Making a slight mistake during live music performance can easily be spotted by an astute listener, even if the performance is an improvisation or an unfamiliar piece. An example might be a highly dissonant chord played by mistake in a classical-era sonata, or a sudden off-key note in a recurring motif. The problem of identifying and correcting such errors can be approached with artificial intelligence – if a trained human can easily do it, maybe a computer can be trained to spot the errors quickly and just as accurately. The ability to identify and auto-correct errors in real-time would be not only extremely useful to performing musicians, but also a valuable asset for producers, allowing much fewer overdubs and re-recording of takes due to small imperfections. This paper examines state-of-the-art solutions to related problems and explores novel solutions for music error detection and correction, focusing on their real-time applicability.

Making a slight mistake during live music performance can easily be spotted by an astute listener, even if the performance is an improvisation or an unfamiliar piece. An example might be a highly dissonant chord played by mistake in a classical-era sonata, or a sudden off-key note in a recurring motif. The problem of identifying and correcting such errors can be approached with artificial intelligence -- if a trained human can easily do it, maybe a computer can be trained to spot the errors quickly and just as accurately. The ability to identify and auto-correct errors in real-time would be not only extremely useful to performing musicians, but also a valuable asset for producers, allowing much fewer overdubs and re-recording of takes due to small imperfections. This paper examines state-of-the-art solutions to related problems and explores novel solutions for music error detection and correction, focusing on their real-time applicability. The explored approaches consider error detection through music context and theory, as well as supervised learning models with no predefined musical information or rules, trained on appropriate datasets. Focusing purely on correcting musical errors, the presented solutions operate on a high-level representation of the audio (MIDI) instead of the raw audio domain, taking input from an electronic instrument (MIDI keyboard/piano) and altering it when needed before it is sent to the sampler. This work proposes multiple general recurrent neural network designs for real-time error correction and performance aid for MIDI instruments, discusses the results, limitations, and possible future improvements. It also emphasizes on making the research results easily accessible to the end user - music enthusiasts, producers and performers -- by using the latest artificial intelligence platforms and tools.

Many of the recent approaches to polyphonic piano note onset transcription require training a machine learning model on a large piano database. However, such approaches are limited by dataset availability; additional training data is difficult to produce, and proposed systems often perform poorly on novel recording conditions. We propose a method to quickly synthesize arbitrary quantities of training data, avoiding the need for curating large datasets. Various aspects of piano note dynamics - including nonlinearity of note signatures with velocity, different articulations, temporal clustering of onsets, and nonlinear note partial interference - are modeled to match the characteristics of real pianos. Our method also avoids the disentanglement problem, a recently noted issue affecting machine-learning based approaches. We train a feed-forward neural network with two hidden layers on our generated training data and achieve both good transcription performance on the large MAPS piano dataset and excellent generalization qualities.