Apple working on how to read back iMessages in the sender’s voice

Messages

AppleInsider could earn an affiliate fee on purchases made by way of hyperlinks on our website.

Apple customers can already ship audio recordings in iMessage or have Siri learn textual content messages again to them, however the patent describes a solution to have the gadget learn the textual content message within the sender’s voice as an alternative of Siri, utilizing a voice file.

“The voice mannequin is supplied to a second digital gadget,” based on the patent. “In some examples, a message is obtained from a respective person of a second digital gadget.”

Which means, when somebody sends an iMessage, they will select to connect a voice file, which might be saved on the gadget. If this occurs, the receiver might be prompted to determine in the event that they wish to obtain each the message and the voice recording.

“In response to receiving the message, a voice mannequin of the respective person is obtained,” the patent reads. “Primarily based on the voice mannequin, an audio output akin to the obtained message is supplied.”

Based on the patent, the iPhone in query would then construct a Siri-like profile for the sender’s voice, after which simulated it when studying that message and any future messages they obtain from that sender. The voice simulation mannequin is also despatched by itself so an individual’s contacts can obtain it earlier than messages.

It might provide extra personalization when family and friends textual content one another, as an alternative of listening to Siri’s voice when it reads messages. {Couples} may additionally hear messages in a extra private method, comparable to listening to “I really like you” of their associate’s voice.

The patent’s inventors are Qiong Hello, Jiangchuan Li, and David A. Winarsky. Winarsky is Apple’s director of text-to-speech expertise, whereas Li is a senior Siri software program engineer for machine studying at Apple, and Hu previously labored on Siri on the firm.

As typical, with patents, it will not essentially develop into a actuality, however it’s attainable given Apple’s current work with synthetic intelligence and voices. For instance, with iOS 11, Apple switched Siri’s voice from counting on recordings from voice actors to a text-to-speech mannequin utilizing machine studying.

In 2020, Apple acquired an organization referred to as Voysis that labored on bettering pure language processing in digital assistants. They used WaveNet expertise, which was launched by Google’s DeepMind program in 2016.

WaveNets are “deep generative fashions of uncooked audio waveforms” that can be utilized to generate speech that mimics any human voice.

Apple has additionally started using synthetic intelligence to relate particular audiobook genres as an alternative of utilizing people. So the patent is totally within the realm of risk of getting an Apple gadget ultimately study to learn messages in an individual’s voice.