WAVETWEAK
the future of music industry
Our mission is to bring state-of-the-art AI to the music production, unification and fairness to audio streaming services and decentralization to music distribution.
At the moment, our main research focus is put on development of optimal and state-of-the-art algorithms which for a given song can perform:
- Extraction of audio samples.
- Source separation into volatile tracks.
- Synthesizer parameter prediction for building virtual instruments.
- Reverse engineering VST plugins and finding their parameters
- Predictions of effect chains and their decompositions
Combined, these are sufficient to construct a program which given a song returns DAW-like project file. These functionalities will provide music producers with an incredible power in the area of sound design and discovery. We consider those functionalities to be essential utilities needed for creating unified music system.
We envision singular, decentralized system for music production and distribution. In the core of our system, we employ novel audio representation method. We provide set of audio processing functions, with support of automated GUI design. We deliver fully configurable DAW of new generation, powered by your personal AI assistant(think Github copilot for music production).
Along with that, we envision smart-contracts based system, for music distribution, where publishers and artists have full control over their earnings. This naturally comes with audio streaming utilities and one of our main goals will be to develop unification between different streaming services.
We find current state of music industry lamentable. Every aspect is governed by soulless corporations, that take advantages of artists and users. We believe that only establishment of the introduced system can change current order of things and make people regain control over their art and earnings.
We got to the point, where artists have to pay 11$ per month on SoundCloud to be able to pin a song on the top of their profile, while on Spotify artists earn 3,500$ for one million streams and artists must pay 20$ to DistroKid per year, because there is 30 different streaming services, that are just pathetic copies of the copies of the copies of each other.
Apart from audio related utilities, Wavetweak aims to provide universal extensive identity protocol, that will guarantee security, reliability and copyright control. It will play key role in learning user profile across different applications. This protocol will play an interesting role in the process of AI assistance during art creation. It will be used to provide information on actual human input to the piece of art. We believe that object of art could be defined as "a product of thought" and thought is something exclusively human. Given the speed of development of generative models, I fear that in the future we might come to the point where majority of "digital art" could be generated with GANs. Imagine, that 80% of your last saved songs are raw model outputs with vocals synthesized and lyrics generated with GPT3 and you can't tell if there is anything human in it. Identity protocol will allow us to avoid such situation.