Strangers in Silence: Behind-The-Scenes Chroma Awards Entry

This is my entry into to the BEHIND-THE-SCENES category, looking at how I created my entry for the FILM TRAILER category. I wanted to showcase that with these AI tools, we can now create spec trailers/sizzles which build worlds that feel real, cinematic and tonally in line with what the final vision for our show or movie could be. I haven't used any real footage, sounds or music in this trailer, everything is AI-generated. For ideation, this is such an exciting time to be a creative in this field!

STRANGERS IN SILENCE: After a devastating loss leaves her untethered, a grieving woman takes an unusual job as a “silent companion” — wordlessly sitting beside clients on their train journeys to keep them company. But as the quiet she once sought begins to consume her, she must choose between fading into other people’s lives… or finding the courage to reclaim her own.

Tools used:

• ChatGPT 4o - helping create the idea and script. Then generating JSON prompts for Veo3.
• Google Veo3 - Essentially all the shots not of humans or of humans that don’t need to speak. The occasional generated audio from these videos have been kept in but most haven't.
• Runway - Consistent character image gen for main characters via Runway References and then Runway's Act-Two for 99% of the speaking shots, using motion capture on my iPhone.
• Sync.so - Fixing all the lip-sync moments to look and feel more accurate and realistic.
  
• Higgsfield - For the initial shot of Claire (the lead character) on the train via the Flux Context Pro tool. I then used that generated image in Runway References to build her out. I also generated 2 x speaking shots using the Higgsfield Speak tool.
• Google Gemini 2.5 Flash Image Preview - There are three shots I used this tool for. When Claire sees the poster in the train station, the exterior of the shop and then the close up of the shop sign. I then animated the images using Veo3 or Runway (as the exterior shot features people and in the UK we can’t upload shots of people’s faces to Flow)
• Elevenlabs Text-To-Speech V3 - I created the voice of Claire and Joe using the voice creation tool and then the bonus voices are Elevenlabs pre-made voices.
• Elevenlabs Music - Both music tracks are created using Elevenlabs. I then blended them together in the edit.
• Elevenlabs Sound Effects - ALL SFX and foley sounds are created using the Elevenlabs SFX generator.
• Adobe Premiere Pro - I have edited the entire video using Premiere Pro, as well as colour-corrected and graded it myself.

Built With

• adobe
• adobe-premiere-pro
• chatgpt
• elevenlabs
• higgsfield
• nano-banana
• premiere-pro
• runway
• sync.so
• veo3