Compare the Sign of the Times AI Video references
These complete external source performances are credited for reference. They are not results made with your photos on this website.
Native wide sea cliff
The cap-wearing adult on the initial left and red-haired adult on the initial right, in the source’s native wide sea-cliff frame.
@rybka_18 · external source reference · 14.8 seconds
Portrait-window source
The same complete couple scene in the supplied portrait frame, with its original black space around the wide image.
@rybka_18 · external source reference · 14.8 seconds
Make your video from photos
Watch the full reference and review its actual length, framing and performer roles.
Choose Create my video and assign photos to the labeled positions. For a duo, use separate photos or one shared two-subject photo.
Review the selected clip, background option and photo assignments, then sign in and create using one video credit.
Follow the task in My videos, review the full result, prepare its available reference audio and download the MP4.
Everything runs in your browser. There is no dance to learn or performance to record. Processing can take several minutes; check the existing task in My videos before submitting again.
Price and what your video includes
Each selected reference costs one video credit to personalize. Watching is free, and downloading a completed result is included. There is no subscription.
Discounted packs: Previously $9.90 $6.99 for 1 video, Previously $19.99 $16.99 for 3 videos, Previously $27.99 $24.99 for 5 videos. See Pricing for validity and purchase terms. A confirmed failed generation releases its credit; an uncertain task remains under review.
Generation is upgraded from 480p to a 720p request using Higgsfield Genjutsu Light. Credits never expire. The selected reference supplies its timing and original audio; completed downloads are MP4 files. Source preview resolution is not a guaranteed export resolution.
Choose your photo
Assign clear consenting-adult portraits to the cap-wearing and red-haired partner positions, or use one shared picture with exactly two readable subjects. Chest-up images work well because profile views and the final embrace bring faces close together. Avoid sunglasses, a phone across the face or another person overlapping the selected subject. The role labels identify where each character starts, not a promise that it stays on that side after turning. Both frame options use the same character mapping. This page does not add the separate parent-and-child, custom ruined-city or person-and-pet scene advertised elsewhere in the wider trend.
Use JPG or PNG up to 10 MB. Select the reference first; changing it clears earlier photo assignments. Use your own pet or consenting adults.
About this reference
The researched page credits this particular AI sea-cliff performance to @rybka_18’s May 6, 2026 post. The song is Harry Styles’ Sign of the Times, and the wider format became known as the couple-apocalypse trend. The selected scene combines a tense exchange, a flash over the sea and an embrace before the wave reaches the couple. Two supplied files present the same complete performance in different frames. The 800 by 450 native-wide file keeps the sea visible across the shot; the portrait file retains the source’s black space around that wide image. Neither has been recropped, trimmed or converted here.
Historical account and measurements: Source report.
Watch the performance
Play the complete source before assigning photos. Source dimensions describe the preview, not a guaranteed output format.
| Part | What to watch |
|---|---|
| Initial exchange | The cap-wearing partner starts on the left, facing the red-haired partner. |
| Sea flash | The couple turns attention toward the light over the water. |
| Wave rise | The sea fills more of the background as the camera view changes. |
| Embrace | The partners move close together as the wave breaks behind them. |
Casting and caption ideas
Use the scene as a dramatic fictional character edit rather than presenting it as footage of a real disaster. A caption about holding on through an exaggerated bad day matches the emotional arc. Compare identity in the side profiles and final embrace, where one face can hide the other. Select the wide reference when you want the full sea-cliff composition; select the portrait-window file only if you prefer its supplied framing. The generator requests its configured output settings, so source dimensions do not guarantee the finished video’s exact framing or effects.
Sound and sharing
Sign of the Times · Harry Styles. The download prepares the audio actually present in the selected reference. It does not clone your voice or grant music rights. Check the posting platform’s music permissions and keep its required AI label.
The external preview is not evidence of a completed generation with your photo. Review the existing task before spending another credit; retry audio preparation on that task if needed.
The recording is Harry Styles’ song, not new dialogue from your input portraits. Share the scene as an AI fiction edit and review the posting platform’s music rules.
Review the result
| Issue | What to check |
|---|---|
| Faces swap during the embrace | Check both initial character assignments and review the close contact. |
| The wave covers a face | Review identity before and during the final shot; dramatic effects may vary. |
| The portrait version has black space | That space is part of the supplied file and is intentionally preserved. |
| A parent-and-child scene is expected | Only the researched adult couple references are registered on this page. |
Sign of the Times AI Video FAQ
Which frame option should I choose?
Native wide keeps the sea-cliff composition across the full picture. The portrait-window option keeps the supplied black space around that same scene. Neither file is cropped here, and source framing does not guarantee the final generated framing.
Is the portrait file cropped from the wide file here?
No. Both were downloaded as supplied. The portrait version’s surrounding black space is preserved.
Can I replace one partner only?
Yes. With separate photo assignments, fill the desired role and leave the other unassigned.
Does this make a real disaster recording?
No. The preview is an external AI scene, and a personalized result is another fictional video edit.
Is watching the preview free?
Yes. Watching is free. Creating a personalized video requires signing in and one video credit, including a two-role edit. See the current one-time packs on the pricing page.
What does the generator request?
The existing generator requests 720p and prepares an MP4 with the selected reference’s audio. It does not provide the source site’s model menu or guarantee every background, effect or clothing detail.
Where can I follow my video?
Use My videos in the same signed-in account. A confirmed failed generation releases its reserved credit; an uncertain task remains under review. No new paid generation is needed merely to retry a download.