PSYCHE AI Dynamic Digital Human Avatar Customization- Recording and Upload Guidelines for Video Materials
1. Cloning Your Digital Human
PSYCHE AI cloned digital humans are designed for various professionals, including:
- Entrepreneurs and Brand Founders: Those who wish to convey their company’s vision without being on camera constantly.
- Short-Form Content Creators: Video bloggers and social media managers aiming to enhance content production efficiency and reduce costs.
- Specialists in Vertical Fields: Educators, lawyers, doctors, and other experts who want to quickly transform their knowledge into engaging video content.
In short, if you have ever wanted a digital double, this is for you. Your submitted video material is used to train an AI-driven model that accurately replicates your appearance, expressions, gestures, and voice. Rest assured, PSYCHE AI ensures your digital human model comes with privacy protection and exclusive usage rights.
2. Pre-Recording Requirements
1. Recording Equipment
For best results, use a high-performance smartphone with a high-resolution camera. We recommend devices such as the iPhone 14 Pro Max or newer models. Key settings include:
- Camera Settings:
- Resolution: 1080p or higher (4K is recommended for optimal quality)
- Frame Rate: At least 30 FPS (ensure your device is set to a stable 30 FPS or above)
- Aspect Ratio: 16:9 (landscape) or 9:16 (portrait)
- Format: MP4 or MOV
- Additional Tips:
- Disable HDR mode on iPhones.
- Use a tripod or phone stand to maintain stability during recording.
- Clean your camera lens prior to recording.
- Lock exposure and white balance, and use manual focus if available.
Note: Some devices may have different settings. Always choose the highest available resolution and frame rate to improve training outcomes.
2. Recording Environment & Lighting
A consistent, well-lit environment is crucial:
- Lighting:
- Ensure abundant, stable, and evenly distributed lighting.
- The subject’s face should be illuminated softly and uniformly.
- Environment:
- Choose a spacious area where your full upper body is visible.
- Record in a quiet room free from echoes and background noise (including music and other human voices), as these can interfere with lip-sync training.
- Background Considerations:
- Since the background in your video will be faithfully reproduced, select a well-lit and uncluttered setting. Avoid capturing identifiable information (e.g., other people’s faces, license plates, or door numbers).
3. Clothing and Makeup Guidelines
To ensure the best training quality:
- Appearance:
- Wear clean, non-reflective clothing that contrasts with your background.
- Avoid garments or accessories made of reflective materials, particularly metallic items.
- Maintain a neat hairstyle—secure stray hair with styling products if necessary.
- Opt for natural, even makeup and avoid excessive shine.
- Eyewear:
- Do not wear glasses with visible frames (contact lenses are preferable). If you must wear glasses, minimize the chance of glare.
- Post-Processing:
Any desired effects, such as face slimming or brightening, should be applied during or after recording. Note that once the digital human model is trained, modifications cannot be made.
4. Speech Material
Prepare a 5-minute speech on a topic you are familiar with. The content should ideally align with your future usage—for instance, product marketing should focus on discussing key product features.
3. Recording Content and Process
1. Subject Positioning and Presentation
When recording your video:
- Positioning:
- Sit or stand in the center of the frame with your eyes directly facing the camera.
- Avoid tilting your head up or down.
- Ensure your entire face and shoulders are visible throughout the video.
- Speech and Gestures:
- Speak clearly, at a measured pace with natural pauses.
- Avoid distracting mouth movements (e.g., lip licking, tongue protrusion, or pouting).
- Allow for subtle head movements (within ±10°) that naturally follow your speech.
- Use natural hand gestures without blocking your face or neck.
2. Recording Process
Follow these steps to capture high-quality footage:
- Starting the Video:
- Begin with a few seconds (approximately 5 seconds) of a neutral state—mouth closed, body relaxed.
- During the Recording:
- Walk and speak naturally for around 3 minutes.
- Record in either vertical (9:16) or horizontal (16:9) orientation at a minimum resolution of 1080p (4K is preferable).
- Ensure audio is synchronized with the video.
- Additional Considerations:
- If using a teleprompter or similar aids, avoid constant screen-gazing to maintain natural eye contact.
- If you make a mistake, pause for 3 seconds and then continue speaking without stopping the recording.
- Ending the Video:
- Conclude by returning to the initial neutral state (mouth closed, relaxed posture) for at least 1 second.
- Strive for consistency between the starting and ending poses.
Remember: The digital human’s head movements, body gestures, beauty enhancements, and background are dynamically selected from your footage. This means the quality and naturalness of your recording directly impact the final digital human model.
4. Video Processing and Submission
- After recording, you may apply light beauty adjustments, but make sure the image quality remains intact (1080p and 30 FPS or higher).
5. Review Standards
Your video will undergo the following review process:
- Video Quality: We will check the clarity, lighting, resolution, and frame rate to ensure they meet our standards.
- Movement and Expression: We’ll review the subject’s gestures and facial expressions to ensure they align with the training requirements.
- Lip-sync Accuracy: We’ll assess lip movements and overall language synchronization, avoiding unwanted actions (e.g., licking lips, sticking out the tongue) that may affect training.
6. Technical Support and Contact
For any technical support, please email us at: cs@psyai.net
Please include the subject line: [Digital Human Training Material Recording Issue]
Thank you for your support and participation!