Whisk (https://labs.google/fx/fr/tools/whisk) generates sound images and videos from “inspirational” images that can be supplemented with an instruction.

“Inspirational” images represent object, scene and style. Many can be transmitted and they can be generated.

Models not specified: images probably with Nano banana and Veo video certainly version 2.

Experimental version, free, images seem unlimited and 5 videos (per month?).

Images for testing:
– https://commons.wikimedia.org/wiki/File:2019-04-06_Dieter_M._Weidenbach_StP_4058_by_Stepro.jpg
– https://commons.wikimedia.org/wiki/File:Lake_Willoughby_October_2021_003.jpg
– https://commons.wikimedia.org/wiki/File:Pez_murci%C3%A9lago_orbicular_(Platax_orbicularis),_mar_Rojo,_Egipto,_2023-04-19,_DD_01.jpg