Subword tokenization has been widely successful in text-based natural language processing (NLP) tasks with Transformer-based models. As Transformer models become increasingly popular in symbolic music-related studies, it is imperative to investigate the efficacy of subword tokenization in the symbolic music domain. In this paper, we explore subword tokenization techniques, such as byte-pair encoding (BPE), in symbolic music generation and its impact on the overall structure of generated songs. Our experiments are based on three types of MIDI datasets: single track-melody only, multi-track with a single instrument, and multi-track and multi-instrument. We apply subword tokenization on post-musical tokenization schemes and find that it enables the generation of longer songs at the same time and improves the overall structure of the generated music in terms of objective metrics like structure indicator (SI), Pitch Class Entropy, etc. We also compare two subword tokenization methods, BPE and Unigram, and observe that both methods lead to consistent improvements. Our study suggests that subword tokenization is a promising technique for symbolic music generation and may have broader implications for music composition, particularly in cases involving complex data such as multi-track songs.
Intelligent music generation, one of the most popular subfields of computer creativity, can lower the creative threshold for non-specialists and increase the efficiency of music creation. In the last five years, the quality of algorithm-based automatic music generation has increased significantly, motivated by the use of modern generative algorithms to learn the patterns implicit within a piece of music based on rule constraints or a musical corpus, thus generating music samples in various styles. Some of the available literature reviews lack a systematic benchmark of generative models and are traditional and conservative in their perspective, resulting in a vision of the future development of the field that is not deeply integrated with the current rapid scientific progress. In this paper, we conduct a comprehensive survey and analysis of recent intelligent music generation techniques,provide a critical discussion, explicitly identify their respective characteristics, and present them in a general table. We first introduce how music as a stream of information is encoded and the relevant datasets, then compare different types of generation algorithms, summarize their strengths and weaknesses, and discuss existing methods for evaluation. Finally, the development of artificial intelligence in composition is studied, especially by comparing the different characteristics of music generation techniques in the East and West and analyzing the development prospects in this field.
In this work, we investigate the personalization of text-to-music diffusion models in a few-shot setting. Motivated by recent advances in the computer vision domain, we are the first to explore the combination of pre-trained text-to-audio diffusers with two established personalization methods. We experiment with the effect of audio-specific data augmentation on the overall system performance and assess different training strategies. For evaluation, we construct a novel dataset with prompts and music clips. We consider both embedding-based and music-specific metrics for quantitative evaluation, as well as a user study for qualitative evaluation. Our analysis shows that similarity metrics are in accordance with user preferences and that current personalization approaches tend to learn rhythmic music constructs more easily than melody. The code, dataset, and example material of this study are open to the research community.
WikiMT++ is an expanded and refined version of WikiMusicText (WikiMT), featuring 1010 curated lead sheets in ABC notation. To expand application scenarios of WikiMT, we add both objective (album, lyrics, video) and subjective emotion (12 emotion adjectives) and emo\_4q (Russell 4Q) attributes, enhancing its usability for music information retrieval, conditional music generation, automatic composition, and emotion classification, etc. Additionally, CLaMP is implemented to correct the attributes inherited from WikiMT to reduce errors introduced during original data collection and enhance the accuracy and completeness of our dataset.
Fast and user-controllable music generation could enable novel ways of composing or performing music. However, state-of-the-art music generation systems require large amounts of data and computational resources for training, and are slow at inference. This makes them impractical for real-time interactive use. In this work, we introduce Musika, a music generation system that can be trained on hundreds of hours of music using a single consumer GPU, and that allows for much faster than real-time generation of music of arbitrary length on a consumer CPU. We achieve this by first learning a compact invertible representation of spectrogram magnitudes and phases with adversarial autoencoders, then training a Generative Adversarial Network (GAN) on this representation for a particular music domain. A latent coordinate system enables generating arbitrarily long sequences of excerpts in parallel, while a global context vector allows the music to remain stylistically coherent through time. We perform quantitative evaluations to assess the quality of the generated samples and showcase options for user control in piano and techno music generation. We release the source code and pretrained autoencoder weights at github.com/marcoppasini/musika, such that a GAN can be trained on a new music domain with a single GPU in a matter of hours.
Autoregressive models based on Transformers have become the prevailing approach for generating music compositions that exhibit comprehensive musical structure. These models are typically trained by minimizing the negative log-likelihood (NLL) of the observed sequence in an autoregressive manner. However, when generating long sequences, the quality of samples from these models tends to significantly deteriorate due to exposure bias. To address this issue, we leverage classifiers trained to differentiate between real and sampled sequences to identify these failures. This observation motivates our exploration of adversarial losses as a complement to the NLL objective. We employ a pre-trained Span-BERT model as the discriminator in the Generative Adversarial Network (GAN) framework, which enhances training stability in our experiments. To optimize discrete sequences within the GAN framework, we utilize the Gumbel-Softmax trick to obtain a differentiable approximation of the sampling process. Additionally, we partition the sequences into smaller chunks to ensure that memory constraints are met. Through human evaluations and the introduction of a novel discriminative metric, we demonstrate that our approach outperforms a baseline model trained solely on likelihood maximization.
Music is essential when editing videos, but selecting music manually is difficult and time-consuming. Thus, we seek to automatically generate background music tracks given video input. This is a challenging task since it requires plenty of paired videos and music to learn their correspondence. Unfortunately, there exist no such datasets. To close this gap, we introduce a dataset, benchmark model, and evaluation metric for video background music generation. We introduce SymMV, a video and symbolic music dataset, along with chord, rhythm, melody, and accompaniment annotations. To the best of our knowledge, it is the first video-music dataset with high-quality symbolic music and detailed annotations. We also propose a benchmark video background music generation framework named V-MusProd, which utilizes music priors of chords, melody, and accompaniment along with video-music relations of semantic, color, and motion features. To address the lack of objective metrics for video-music correspondence, we propose a retrieval-based metric VMCP built upon a powerful video-music representation learning model. Experiments show that with our dataset, V-MusProd outperforms the state-of-the-art method in both music quality and correspondence with videos. We believe our dataset, benchmark model, and evaluation metric will boost the development of video background music generation.
The launch of ChatGPT has garnered global attention, marking a significant milestone in the field of Generative Artificial Intelligence. While Generative AI has been in effect for the past decade, the introduction of ChatGPT has ignited a new wave of research and innovation in the AI domain. This surge in interest has led to the development and release of numerous cutting-edge tools, such as Bard, Stable Diffusion, DALL-E, Make-A-Video, Runway ML, and Jukebox, among others. These tools exhibit remarkable capabilities, encompassing tasks ranging from text generation and music composition, image creation, video production, code generation, and even scientific work. They are built upon various state-of-the-art models, including Stable Diffusion, transformer models like GPT-3 (recent GPT-4), variational autoencoders, and generative adversarial networks. This advancement in Generative AI presents a wealth of exciting opportunities and, simultaneously, unprecedented challenges. Throughout this paper, we have explored these state-of-the-art models, the diverse array of tasks they can accomplish, the challenges they pose, and the promising future of Generative Artificial Intelligence.
Dance-driven music generation aims to generate musical pieces conditioned on dance videos. Previous works focus on monophonic or raw audio generation, while the multiinstruments scenario is under-explored. The challenges of the dance-driven multi-instruments music (MIDI) generation are two-fold: 1) no publicly available multi-instruments MIDI and video paired dataset and 2) the weak correlation between music and video. To tackle these challenges, we build the first multi-instruments MIDI and dance paired dataset (D2MIDI). Based on our proposed dataset, we introduce a multi-instruments MIDI generation framework (Dance2MIDI) conditioned on dance video. Specifically, 1) to model the correlation between music and dance, we encode the dance motion using the GCN, and 2) to generate harmonious and coherent music, we employ Transformer to decode the MIDI sequence. We evaluate the generated music of our framework trained on D2MIDI dataset and demonstrate that our method outperforms existing methods. The data and code are available on https://github.com/Dance2MIDI/Dance2MIDI
Lead sheets have become commonplace in generative music research, being used as an initial compressed representation for downstream tasks like multitrack music generation and automatic arrangement. Despite this, researchers have often fallen back on deterministic reduction methods (such as the skyline algorithm) to generate lead sheets when seeking paired lead sheets and full scores, with little attention being paid toward the quality of the lead sheets themselves and how they accurately reflect their orchestrated counterparts. To address these issues, we propose the problem of conditional lead sheet generation (i.e. generating a lead sheet given its full score version), and show that this task can be formulated as an unsupervised music compression task, where the lead sheet represents a compressed latent version of the score. We introduce a novel model, called Lead-AE, that models the lead sheets as a discrete subselection of the original sequence, using a differentiable top-k operator to allow for controllable local sparsity constraints. Across both automatic proxy tasks and direct human evaluations, we find that our method improves upon the established deterministic baseline and produces coherent reductions of large multitrack scores.