.

Good Good framing: centered head and shoulders, even lighting

Front-facing with even lighting, clean background, hair behind shoulders, and no accessories.

Bad Bad framing: subject too far from camera

Too far from the camera. Frame yourself closer for best results.

Bad Bad framing: subject too close, head cropped

Too close to the camera. Head and shoulders should be fully visible.

Recommended framing

  • Head and upper chest visible; leave space above the head.
  • Person centered; avoid extreme close-ups.

Section 4

Lighting

Good lighting keeps your face clear and dimensional on a landscape recording. The standard approach is three-point lighting — three sources that work together so the camera sees an even, natural face.

Three-point lighting diagram: key light at 45 degrees front-left, fill light at 45 degrees front-right, back rim light behind subject, camera in front. Warning to avoid a backlit window.

1. Key light

The main source — brightest, placed front-left or front-right at roughly 45° and slightly above eye level. A window, ring light, or desk lamp with a diffuser works.

2. Fill light

A softer, dimmer source on the opposite side. It lifts shadows the key creates (under cheek, nose). A second lamp, reflector, or light wall can be your fill — it should never overpower the key.

3. Back / rim light

Optional but helpful — a small light behind you, aimed at shoulders/hair to separate you from the background. Keep it subtle; avoid shining into the lens.

No studio? Approximate it at your desk

  1. Key — ring light or window in front of you, off to one side, slightly above eye level.
  2. Fill — a second lamp or white wall on the opposite side to soften shadows.
  3. Back — skip if the background is plain; otherwise a small lamp behind you, out of frame.
  4. Check the preview — both cheeks and forehead similarly bright; no silhouette, no ceiling-only shadows.

Do

  • Key + fill on the face (three-point or a simple ring-light setup)
  • Soft, even light — no harsh shadows on cheeks or under eyes
  • Landscape frame with stable exposure throughout

Avoid

  • Backlight from windows
  • Dark rooms
  • Harsh overhead light
  • Colored or flickering lights

Section 5

Background

Background setup: good example with plain neutral wall and centered subject; bad example with cluttered room, moving people, patterns, mirror, and screen distractions.

Recommended

  • Plain wall or office setup
  • Clean, professional, light neutral background

Avoid

  • Busy or cluttered rooms
  • Moving people, strong patterns, mirrors
  • Screens playing video
  • Logos you do not own rights to

Section 6

Audio

Audio quality is as important as video quality.

Audio recording environment: good quiet room with clear voice and microphone; bad room with traffic noise, fan or AC hum, echo, and interruptions.

Do

  • Record in a quiet room
  • Use a good microphone if available
  • Keep the mic close but not visible if possible
  • Speak clearly, naturally, at consistent volume

Avoid

  • Background noise, echo, fan/AC noise
  • Music or traffic
  • Interruptions or speaking too fast

Section 7

Appearance

Recommended

  • Professional or brand-appropriate clothing
  • Solid colors; hair away from face
  • Avoid reflective glasses if possible
  • Natural facial expression

Avoid

  • Heavy patterns, stripes, shiny accessories
  • Face coverings, hats, sunglasses
  • Large distracting jewelry

Section 8

Speaking

Speak naturally, confidently, and clearly. Example lines:

“Hi, I’m here to help you understand our services and answer your questions.”

“You can ask me about pricing, features, support, or how our product works.”

Do

  • Normal speed; natural pauses
  • Friendly tone; slight expressions
  • Smile naturally; eye contact with camera

Avoid

  • Reading too fast; looking away
  • Long silent pauses; overacting
  • Whispering or shouting

Section 9

Body movement

Recommended

  • Sit or stand still; minimal natural head movement
  • Slight hand movement okay if visible
  • Stable posture

Avoid

  • Walking, excessive movement, leaving frame
  • Turning away; covering mouth; sudden gestures

Section 10

Technical requirements

  • File type: MP4
  • Resolution: 1080p minimum
  • Frame rate: 25fps or 30fps
  • Audio: clear voice, no background music
  • Optimize file size for upload

Avoid

  • Low resolution, blur, screen recording, heavy compression
  • Filters, beauty effects, watermarks, AI face filters

Section 12

Quality checklist before upload

Pre-upload verification

Image is sharp and well-lit
Subject appears 18 or older
Only one person in the video
Person centered in frame
Head fully visible
Looking straight at the camera
Pose natural and upright
Face clearly visible
Neck clearly visible
Hair behind shoulders
No accessories
Not a celebrity or copyrighted character
  • Audio clear; no background noise
  • At least 90 seconds of speech
  • No copyrighted music or third-party visuals
  • Consent confirmed

Section 13

Ideal recording setup

  • Quiet room · plain background · front-facing soft light
  • Camera at eye level · 1080p · external microphone
  • Person centered · natural speaking tone

Section 14

Sample recording scripts

Option 1 — General business

Hi, I’m your AI assistant. I’m here to help you understand our services, answer your questions, and guide you to the right solution.

You can ask me about our products, pricing, support, or how our company can help your business.

I’ll do my best to provide clear, helpful, and accurate information in a friendly and professional way.

Option 2 — Sales

Hi, I’m here to help you explore the right solution for your business.

You can ask me about features, pricing, use cases, or how our product works.

I can also help qualify your needs and guide you toward the best next step.

Option 3 — Support

Hi, I’m your support assistant.

I can help answer common questions, guide you through product features, and provide step-by-step assistance.

If your issue needs human support, I can help collect the right details and guide you to the next step.

Section 15

Error / rejection reasons

Uploads may be rejected if:

  • Face not clearly visible or person not centered
  • Unclear audio or video too short
  • Multiple people speaking
  • Distracting background or poor lighting
  • Copyrighted content, missing consent, or heavy filters/manipulation