How to make a Bad Times AI video

What does the Bad Times AI preset make?

The preset creates a new portrait video of the person in your photo performing a dramatic, camera-facing kitchen routine. A built-in 15-second reference guides the setting, camera and gestures. The photo supplies the new performer’s appearance; it does not need to show the kitchen or any particular pose.

The example above is a real Vismuse generation made from a fictional adult photo. Compare the starting image on the tool page with the generated result to see what the model preserved and what changed. A new take will not match every frame or mouth movement exactly.

1. Choose one clear person photo

Start with one person whose face, hair and upper body are easy to see. Include the outfit you want the generated performer to wear. A clear frontal photo gives the model more useful identity detail than a distant group shot.

• Use even light and a sharp, unobstructed face.
• Show enough of the upper body and clothing to guide the outfit.
• Avoid sunglasses, heavy filters, hands over the face and multiple people.
• Use a photo you have permission to upload and share.

2. Use the included kitchen reference

Open the Bad Times tool and upload your photo. The 15-second performance reference is already part of this preset, so you do not need to find or upload a video. The reference guides the kitchen, camera and motions; your photo guides the performer’s face, hair and clothing.

For your first take, leave the scene direction alone. That makes it easier to judge likeness and motion before asking for a visual adjustment.

3. Add a short instruction only when needed

A prompt is optional. If an appearance detail matters, add one focused request such as keeping your glasses or the color of your jacket. Avoid conflicting requests for a different room, camera and performance in the same first take; the included reference already supplies those elements.

4. Check duration, format and credits

This preset uses a 15-second portrait 9:16 video and exports an MP4. It is a short excerpt, not a full-length music video. With sound enabled, the matching original audio is attached to the export; the photo does not provide or clone a voice.

Watching the example is free. Personalized generation requires paid credits. Review the cost shown in the workbench before submitting. Each new generation uses credits, including another take after a completed result.

5. Review the full video and downloaded sound

Play the complete result rather than judging its poster frame. Check the face, hair and outfit while the performer moves and approaches the camera. Look for hand or gesture changes in the closer shot.

Listen to the downloaded MP4. The original sound can be included, but generated mouth movements and gesture timing may differ from the reference. Use photos and audio you have permission to publish, especially in commercial posts.

What if the result does not look like the photo?

Try a brighter, sharper, more frontal image with the hair and desired outfit visible. Remove competing appearance instructions and keep one specific request if needed. The reference guides motion but cannot guarantee exact likeness or timing.

Wait for a pending task to finish before submitting another request. If you want another completed take, check the current credit cost first. Results can vary even when the inputs are unchanged.

Can I use a pet or group photo?

This Vismuse preset is designed for one person from one photo. Pet and group results are not promised by this workflow. Choose a clear single-person image when you want a predictable lead performer.