{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/midinet-a-convolutional-generative","title":"MidiNet: A Convolutional Generative Adversarial Network for Symbolic-domain Music Generation","arxiv_id":"1703.10847","date":"2017-03-31","proceeding":null,"authors":["Li-Chia Yang","Szu-Yu Chou","Yi-Hsuan Yang"],"abstract":"Most existing neural network models for music generation use recurrent neural\nnetworks. However, the recent WaveNet model proposed by DeepMind shows that\nconvolutional neural networks (CNNs) can also generate realistic musical\nwaveforms in the audio domain. Following this light, we investigate using CNNs\nfor generating melody (a series of MIDI notes) one bar after another in the\nsymbolic domain. In addition to the generator, we use a discriminator to learn\nthe distributions of melodies, making it a generative adversarial network\n(GAN). Moreover, we propose a novel conditional mechanism to exploit available\nprior knowledge, so that the model can generate melodies either from scratch,\nby following a chord sequence, or by conditioning on the melody of previous\nbars (e.g. a priming melody), among other possibilities. The resulting model,\nnamed MidiNet, can be expanded to generate music with multiple MIDI channels\n(i.e. tracks). We conduct a user study to compare the melody of eight-bar long\ngenerated by MidiNet and by Google's MelodyRNN models, each time using the same\npriming melody. Result shows that MidiNet performs comparably with MelodyRNN\nmodels in being realistic and pleasant to listen to, yet MidiNet's melodies are\nreported to be much more interesting.","url_abs":"http://arxiv.org/abs/1703.10847v2","url_pdf":"http://arxiv.org/pdf/1703.10847v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"midinet-a-convolutional-generative","repo_url":"https://github.com/RichardYang40148/MidiNet","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null},{"paper_slug":"midinet-a-convolutional-generative","repo_url":"https://github.com/KarenMars/TUD_MMSR","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"midinet-a-convolutional-generative","repo_url":"https://github.com/annahung31/MIdiNet-by-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"midinet-a-convolutional-generative","repo_url":"https://github.com/dongmingli-Ben/MidiNet-by-pytorch-Debugged","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":null,"task_name":"Generative Adversarial Network"},{"task_slug":"music-generation","task_name":"Music Generation"}],"methods":[{"method_slug":"dilated-causal-convolution","method_name":"Dilated Causal Convolution"},{"method_slug":"mixture-of-logistic-distributions","method_name":"Mixture of Logistic Distributions"},{"method_slug":"wavenet","method_name":"WaveNet"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1703.10847","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}