ad
ad
Topview AI logo

How I made AI Avatar with Open-Source Tools | Ultimate Guide

Science & Technology


Introduction

Creating an AI avatar can seem like a daunting task, especially with the high costs and quality inconsistencies associated with many AI Avatar services. However, I've taken the time to meticulously craft a method that utilizes open-source tools to generate a satisfactory result without breaking the bank. This guide shares my process in creating an engaging AI avatar while keeping expenses to a minimum.

Getting Started: The Tools Needed

To begin with, I recommend using Live Portrait. This tool produces impressive results and works efficiently even on low-performance PCs, although a GPU is still necessary. An important limitation of Live Portrait is that it primarily focuses on animating the facial features and head gestures. However, without corresponding body movements, the result can seem unnatural. To remedy this, I devised a way to animate the body separately using various AI services.

Why Not Other Tools?

While there are alternatives like Hollow and Set Talker, their inefficiencies made them less appealing. For example, using Set Talker on my 390 Ti GPU resulted in slow processing, taking around 30 minutes for minimal output. Additionally, installing the latest version of Set Talker proved challenging due to numerous dependency issues.

Steps to Create Your Avatar

1. Generating Your Avatar Image

Begin by generating an image for your avatar using an image generator of your choice. I utilized Focus UI to create a detailed image and extend it to a satisfactory result.

2. Animating the Body

Since Live Portrait can’t animate the body, I turned to options like Clink AI and Luma AI for generating videos from images. Although both are free, the wait times can vary significantly, with Clink AI taking as long as three days for results, while Luma AI provided results after around 15 minutes.

3. Editing the Video

The final video will initially have a watermark and a short duration of only five seconds. I recommend using software like DaVinci Resolve, even the free version, to edit your video. Here’s a brief on what to do:

  • Import your video and use the Patch Replacer to hide the watermark.
  • Extend the length of your video by copying and reversing the speed of the clips.

4. Creating a Driving Video

For the animation, you need a driving video that can be made by recording yourself on a smartphone or using a service that creates animation based on audio. Set Talker is again a free option for generating such animations.

5. Finalizing with Voice Modulation

Once the animation is done, you need to ensure the accompanying audio fits. I extracted the audio from the video and used 11 Labs to generate new voice audio. After adjusting audio parameters, you can download and sync it with your new animated video.

6. Reviewing the Final Product

Combine the various elements in an editing tool to get your final result, tweaking until you're satisfied. The final output will showcase your AI avatar in action, with animations that feel more natural and cohesive.

Conclusion

With some time and effort, creating an AI avatar using open-source tools is absolutely achievable. While it may seem overwhelming initially, the process becomes manageable after you repeat the steps a couple of times. You can create professional-looking avatars with minimal investment.


Keywords

AI Avatar, Open-Source Tools, Live Portrait, Clink AI, Luma AI, DaVinci Resolve, Voice Modulation, Set Talker, Animation, Image Generation.


FAQ

Q1: What is Live Portrait and how does it work?
A1: Live Portrait is a tool that animates facial features and head movements based on a static image, enabling the creation of AI avatars.

Q2: What alternatives to Live Portrait are available?
A2: Alternatives include Hollow and Set Talker, but these have limitations related to efficiency and ease of installation.

Q3: How can I generate a video from my avatar image?
A3: You can use AI services like Clink AI and Luma AI to create a video from a still image, though they may require significant wait times for processing.

Q4: What software do I need for editing the final video?
A4: DaVinci Resolve is recommended as a free tool for video editing, where you can remove watermarks and adjust video lengths.

Q5: Can I use my voice for the avatar?
A5: Yes, you can record your voice and use voice modulation services like 11 Labs to create a fitting voice for your avatar.

ad

Share

linkedin icon
twitter icon
facebook icon
email icon
ad