It is possible to convert an iMessage into a voice note, which may subsequently be played again using a bespoke voice produced by the recipient's phone using samples from the original sender's voice. According to Apple Insider, this is the technology that was recently detailed in an Apple patent application.
It's a new patent
Customers of Apple already have access to text message readings by Siri and the ability to transmit audio recordings via iMessage. The new patent does, however, include a method to having the device read the text message using the sender's voice as opposed to Siri utilizing a voice file.
According to the patent, a second electronic device receives the voice model specifically. A message from a user of a different electronic device would serve as an illustration.
The concept that served as the foundation for the invention was created by Qiong Hi, Jiangchuan Li, and David A. Winarsky. Hu is a former Apple employee who worked on Siri, and Li is a senior Siri software developer for machine learning at Apple. Director of text-to-speech technology is Winarsky.
How will it achieve the function?
An associated voice recording that stays on the device can be included in an iMessage sent by the user. In that situation, the recipient will have the choice of either getting the message or the audio recording.
A speech model of the impacted user is received after receiving the message, according to the statement. An audio output that corresponds to the message received is produced based on the speech model, according to the patent.
The recipient's iPhone would then utilize that profile to simulate the voice of the sender when reading that message and any subsequent messages from that sender. This would result in a Siri - like profile of the sender's voice being established. So that the recipient's contacts can download and test it before receiving messages, the voice simulation model may be made available independently.
It would give friends and family a more individualized experience when contacting one another rather than reading messages aloud in Siri's voice. In order to give it that special touch, Apple has worked very hard.
The theory is tenable given Apple's recent investments in artificial intelligence (AI) and speech recognition. For instance, Apple's text-to-speech algorithm for Siri in iOS 11 replaced voice acting records. It is based on machine learning.
Apple acquired the business Voysis in 2020, which was primarily devoted to enhancing the usage of natural language processing by virtual assistants. It made use of WaveNet, a technology created by Google's DeepMind division. WaveNets can be trained to produce speech that faithfully replicates any human voice using so-called "deep generative models of raw audio waveforms.
Apple has also begun employing AI narration in place of human narrators for some types of audiobooks. The patent indicates that there is a high likelihood that Apple devices will eventually be able to learn to identify spoken communication.