While there has been an immense anticipation for the next iPhone, Apple is planning even to uplift Siri with a human touch to make the overall user experience much better than ever before. The new Siri voice will be a part of iOS 11 that will debut on the new iPhone range, including the highly rumoured iPhone 8.
Apple has released a white paper that reveals a deep learning integration within Siri's unique voice. The new voice is more natural and smoother than what you hear on your iOS 10 device. "The quality of the new TTS (text-to-speech) system is superior to the previous Siri system," Apple's Siri team writes in the paper.
The Siri team has used a deep MDN (mixture density network) that combines conventional deep neural networks with a Gaussian mixture model. This fusion enhances the voice output with human-like speech features such as formants and voiced and unvoiced speech sound transitions. The new engine generates high-quality speech into multiple audio segments -- as many as one to two million units -- and trains the unified target that resulted in through the mix with natural and lively prosody traits. Apple has also added a higher sampling rate of 48kHz over the existing 22kHz and accomplished the entire process with better audio compression.
Notably, the fresh TTS system has several optimisations to suit mobile devices such as the new iPhone models or even the older ones that will be compatible with iOS 11. These optimisations are specific to the runtime performance for speed, memory usage and footprint.
New talent behind the scenes
Not just bringing some tweaks using deep learning algorithms, but Apple has also hired a new female voice talent to enhance Siri your iOS device. "We evaluated hundreds of candidates before choosing the best one. Then, we recorded over 20 hours of speech and built a new TTS voice using the new deep learning based TTS technology," the Siri team states in the white paper.
Also read: Samsung set to answer Apple's HomePod with its smart speaker
The new voice talent is aimed to improve naturalness, personality and expressivity of Siri's new voice. Apple has provided a few samples to show the difference in reality.
Advanced translation support
Apple's Siri engineers are additionally working hard to bring as many international languages to the digital assistant as they can. The iOS 11 update is expected to enable you to translate phrases to languages such as Spanish, French, Italian and German. These translation changes are presently available in beta for users with American English.
With all the speech and translation related changes, Apple is indeed intended to give a competitive edge to Siri. These upgrades will help in winning the race against Amazon's Alexa, Google Assistant, Samsung's Bixby Voice and Microsoft's Cortana.
