Reference 1 is a clear AI avatar photo of the woman, used as her identity anchor — her face must exactly match this avatar photo throughout. Reference 2 is the source video of her speaking. Keep the woman completely unchanged from the source video — her face must exactly match the identity avatar (same facial features, same likeness), and her expression, hair, clothing, body, position and her n...
Бесплатный старт · Генерация видео и картинок с ИИ за секунды
