How to Make a Smalltown Boy AI Video with Two Photos

1. Prepare two separate identity photos

Use one subject per image. Start with an unobstructed face, similar camera angles and enough visible clothing or markings to distinguish the pair. Avoid a group crop that includes another face. Keep a copy of the original inputs so you can compare them with the output.

Our preview accepts JPG, PNG and WebP files up to 10 MB each. A video provider may have different limits. The photo selection guide covers framing in more detail.

2. Assign the two roles before rendering

Write down Subject A and Subject B. Decide who has the first foreground reaction and who starts behind them. The exact left/right convention varies by tool, so follow the labels on the selected preset rather than assuming a universal order.

For couples, make sure the references are easy to tell apart. For pets, use distinctive fur patterns or collars. Similar inputs make it harder to notice an identity swap during review.

3. Choose a preset or a supported motion reference

A named template is usually the most direct route to a fixed sequence. With a general video model, you may need identity images, a reference clip and a written movement description. Confirm that the model accepts those inputs together; a text prompt alone does not guarantee matching cuts or choreography.

Use a reference clip you are allowed to use. Keep notes about its shot order and role exchange. The prompt guide provides editable examples and a shot checklist.

4. Check the entire result

• Opening close-up: is Subject A recognizable?
• Role exchange: do the identities follow the intended roles?
• Wide shot: are both subjects still present and distinct?
• Movement: do limbs, faces or markings merge or change?
• Final frame: does the clip finish cleanly rather than cutting off a gesture?

Change one variable at a time when retrying: the input image, role order, reference or prompt. This makes it easier to tell what improved the result.

5. Add the song and prepare the export

For searches such as “how to create AI video with Jimmy Somerville Smalltown Boy,” the first decision is the dance format and the audio version. You do not need to generate or imitate a singer’s voice to prepare a dance edit. Match the chosen audio to the sequence in your editor.

Review the exported file on a phone. Check framing, sound, duration and captions before uploading. See the song and remix guide if the timing feels different from the reference.

Prepare the photo pair here

Use our browser-local studio to select and compare the two images, then download a PNG composition. AI video generation is coming soon in this studio. The tool-selection guide links to public provider pages for the rendering step.