Kena.AI launches Artificial-Intelligence based piano music transcription.
[San Francisco, California] 03/11/2020 – Kena.ai is excited to launch its Artificial-Intelligence based piano music transcriber which listens into any freestyle, improv melody and automatically transcribes to sheet music. Unlike competitors who transcribe the piano music into a single voice, Kena performs 2-voice transcription which is more accurate. Kena also solves for left and right-hand annotations for clef changes and playability.
Music transcription is extremely hard given that music as an expressive form has layers and layers of mixed frequencies produced from playing multiple keys or notes simultaneously. The ability to separate these frequencies and determining what keys were played at the same time is challenging, especially when the sound lingers on even when the next set of keys are played.
In addition to the complexity of the music, the ability to transcribe a phone recording in a non-studio setup adds echo and noise which needs to be addressed prior to transcription. Kena uses AI to eliminate noise and amplify signals through Machine Learning based music mastering and then enables transcription. This way even the final quality of recorded music is higher.
“Accurate music transcription is the holy grail of the industry. Especially when you are transcribing music to sheet music that is readable and playable by humans. While our AI-based transcription is extremely accurate in capturing all dimensions and dynamics of complex musical renditions, we still have a lot to improve on the readability of the sheet music similar to how a human would transcribe though. That said, our transcriptions are best in class and beat most of the competitors among dimensions of rhythm, clarity, dynamics, multiple voices, and clef changes.” states CEO and Founder Preetham Vishwanatha
Here is a sample transcription with full articulations, embellishments and dynamics for Fur Elise by Ludwig Van Beethoven
The AI-based Piano transcriber is free as part of the Kena Opus platform and can be downloaded from the Apple app store and Google play store. Kena Opus enables real-time artificial intelligence feedback across vocals and 12 instruments. Musicians use it for real-time artificial intelligence feedback for practice sessions. As an audience, you snack on delightful musical expressions published by global musicians. Kena Opus is available for download for free!
Kena is creating a music neural engine based on deep neural networks and musical acoustics to enable a “music expression graph” based on the tonality, compositional architecture and the feel of the music. The idea is to use this expression graph to teach music and to compose AI-generated music closer to human composers.
For further press inquiries contact: