coursesfromhome could earn affiliate commissions on purchases made by means of hyperlinks on our website.
Apple customers can already ship voice recordings in iMessage or have Siri learn textual content messages to them, however the patent states that utilizing audio information, the system can converse the sender’s voice as a substitute of Siri. It tells you the right way to learn textual content messages.
In line with the patent, “speech fashions are offered to a second digital system.” “In some cases, messages are acquired from respective customers of second digital units.”
Which means that when somebody sends an iMessage, they will select to connect an audio file that’s saved on their system. On this case, the recipient is requested to determine whether or not to obtain each the message and the voice recording.
“In response to receiving a message, every consumer’s voice mannequin is acquired,” the patent reads. “Primarily based on the voice mannequin, a voice output akin to the acquired message is offered.”
In line with the patent, the iPhone in query builds a Siri-like profile of the sender’s voice and simulates it when studying that message and future messages it receives from that sender. You may also ship with , so contacts could be downloaded earlier than the message.
As an alternative of listening to Siri’s voice when studying messages, it turns into extra personalised when family and friends textual content one another. We have been additionally capable of hear the messages in a extra private manner, reminiscent of.
The inventors of this patent are Qiong Hello, Jiangchuan Li, and David A. Winarsky. Winarsky is Apple’s director of text-to-speech know-how, Li is his senior Siri software program engineer in Apple’s machine studying, and Hu beforehand labored on Siri on the firm.
As typical, patents do not essentially make it a actuality, however given Apple’s current work on synthetic intelligence and voice, it is attainable. For instance, in iOS 11, Apple switched Siri’s voice from counting on voice actor recordings to a text-to-speech mannequin utilizing machine studying.
In 2020, Apple acquired an organization referred to as Voysis, which was working to enhance pure language processing for digital assistants. They used WaveNet know-how launched by Google’s DeepMind program in 2016.
WaveNet is a “deep generative mannequin of uncooked audio waveforms” that can be utilized to generate speech that mimics the human voice.
Apple has additionally began utilizing synthetic intelligence as a substitute of people to relate sure audiobook genres. So this patent is completely within the realm of the potential for Apple units to finally study to learn messages with a human voice.