Cloud Service for Audio and Music Analysis & Recognition


Focusing on using AI and machine learning for audio/music analysis, retrieval, and recognition.

Singing Voice Seperation  

Our SVS system can achieve state-of-the-art performance, and the extracted singing voices can be used for various purposes.


Music Rhythm Game  

Genre/ Mood Classification  

Audio Fingerprinting  

Query by

Beat Tracking  

Active Noise Cancellation  

We have been working on these areas for over 20 years, and we have won several prestigious international competitions of audio & music analysis. Our primary misson and vision is to provide brand new user experience of music and audio based on our technologies.

— CSO Roger Jang



Vocal/Instrument Sound Extration

We provide a novel vocal/instrument extraction system, which can identify individual instrument sound and human vocal, respectively. All these separated sound and vocal can be utilized as many usages, includes music education, karaoke, singing assessment, etc.


Music Game Sheet Generation

Combine with onset detection/beat tracking and other machine learning techniques, our music game sheet generation system can provide a rhythm-game music sheet for any given song in a customizable format. Just plug-in and create your own music game!


Music Database Management

Our music database management system can classify the given music clips by the pre-defined genre/mood labels. Use these information to build up your CRM or recommendation system easily.



ChordPool generates sheet music for guitar/ukulele by utilizing beat tracking and chord estimation results. ChordPool can let user choose their favorite song to play guitar/ukulele as they want to.

Interested in working with us?

Join Us!