An efficient method to build music generative model by controlling both general and local note characteristics
It has been shown that since the rapid development of the entertainment industry, music generation has become a focused research topic. Numerous methods for creating music, or musical notes specifically have been announced, each with distinct characteristics and advantages. These methods usually con...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2023-10-01
|
| Series: | Journal of King Saud University: Computer and Information Sciences |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S1319157823003154 |
| _version_ | 1797629747093569536 |
|---|---|
| author | Thinh Do Quang Trang Hoang |
| author_facet | Thinh Do Quang Trang Hoang |
| author_sort | Thinh Do Quang |
| collection | DOAJ |
| description | It has been shown that since the rapid development of the entertainment industry, music generation has become a focused research topic. Numerous methods for creating music, or musical notes specifically have been announced, each with distinct characteristics and advantages. These methods usually concentrated on these two aspects: the overall harmony of the whole music score and the link between adjacent notes, which this research referred respectively as the general and local aspects. This study proposes a model with combined methods that is capable of deriving benefits from these both aspects, hence creating music with good quality in terms of both quantitative and qualitative evaluations. Various results based on those have been discussed and judged for efficient enhancing as well as for future development opportunities. The value of Average Pitch Interval (API) achieved a remarkable value of 1.43, along with the note range of 12.145; while on the subjective aspect, survey participants gave 6.81 score for the generated music, yet only about 70% of them can distinguish the generated from the genuine pieces of music. |
| first_indexed | 2024-03-11T10:58:20Z |
| format | Article |
| id | doaj.art-84ffa71e998a44ca8acf719954c80276 |
| institution | Directory Open Access Journal |
| issn | 1319-1578 |
| language | English |
| last_indexed | 2024-03-11T10:58:20Z |
| publishDate | 2023-10-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Journal of King Saud University: Computer and Information Sciences |
| spelling | doaj.art-84ffa71e998a44ca8acf719954c802762023-11-13T04:09:00ZengElsevierJournal of King Saud University: Computer and Information Sciences1319-15782023-10-01359101761An efficient method to build music generative model by controlling both general and local note characteristicsThinh Do Quang0Trang Hoang1Ho Chi Minh City University of Technology (HCMUT), Vietnam National University Ho Chi Minh City (VNU-HCM), Viet NamCorresponding author.; Ho Chi Minh City University of Technology (HCMUT), Vietnam National University Ho Chi Minh City (VNU-HCM), Viet NamIt has been shown that since the rapid development of the entertainment industry, music generation has become a focused research topic. Numerous methods for creating music, or musical notes specifically have been announced, each with distinct characteristics and advantages. These methods usually concentrated on these two aspects: the overall harmony of the whole music score and the link between adjacent notes, which this research referred respectively as the general and local aspects. This study proposes a model with combined methods that is capable of deriving benefits from these both aspects, hence creating music with good quality in terms of both quantitative and qualitative evaluations. Various results based on those have been discussed and judged for efficient enhancing as well as for future development opportunities. The value of Average Pitch Interval (API) achieved a remarkable value of 1.43, along with the note range of 12.145; while on the subjective aspect, survey participants gave 6.81 score for the generated music, yet only about 70% of them can distinguish the generated from the genuine pieces of music.http://www.sciencedirect.com/science/article/pii/S1319157823003154Artificial intelligenceGenerative adversarial networkLong short-term memoryMusic generationNote generation |
| spellingShingle | Thinh Do Quang Trang Hoang An efficient method to build music generative model by controlling both general and local note characteristics Journal of King Saud University: Computer and Information Sciences Artificial intelligence Generative adversarial network Long short-term memory Music generation Note generation |
| title | An efficient method to build music generative model by controlling both general and local note characteristics |
| title_full | An efficient method to build music generative model by controlling both general and local note characteristics |
| title_fullStr | An efficient method to build music generative model by controlling both general and local note characteristics |
| title_full_unstemmed | An efficient method to build music generative model by controlling both general and local note characteristics |
| title_short | An efficient method to build music generative model by controlling both general and local note characteristics |
| title_sort | efficient method to build music generative model by controlling both general and local note characteristics |
| topic | Artificial intelligence Generative adversarial network Long short-term memory Music generation Note generation |
| url | http://www.sciencedirect.com/science/article/pii/S1319157823003154 |
| work_keys_str_mv | AT thinhdoquang anefficientmethodtobuildmusicgenerativemodelbycontrollingbothgeneralandlocalnotecharacteristics AT tranghoang anefficientmethodtobuildmusicgenerativemodelbycontrollingbothgeneralandlocalnotecharacteristics AT thinhdoquang efficientmethodtobuildmusicgenerativemodelbycontrollingbothgeneralandlocalnotecharacteristics AT tranghoang efficientmethodtobuildmusicgenerativemodelbycontrollingbothgeneralandlocalnotecharacteristics |