If the player does not wear a full set of motion capture equipment, it will be a little difficult for their avatar to dance in the metaverse in the future. Secondly, the production of complex and aesthetic dance moves is also a threshold for digital human manufacturers. 36Kr learned that "Ta Ge" can provide virtual human dance and music solutions for game manufacturers, advertisers, etc.
Founded in April 2021, Tinge (Beijing Wege Technology Co., Ltd.) provides a Saas platform for dance studios is the main business of Tingge, based on the SaaS platform, Tingge has derived hyper-realistic dance digital human solutions, digital collections in the field of dance and music and other businesses.
"Ta Song" creates a hyper-realistic dance digital person "Dancing Ta si" for SPY Dance Studio
In terms of SaaS business, the dance studio uploads self-made dance courses to the Tingge Dance Cloud Platform for its students to use the Tingge APP to warm up, review and punch in after class. On the other hand, the dance studio students open the "Stepping Song APP", point the camera at themselves, learn to dance like a game, the mobile phone camera posture recognition algorithm automatically recognizes the students' dance movements, and the "teacher" disassembles the dance movements, recognizes the dance style, scores, and corrects errors.
This set of action posture recognition algorithm is based on a large number of dance movement accumulation, "the dance room teacher will put on our motion capture equipment, or go to the motion capture base for dance recording", according to the founder Li Di, ta ge has accumulated jazz dance action posture data content library, and jazz dance is one of the most widely audienceed dances in the female group.
In terms of technology, the dance movements are complex and change fast, and there are high requirements for the processing speed of visual algorithms. The gesture recognition algorithm of the mobile phone camera used by Tingge adopts the "motion capture + visual recognition" scheme to collect dance movements. The algorithm processes a frame for 20ms on a medium-sized configuration phone, which is 1/3 of the time consumed by google ML kits on the same hardware platform. Relying on this efficient mobile platform camera attitude recognition algorithm, Ta Ge can derive other gameplay to enhance the moat, such as dancing digital people.
According to Li Di, from the perspective of the application scenario of dancing digital people, digital people can dance in the music debut MV, and in the future, stepping songs can also provide dance solutions for game developers, advertisers, etc. "Making hyper-realistic digital human models is an increasingly competitive and labor-intensive business. In order to make a special feature and depth, Ta Ge focuses on the subdivision of making hyper-realistic digital people dance the most professional and fluent dance. Now, every consumer goods company is making its own hyper-realistic digital person, and if their digital person has been made by other companies, but wants to dance the most professional dance moves, they can adopt the solution provided by Tingge. ”
Around the dance, Li Di also hopes that the song can bring dancers continuous creator income. To this end, Tingge has developed a digital collection mini program, which produces photos of dancers, action videos, etc. into digital collections and distributes them in limited quantities. At the same time, Tingge believes that the digital collection platform should carry out more business innovation when it meets the regulatory requirements. It is reported that the dance movements of a head dancer who became popular on the short video platform were motion capture, and the action files generated were sold to advertisers and game developers as advertising assets and game assets through the digital collection platform of the song, and the relevant details will be disclosed in the near future.
"Tread Song" digital collection mini program
Li Di introduced to 36Kr, "What Ta Ge does, in one sentence, is to do the C-end application of dance and music in the metaverse field. In the short term, the path to be achieved is to link KA users such as dance studios through the SaaS platform, so as to obtain the C-end users behind the dance studios. In the future, Tingge will also expand pure C-end business, create a C-end-oriented dance online game platform, introduce rich game mechanisms, allow players to complete the jump, pk, record, and even participate in online disco through mobile phones, and eventually form a dance online community similar to 'K song for all'. ”
In terms of sales expansion, Through the establishment of a deep cooperative relationship with "SPY Dance Studio", Tingge has been linked to hundreds of head dance studios across the country. SPY Dance Studio is the standing director unit of the Street Dance Committee of China Dancers Association, which has a demonstration effect. According to Li Di, offline dance learning is a high-priced business. The typical portrait of a head dance room in a second- and third-tier city is that it has 10 to 20 chain stores, with an annual unit price of 4,000 to 8,000, and an annual revenue of about 30-50 million yuan. "The dance studio is prepaid and has good cash flow. If we serve KA customers such as the head dance studio in the big city, we will get 100,000 students in the next year, and each student will contribute an average of 200 yuan, and there will be 20 million yuan in revenue in one year. ”
The Tingge project began in 2020, and founder Li Di contributed most of the early code of Tingge's computer vision algorithm, gesture matching algorithm, SaaS platform, game client development, hyper-realistic digital human blockchain platform and other technologies. At present, the 1.0 version of the Tuoge APP and the Tuoge Digital Collection Mini Program have been tested among seed users and will be launched soon.
The founding team of Tage comes from Tencent, Facebook (Meta), Baidu, Ping An Technology, Warner Music, and Union Hospital, and there are many professional dancers and top musicians among the team members. Founder Li Di is a master's degree from the School of Artificial Intelligence of Beijing University of Posts and Telecommunications, and has done algorithm and engineering work in NLP and CV at ByteDance and Tencent; the person in charge of CV algorithm is a master of computer vision from the Institute of Computing of the Chinese Academy of Sciences, a senior algorithm engineer in the intelligent vision department of Ping An Technology, mainly responsible for deepening the direction of computer vision; the person in charge of vr platform is an optical module engineer of the University of Rochester, the former Apple (Silicon Valley) company, and the former Meta Oculus engineer. He was responsible for the supply chain development and manufacturing of optical components and modules for VR products such as Oculus Quest, and was responsible for the integration of VR hardware and offline dancing machine hardware solutions for the Tingge project; Lizzy, head of music products, was the first female DJ in Asia.