Hedra: Make Photo Portraits Speak and Sing 255

Upload any person's photo along with a voice clip, and Hedra generates a dynamic video of that person speaking or singing — with lip movements, expressions, and posture perfectly synced to the audio.

#Ai工具箱 #Ai视频生成

Visit Hedra: Make Photo Portraits Speak and Sing

Hedra: Make Photo Portraits Speak and SingIntroduction

Hedra Labs has launched a research preview of Character-1. By uploading any photo of a person and a voice, it can generate a personal dynamic video that talks and sings—in other words, a powerful tool that generates talking and singing videos from text and images, while keeping the lip movements, expressions, posture, and spoken content perfectly matched.

Hedra can be used on desktop and mobile devices, with unlimited duration (30 seconds available in the open preview). It supports expressive dialogue, singing, and rapping characters, with natural posture and expressions.

Hedra Features:

Text and image generation: Character-1 can generate talking and singing videos from text and images.
Unlimited duration: During the open preview, Character-1 supports video generation of unlimited duration, up to a maximum of 30 seconds.
High efficiency: Every 60 seconds of input can generate 90 seconds of video.
Expression sync: The expressions and lip movements of the generated video characters match the spoken content.
Diverse characters: Character-1 supports generating character videos with expressiveness and controllability.
Virtual world building: Hedra also plans to launch a Worlds feature, allowing users to build their own virtual worlds.
Creative control: Hedra provides complete creative control, letting users imagine worlds, characters, and stories.

How to Use Hedra:

1. Upload voice or text prompt: It supports text-to-speech with currently 6 voices available. Enter or paste the text you want the character to say and select a preset voice, or upload audio directly.

2. Upload an avatar: The avatar image can be your own photo or automatically generated by entering prompts. Currently only English is supported; Chinese is not supported.

3. Generate.