Adobe has taken advantage of its conference MAX to present some of his work. The project VoCo has impressed the audience. Has the image of famous software Photoshop, VoCo has the ambition of becoming the benchmark of speech synthesis. One of its features allows you to mimic the user’s voice to utter words.
Project Voco Adobe : speech synthesis, with any voice,
VoCo, a voice-print of a user imitated to perfection ?
The Project VoCo is under development. It is the fruit of a collaboration between the research teams of Adobe and the university of Princeton. In terms of appearance, it is quite similar to other specialized applications in the area, but it is just an appearance.
one of its features was particularly surprised. Adobe the summarized
” We have developed a technology that allows to simply type the words you want to change in the voice synthesis [...] The algorithm does the rest and generates words that seem to come from the original person “.
In the clear, the software is able to edit phrase to add or remove words while keeping the original voices. The diction, the tone, the effects would be almost perfect although the lighthouse changed was never pronounced by the user.
The program has the ability to modify audio content by generating, for example, of new words which appear to come from the speaker. The most impressive one being the fact that the voice-print of the original is preserved.
VoCo, a project in development
For the moment VoCo is a project in development. His prowess ask for the power calculation and time. To achieve to master the voice-print of a user an analysis of a recording of his voice twenty minutes is necessary.
there are also a lot of question around its arriving on the market. Her ability to transform is not trivial. What is going to happen around the credibility of the statements and records, said the official. How can I prove it ? How to ensure veracity and authenticity ?
No comments:
Post a Comment