They look unnatural when the source photos have poor lighting, obscured faces, or mismatched angles, causing the AI to struggle with mapping the faces onto the moving 15-second video template.
AI video generation takes static images and animates them. When you make a duo dance video, the system extracts the facial features from the photos you provide and maps them onto the moving bodies in the template. The ba ba ba dance is a fast-paced routine. The bodies in the template move quickly, turn, and bounce. If the source photos do not provide clear, front-facing data, the software has to fill in the missing information. This estimation is what causes unnatural results, such as warped faces, flickering textures, or floating heads.
Lighting plays a major role in how the final 15-second video looks. If you upload a photo with a strong shadow on one side of the face, the AI maps that shadow as a permanent feature. When the character in the video turns their head, the shadow stays fixed to the face, ignoring the lighting in the video environment. This creates a mask-like effect. To prevent this, you need photos with flat, even lighting across the entire face.
The angle of the face in the original photo also determines the quality of the result. The ideal photo shows the subject looking directly at the camera. If the subject is looking down, up, or to the side, the AI struggles to align the eyes, nose, and mouth with the template. When the video template requires the character to face forward, the AI stretches the side-profile photo to fit a front-facing head, resulting in severe distortion.
Resolution and clarity matter. A blurry photo gives the AI poor data. The resulting face in the video will look muddy or out of focus compared to the rest of the video. The software cannot add detail that does not exist in the original image. You must start with a sharp, well-lit photo for the face mapping to look natural during the rapid movements of the duo dance.
What photos do you need to prepare for a two-person dance?
To use the 2 Person Dance AI, you must prepare two separate photos. The tool requires one photo for each person in the video. You cannot upload a single photo of two people standing together; you must crop them into two distinct image files or choose two completely different pictures.
When selecting these two photos, look for images where the subjects are well-lit and facing the camera. Passport-style photos or clear selfies taken in daylight work well. The face should take up a significant portion of the image. If you use a wide shot where the person is standing far away in the background, the face will be too small. When the software enlarges a small face to fit the template, it loses resolution and looks unnatural.
Avoid photos where hair, hands, or accessories cover the face. Sunglasses hide the eyes, and hats cast heavy shadows over the forehead. The software needs to see the boundaries of the face, the eyes, the nose, and the mouth to map them accurately. If a hand is resting on the chin in the original photo, the AI might map the hand texture onto the chin in the video, making it look like a growth rather than skin.
Here is a summary of what makes a good photo versus a bad photo for this process:
| Feature | Good Photo | Bad Photo |
|---|---|---|
| Lighting | Even, flat lighting across the face | Harsh shadows on one side, dark rooms |
| Angle | Subject looking directly at the camera | Subject looking away, profile shots |
| Obstructions | Clear face, hair pulled back | Sunglasses, hands on face, heavy bangs |
| Resolution | Sharp, in-focus, high resolution | Blurry, pixelated, low resolution |
| Framing | Head and shoulders visible | Full body shot from far away |
How do you make your own ba ba ba dance video?
Creating the video requires a few specific steps on the site. You must upload your prepared photos, create an account, and purchase credits to generate the final file.
First, you need to visit the main page to make yours. The interface is straightforward. You will see areas to upload the two photos you prepared. Upload the photo for the first person, and then upload the photo for the second person. Make sure you have the right files selected.
Next, you need to sign in to the site. The platform offers two ways to access your account. You can sign in using your Google account, or you can sign in using an email address. Choose the method you prefer. Once you sign in, the site will keep track of your uploads and your video generations. The site interface is available in two languages: English (en) and Chinese (zh).
Before you can generate the video, you need to have credits in your account. The pricing structure is simple. You pay $5.00 one-time. This payment gives you 60 credits. These credits are valid for 365 days from the date of purchase. You do not need to use all 60 credits at once; you can keep them in your account and use them over the course of the year. For more details on the cost, you can check the pricing page.
Payments on the site are processed through FloatPass. When you click to buy the credits, you will be directed through the FloatPass system to complete the transaction. Once the payment is successful, the 60 credits will appear in your account immediately. If you need assistance with your account or run into an issue, you can reach out via email at contact@bababadance.com.
With your photos uploaded and your credits ready, you can start the generation process. Click the button to create the video. The software will process the two photos, extract the faces, and map them onto the dancers in the template. The output will be one 15-second video featuring the faces from your photos performing the viral duo dance.
What mistakes should you avoid when uploading photos?
Many people experience unnatural results because they rush the photo selection process. The most common mistake is using a screenshot from a video or a heavily compressed image sent through a messaging app. These files often contain compression artifacts that confuse the AI. Always try to use the original photo file from your camera roll.
Another mistake is ignoring the expressions on the faces in the photos. While a slight smile is fine, extreme expressions like a wide open mouth or a deep frown can cause issues. The AI maps the static expression onto the moving body. If the person has their mouth wide open in the photo, they will have their mouth wide open for the entire 15-second video, which looks unnatural. A neutral expression or a soft smile maps best onto the moving template.
People also make the mistake of using photos with heavy filters. Filters that alter the shape of the jaw, enlarge the eyes, or smooth the skin to an extreme degree remove the natural textures the AI needs to create a realistic mapping. Use unfiltered, natural photos for the best results. If you want to read more about the general process, you can check the blog or read the specific guide.
How does the 365-day credit validity work?
When you make the $5.00 one-time payment through FloatPass, you receive 60 credits. The 365 days of validity starts the moment the transaction is complete. This means you have a full year to use those 60 credits.
If you only want to make one 15-second video today, you will use a portion of those credits. The remaining credits stay in your account. You can return to the site next week, next month, or six months later, sign in with your Google or email account, and use the remaining credits to make another video with different photos. You just need to remember to use them before the 365 days pass.
Why is the video exactly 15 seconds long?
The output of this specific tool is one 15-second video. This duration is not random. The viral ba ba ba dance routine itself is a short, high-energy sequence that fits perfectly into this timeframe. Fifteen seconds is the standard length for short-form video content on most major social media platforms. It is long enough to show the full dance routine, but short enough to keep the viewer's attention from start to finish.
When the AI processes your two photos, it generates exactly 15 seconds of footage. You do not need to trim or cut the video after you download it. It is ready to be saved to your device or shared immediately. Because the generation process maps the faces frame by frame, keeping the video to 15 seconds ensures the processing time remains reasonable while still delivering the complete duo dance experience.
Where else do people look for AI video tools?
While you can make the viral duo dance here, people searching for AI video generation often look at a variety of different platforms for different types of video editing and generation tasks. Some of the domains people also look at include kling.ai, viggle.ai, freebeat.ai, mangoanimate.com, media.io, and capcut.com. These sites offer various tools, and users often explore them when researching how to animate photos or edit video content.
Frequently asked questions
How much does it cost to make a video?
The cost is a $5.00 one-time payment. This payment gives you 60 credits to use on the site.
How long do my credits last?
Your 60 credits remain valid for 365 days from the time of your purchase.
How long is the final dance video?
The tool generates one 15-second video featuring the two photos you uploaded.
What payment processor does the site use?
All payments on the site are processed securely through FloatPass.
How do I sign in to the platform?
You can sign in to your account using either Google or your email address.
What languages is the website available in?
The website is available in two languages: English (en) and Chinese (zh).