ml-talking-face is a free web application that empowers you to create a talking face video using input text. This input text can be supplied in one of four languages: Chinese (Mandarin), English, Japanese, and Korean. The system additionally provides the option to select the target language for the output speech. In cases where the input and target languages differ, the input text will be translated to the target language through the use of the Google Translate API.
Studies that have been recently conducted have always focused on building a train-once-use-everywhere model that aims for generalization from any source speech to any target entity. However, a notable drawback of such models is their limited efficacy when confronted with entirely different training and testing languages. ml-talking-face introduces a different approach in which the generation system has been built on familiar identities which means to say that they target identity is the same as the training identity resulting in enhanced adaptability to diverse languages.
How it Works:
1. Click on this link to navigate to ml-talking-face that has been hosted on the Hugging Face platform. You are not required to sign up or register for a free account to use this tool.
2. At the outset, type or paste the input speech in the corresponding text box in any of the four allowed languages. A maximum of 200 characters are allowed for English and 100 characters for Chinese, Japanese, and Korean.
3. Next, choose the output language by clicking on any of the corresponding buttons. As stated above, if input and target languages are different, Google Translate API will be used to translate the input text to the target language.
4. Adjust the duration of the talking face video using the slider. You must remember that the bigger the value, the slower will be the output speech.
5. If required, select an action that will be performed by the target character during the speech such as Hand, Both Hand, Sorry etc. Additionally, choose a Background Image / Video for the output by selecting any corresponding button and finally click on ‘Submit’.
6. Wait patiently for some time while ml-talking-face analyses and processes all the inputs that you have submitted to render the talking face video without any watermarks in the right-side pane. This process may take up to 30 seconds due to the latency from Hugging Face spaces and video rendering.
7. To download the talking face video to your computer, click on the 3-dots icon at the bottom right and choose ‘Download’.
Closing Comments:
Overall ml-talking-face is a fine online system hosted on Hugging Face spaces that you can use to generate a talking face video by specifying the input speech and choosing the target language that can be the same or different.
Go ahead and try out this tool and do write to us about what you think. Click here to navigate to ml-talking-face.
Are you searching for an AI-based tool to create videos and animations from simple text. You can consider trying Moonvalley that we have covered in another article. This is a very powerful AI-driven text-to-video generative model, allowing you to create stunning high-definition cinematic videos and animations in various styles, including 3D animation, Comic book, Manga, Realism, and more through natural language text prompts. Click here to read more about it.
