We've prepared detailed instructions for you to create a world-class avatar!
Creating your digital avatar is an exciting step towards a new form of expression. To ensure we capture your essence accurately, we ask you to follow these simple guidelines during the recording:
Be Natural: Talk as if you're sharing a story about how any topic that interests you. The content of your speech isn't crucial, but your natural mannerisms are.
Stay Half-Body in Frame: Make sure the video captures you from the waist or chest up. This helps us understand your natural gestures.
Gesture Under the Shoulders: Limit the use of hand movements. We recommend that it be no more than one or two per minute of recording and that they always remain below the shoulders.
Continuous Recording: We require a constant, unedited video of 2-3 minutes. If any errors occur, please start over. This continuity is vital for AI training.
Pause After Speaking: After finishing a sentence, pause for 1 second with your mouth closed before continuing. This small gap is crucial for the AI to learn your speech patterns accurately.
Pause at the beginning: Please at the beginning of the recording we need about 3 to 5 seconds of you not speaking or moving your hands, you can move your head slightly so that it is not completely static. Also avoid making exaggerated facial expressions or gestures.
Hello! I'm feeling good right now with a relaxed tone, and I'm very confident that I can do well on this recording 😃. I'm in front of the camera and ready to start.
<Close your mouth and breath through nose >
The lighting is good, and there aren't any harsh shadows on my face. My pronunciation is clear, and I'm feeling very relaxed because this process has been simple, easy, and enjoyable 🥰.
<Close your mouth and breath through nose >
I'll be making some natural, slight hand movements throughout the recording 👐, but I won't be moving my body or head too much. These subtle gestures👐 will help me look more natural and relaxed.
<Close your mouth and breath through nose >
I'm not wearing glasses or any other accessories, so my face can be clearly displayed. My tone of voice is steady, and I'm speaking in a natural way. I'm taking a pause between sentences, closing my mouth and breathing through my nose. I'm in a relaxed mood, and this recording process is really fun 🥰.
<Close your mouth and breath through nose >
I also want to share something interesting with you. I'm thinking about some happy memories, and it's making me feel like life is very full and interesting. I'm holding onto this relaxed and happy mood 😃 and hoping to keep it forever 🥰.
<Close your mouth and breath through nose >
Now, I'm going to try adding in some slight hand movements 👐 to make things feel more natural. My tone of voice is still steady and emotional, and I've been doing a great job up to this point. I'll continue doing my best until the end of the recording.
<Close your mouth and breath through nose >
It's a great pleasure to meet you today, and I'm looking forward to seeing how this recording turns out. So far, everything has gone well, and I hope things are going well for you too 😃.
Capturing the Perfect Shot
Your role in this journey is critical. Here are the essentials for capturing high-quality footage that meets our AI’s learning needs:
The video has to be recorded in 16:9 horizontal format.
Use a simple background. It highlights the subject without any distractions.
Ensure the audio is clear, with good volume and no background noise or echo. Crisp audio is crucial for creating a realistic avatar.
Aim for bright, even lighting. Avoid any shadows or reflections on the subject to ensure every detail is captured accurately.
2 Clips
Record 2 clips, each lasting 2-3 minutes, where the subject talks and uses hand gestures naturally, but we recommend doing it infrequently, no more than once every 30 seconds as the AI tends to use these movements excessively.
If you detect that the person is making too many hand movements, please stop the recording and start again.
2 Clips
Record 2 clips, each 2-3 minutes long, focusing on the subject talking without hand movements. We recommend that the subject move their head and shoulders slightly during recording so that they do not look static.
2 Clips
Capture 2 clips, each 1 minute long, of the subject not talking, to be used in the avatar’s waiting mode. We recommend that the subject move their head and shoulders slightly during recording so that they do not look static.
Securing Permissions
Please record a short clip where the subject verbally provides consent for the use of their likeness.
Follow this script:
"I, [Name], hereby declare that I authorize Thingy to use the footage of me to build a Thingy Avatar and use it in my Thingy account."
Replace "[Name]" with the subject's actual name. This clip is essential for legal purposes and does not need to follow the above recording rules.
You can either stand or sit while shooting, but please ensure that your face remains clear and visible and that you are facing the camera directly.
Avatar Creation has the voice cloning feature. After cloning your voice, your avatar will use your voice in the generated videos.
You can also customize the voice clone separately recording a 3-5 minutes audio in a professional or semi-professional microphone and send it to us with your video recording to train a model with your voice.
The hand movements and facial expressions of your avatar are derived from your original materials. If your original materials do not include hand movements or contain too many hand movements, your avatar will exhibit the same behavior. The same principle applies to facial expressions; if your avatar did not display any expressions in the original materials, it will reflect the same. Therefore, we recommend that if you want your avatar to have rich and natural hand movements and cheerful expressions, you can adjust your hand movements and facial expressions when capturing the materials.
But we recommend doing it little, no more than once every 30 seconds since the AI tends to use these movements excessively.
