Search

In this paper, a new method that computes the aesthetics of a melody fragment is proposed, starting from dissonances. While music generated with artificial intelligence applications may be produced considerably more quickly than human-composed music, it has the drawback of not being appreciated like a human composition, being many times perceived by humans as artificial. For achieving supervised machine learning objectives of improving the quality of the great number of generated melodies, it is a challenge to ask humans to grade them. Therefore, it would be preferable if the aesthetics of artificial-intelligence-generated music is calculated by an algorithm. The proposed method in this paper is based on a neural network and a mathematical formula, which has been developed with the help of a study in which 108 students evaluated the aesthetics of several melodies. For evaluation, numerical values generated by this method were compared with ratings provided by human listeners from a second study in which 30 students participated and scores were generated by an existing different method developed by psychologists and three other methods developed by musicians. Our method achieved a Pearson correlation of 0.49 with human aesthetic scores, which is a much better result than other methods obtained. Additionally, our method made a distinction between human-composed melodies and artificial-intelligence-generated scores in the same way that human listeners did. [ABSTRACT FROM AUTHOR]

Featured Application: This paper presents an improvement of the Jukebox system for music generation, which significantly reduces the inference time. The Jukebox model can generate high-diversity music within a single system, which is achieved by using a hierarchical VQ-VAE architecture to compress audio in a discrete space at different compression levels. Even though the results are impressive, the inference stage is tremendously slow. To address this issue, we propose a Fast Jukebox, which uses different knowledge distillation strategies to reduce the number of parameters of the prior model for compressed space. Since the Jukebox has shown highly diverse audio generation capabilities, we used a simple compilation of songs for experimental purposes. Evaluation results obtained using emotional valence show that the proposed approach achieved a tendency towards actively pleasant, thus reducing inference time for all VQ-VAE levels without compromising quality. [ABSTRACT FROM AUTHOR]

Language models based on deep learning showed promising results for artistic generation purposes, including musical generation. However, the evaluation of symbolic musical generation models is mostly based on low-level mathematical metrics (e.g., the result of the loss function) due to the inherent difficulty in measuring the musical quality of a given performance due to the subjective nature of music. This work sought to measure and evaluate musical excerpts generated by deep learning models from a human perspective, limited to the scope of classical piano music generation. In this assessment, a population of 117 people performed blind tests with musical excerpts of human composition and musical excerpts generated using artificial intelligence models, including the models PerformanceRNN, Music Transformer, MuseNet and a custom model based on GRUs. The experiments demonstrated that musical excerpts generated using models based on the Transformer neural network architecture obtained the greatest receptivity within the tested population, surpassing the results of human compositions. In addition, the experiments also demonstrated that people with greater musical sensitivity and musical experience were more able to identify the compositional origin of the excerpts heard. [ABSTRACT FROM AUTHOR]

The development of computational artifacts to study cross-modal associations has been a growing research topic, as they allow new degrees of abstraction. In this context, we propose a novel approach to the computational exploration of relationships between music and abstract images, grounded by findings from cognitive sciences (emotion and perception). Due to the problem's high-level nature, we rely on evolutionary programming techniques to evolve this audio–visual dialogue. To articulate the complexity of the problem, we develop a framework with four modules: (i) vocabulary set, (ii) music generator, (iii) image generator, and (iv) evolutionary engine. We test our approach by evolving a given music set to a corresponding set of images, steered by the expression of four emotions (angry, calm, happy, sad). Then, we perform preliminary user tests to evaluate if the user's perception is consistent with the system's expression. Results suggest an agreement between the user's emotional perception of the music–image pairs and the system outcomes, favoring the integration of cognitive science knowledge. We also discuss the benefit of employing evolutionary strategies, such as genetic programming on multi-modal problems of a creative nature. Overall, this research contributes to a better understanding of the foundations of auditory–visual associations mediated by emotions and perception. [ABSTRACT FROM AUTHOR]

With the rapid development of artificial intelligence, the application of this new technology to music generation has attracted more attention and achieved gratifying results. This study proposes a method for combining the transformer deep-learning model with generative adversarial networks (GANs) to explore a more competitive music generation algorithm. The idea of text generation in natural language processing (NLP) was used for reference, and a unique loss function was designed for the model. The training process solves the problem of a nondifferentiable gradient in generating music. Compared with the problem that LSTM cannot deal with long sequence music, the model based on transformer and GANs can extract the relationship in the notes of long sequence music samples and learn the rules of music composition well. At the same time, the optimized transformer and GANs model has obvious advantages in the complexity of the system and the accuracy of generating notes. [ABSTRACT FROM AUTHOR]

Recent studies demonstrate that algorithmic music attracted global attention not only because of its amusement but also its considerable potential in the industry. Thus, the yield increased academic numbers spinning around on topics of algorithm music generation. The balance between mathematical logic and aesthetic value is important in music generation. To maintain this balance, we propose a research method based on a three-dimensional temporal convolutional attention neural network. This method uses a self-collected traditional Chinese pentatonic symbolic music dataset. It combines clustering algorithms and deep learning-related algorithms to construct a three-dimensional sequential convolutional generation model 3D-SCN, a three-dimensional temporal convolutional attention model BoYaTCN. We trained both of them to generate traditional Chinese pentatonic scale music that considers both overall temporal creativity and local musical semantics. Then, we conducted quantitative and qualitative evaluations of the generated music. The experiment demonstrates that BoYaTCN achieves the best results, with a prediction accuracy of 99.12%, followed by 3D-SCN with a prediction accuracy of 99.04%. We have proven that the proposed model can generate folk music with a beautiful melody, harmonious coherence, and distinctive traditional Chinese pentatonic features, and it also conforms to certain musical grammatical characteristics. [ABSTRACT FROM AUTHOR]

Deep learning has given AI-based methods for music creation a boost by over the past years. An important challenge in this field is to balance user control and autonomy in music generation systems. In this work, we present BassNet, a deep learning model for generating bass guitar tracks based on musical source material. An innovative aspect of our work is that the model is trained to learn a temporally stable two-dimensional latent space variable that offers interactive user control. We empirically show that the model can disentangle bass patterns that require sensitivity to harmony, instrument timbre, and rhythm. An ablation study reveals that this capability is because of the temporal stability constraint on latent space trajectories during training. We also demonstrate that models that are trained on pop/rock music learn a latent space that offers control over the diatonic characteristics of the output, among other things. Lastly, we present and discuss generated bass tracks for three different music fragments. The work that is presented here is a step toward the integration of AI-based technology in the workflow of musical content creators. [ABSTRACT FROM AUTHOR]

Prior to the establishment of the Roman rite with its Gregorian chant, in the Iberian Peninsula and Southern France the Mozarabic rite, with its own tradition of chant, was dominant from the sixth until the eleventh century. Few of these chants are preserved in pitch readable notation and thousands exist only in manuscripts using adiastematic neumes which specify only melodic contour relations and not exact intervals. Though their precise melodies appear to be forever lost it is possible to use computational machine learning and statistical sequence generation methods to produce plausible realizations. Pieces from the León antiphoner, dating from the early tenth century, were encoded into templates then instantiated by sampling from a statistical model trained on pitch-readable Gregorian chants. A concert of ten Mozarabic chant realizations was performed at a music festival in the Netherlands. This study shows that it is possible to construct realizations for incomplete ancient cultural remnants using only partial information compiled into templates, combined with statistical models learned from extant pieces to fill the templates. [ABSTRACT FROM AUTHOR]

This paper describes the theory and implementation of a digital audio workstation plug-in for chord sequence generation. The plug-in is intended to encourage and inspire a composer of electronic dance music to explore loops through chord sequence pattern definition, position locking and generation into unlocked positions. A basic cyclic first-order statistical model is extended with latent diatonicity variables which permits sequences to depart from a specified key. Degrees of diatonicity of generated sequences can be explored and parameters for voicing the sequences can be manipulated. Feedback on the concepts, interface, and usability was given by a small focus group of musicians and music producers. [ABSTRACT FROM AUTHOR]

The strong relationship between music and health has helped prove that soft and peaceful classical music can significantly reduce people's stress; however, it is difficult to identify and collect examples of such music to build a library. Therefore, a system is required that can automatically generate similar classical music selections from a small amount of input music. Melody is the main element that reflects the rhythms and emotions of musical works; therefore, most automatic music generation research is based on melody. Given that melody varies frequently within musical bars, the latter are used as the basic units of composition. As such, there is a requirement for melody extraction techniques and bar-based encoding methods for automatic generation of bar-based music using melodies. This paper proposes a method that handles melody track extraction and bar encoding. First, the melody track is extracted using a pitch-based term frequency–inverse document frequency (TFIDF) algorithm and a feature-based filter. Subsequently, four specific features of the notes within a bar are encoded into a fixed-size matrix during bar encoding. We conduct experiments to determine the accuracy of track extraction based on verification data obtained with the TFIDF algorithm and the filter; an accuracy of 94.7% was calculated based on whether the extracted track was a melody track. The estimated value demonstrates that the proposed method can accurately extract melody tracks. This paper discusses methods for automatically extracting melody tracks from MIDI files and encoding based on bars. The possibility of generating music through deep learning neural networks is facilitated by the methods we examine within this work. To help the neural networks generate higher quality music, which is good for human health, the data preprocessing methods contained herein should be improved in future works. [ABSTRACT FROM AUTHOR]