Korean Voice AI Startups innovating technology to cater to wider audiences – koreatechdesk


Voice synthesizes, modulation & recognition, digital speech synthesis, etc., are no more unique technology processes. The vast network of Artificial Intelligence (AI) technology covers multiple possibilities in Voice technology, making it easier to use voice to create numerous options every day. Though the first widespread use of Voice AI is said to have been done in 1961 by IBM with Shoebox, a digital speech recognition tool, today there are numerous Voice AI tech startups worldwide, offering an array of services.

In South Korea, the market of ‘smart speakers’ is only second to China’s largest market. Korea’s largest telecom company SK Telecom had introduced its first AI-virtual assistant for the Korean language in 2016- NUGU and KaKao Corp, Naver also had followed their version of AI voice assistance. Many Korean startups in the AI sector have introduced voice-based assistance technologies that can be used to cloning singing voices of dead artists, act as baby monitors, or even help find a perfect love match.

KoreaTechDesk checks out some of the interesting Voice-AI-driven startups and their technology.

Supertone – Cloning dead singing voices

This music technology startup is at the top of our list for its latest promise in 2021. Supertone is harnessing AI technology’s prowess to bring new songs in the voice of famous artists who have already passed away. Supertone’s technology is all about creating “a hyper-realistic and expressive voice not distinguishable from real humans” using AI. In January 2021, Supertone used its Singing Voice Synthesis (SVS) technology to “resurrect” the voice of South Korean folk superstar Kim Kwang-seok, who died in 1996. Kim Kwang-seok’s ‘cloned’ singing voice was presented on the Korean television show Competition of the Century: AI vs. Human.

The startup founded in 2020 wants to recreate iconic singing start Freddie Mercury’s voice as well. The company was also in the news at the beginning of 2021, for it got the backing of the company behind K-Pop megastars BTS. Big Hit Entertainment invested $3.6 million in Supertone. The startup says it is not limiting to the ‘cloning’ service and wants to help newer artists with their Singing Synthesis technology. Indeed, with AI-created voices, questions of copyright and ethics come to the forefront. Supertone’s team writes on its website that it will “never monetize any synthetic voice without the permission of the right holder.” It suggests the possibility of generating “virtual” celebrities.

Posicube – Giving voice to call centres, restaurants & stores

This Seoul-based startup provides a specialized voice recognition service powered by AI. Posicube has two services: Robi Reception that was developed to replace repetitive tasks from call centers, and Robi Store, mainly used in restaurants and beauty stores to help customers make a reservation.

Posicube
Posicube

Established in 2017 by CEO Oh Sung-jo, the company aims to expand its AI-powered voice recognition service to Western markets in 2021, with an English version. Posicube’s voice AI platforms have a quick response time of less than a second made possible using advanced data-processing technologies. The startup claims its services can understand customers’ voices more understand technical terms and resolve customer issues accurately

Humelo – Speech synthesis for creative pursuits

This is another popular startup in the K-Pop entertainment circle for its AI composition/writing technology to innovate the entertainment industry through speech synthesis and creative AI. Posicube develops technology that includes emotions in speech and text-to-speech (TTS). The company aims to create a natural-sounding voice that is similar to a human voice.

Humelo
Humelo

Humelo has partnered with one of the biggest K-pop entertainment companies SM Entertainment to incorporate its AI composition/writing technology. The startup’s TTP API’s real-time multilingual text-to-speech technology can be used for navigation, smart cars, AI speakers, smart IoT, etc.

HAII – Voice for comfort and love

Korean startup Posicube (Human-AI-Information) has developed a Therapeutic Companion Bot to aid older people and those with mental illness. Jinwoo Kim, the founder of HAII, is a professor at Yonsei University. Kim researched about Human-Computer-Interaction (HCI) and established a startup that focuses on developing digital therapeutics. The startup has four digital therapy technology to cure mental illnesses, such as dementia, depression, anxiety disorder, and ADHD, applying a research-evidenced AI interaction strategy.

The messenger-based interactive treatment is characterized by grafting interactive AI technology into a digital treatment. Posicube has also signed up an exclusive sales contract with Eisai Group, a leading organization in Korea that engages in research and development of prescription drugs to generic drugs. Kim wants to enter the global market with his digital voice therapy service by entering the US in 2021. It has also started a beta service of a voice-based dating app called Chemi, where the app matches your date using your voice compatibility

Deeply – Baby monitoring with voice

Korean startup Deeply has launched a baby monitoring solution based on sound analysis called ‘Waah’ recently. Deeply is a developer of Voice Analysis AI technology, capable of analysing non-verbal sounds and extracting information. With its product the startup aims to solve difficulties in childcare. Waah is an AI –based  crying-aware baby app that can interpret the cries and whimpers of a child and determine their needs.

READ  With shift toward merit-based pay, Japan's Hitachi to drop old ways - Reuters

Deeply developed the product after thorough research and analysis. The team of the startup recorded baby cries in the age group of 3 to 6 months. They collected 100,000 baby voice data from hospitals, postpartum care homes, and so on to determine what emotions and conditions the extracted voice represented, and then used Artificial Intelligence to find out the characteristics of each type of voice. As a result, the baby ‘s six different expressions of the hunger, sleepiness, inconvenience, trim, pain, temperature and humidity problems, were analysed. The current data has put the result accuracy at 90%. Waah consists of a device and a smartphone app.

Also Read,

What’s your thoughts?

+1

Upvote

+1

Clap

+1

Clap

+1

On fire

+1

Meh

+1

Clap

+1



READ SOURCE

LEAVE A REPLY

Please enter your comment!
Please enter your name here