Home Latest Feeds Technology News Voice Engine – technology from OpenAI allows for realistic voice cloning. Just a short sample is enough

Voice Engine – technology from OpenAI allows for realistic voice cloning. Just a short sample is enough

0
Voice Engine – technology from OpenAI allows for realistic voice cloning.  Just a short sample is enough

[ad_1]

Voice Engine - technology from OpenAI allows for realistic voice cloning.  Just a short sample is enoughToday’s AI technology includes not only chatbots and image or video generators, but also the ability to clone virtually any voice. Although there are already many such solutions on the market, the OpenAI organization has just presented the Voice Engine model, which allows you to achieve impressive results. It can create a very realistic voice of a specific person – for this purpose, it only needs a sample of a dozen or so seconds.

OpenAI showed a sample of the capabilities of its Voice Engine model, which allows for text-to-speech conversion using any source voice. However, the organization wants to approach the topic responsibly, so the solution is not yet publicly available.

Voice Engine - technology from OpenAI allows for realistic voice cloning.  Just a short sample is enough [1]

OpenAI’s GPT-4 model has been dethroned. Users have chosen the next king, which is the new product from Anthropic

The Voice Engine model has been in development for a long time, and now OpenAI has decided to present its capabilities. It must be admitted that the results obtained are literally fantastic. The mentioned 15-second sample is not only enough to clone the timbre of the voice. On its basis, you can reproduce various emotions and change the pace of speech. All this adds up to a very realistic voice that is nothing like the old speech synthesizers. Examples shown by OpenAI on this page they show that a cloned voice can read texts in another language while retaining someone else’s accent. This allows for smooth translation of videos. Another useful use of Voice Engine is to help people who, for some reason, cannot speak normally and freely. You can use a short voice sample from before the accident.

Voice Engine - technology from OpenAI allows for realistic voice cloning.  Just a short sample is enough [2]

Stable Diffusion will create graphics in a fraction of a second. MIT researchers have presented a method that will improve any AI model

Of course, the solution in question carries just as many, or maybe even more, risks. OpenAI recognizes this and intends to first discuss how this technology can be wisely rolled out on a broader scale. Additionally, each recording created using the Voice Engine will have a watermark implemented. This is to allow easy identification of the source. Currently, only a select few people have gained access to this model, and only after the implementation of the mentioned security measures will it be released on the market. This is a very good approach, as many similar solutions have already contributed to spreading disinformation. The world is changing beyond recognition and on the one hand it is fascinating, but on the other it is literally terrifying.

Source: OpenAI



[ad_2]

LEAVE A REPLY

Please enter your comment!
Please enter your name here