The application of AI in music is becoming increasingly popular in our country. Some see artificial intelligence as an extension of the creative and performance aspects of music. However, others view virtual music as a danger, as the line between creation and imitation, art and non-art, remains quite blurred.
New advances
Recently, Ann – Vietnam's first officially recognized virtual singer – released her second musical product titled "Cry". Unlike her debut song "How Can I Say I Love You," which was a ballad last year, Ann's return this time shows a change in both musical style and image. "Cry" is written in the rock genre.
Her vocal style was adjusted to a baby voice. Ann's image in the "Cry" music video is striking and individualistic, a stark contrast to her delicate and gentle appearance in her first music video. Ann's "How Can I Say I Love You" music video faced a wave of criticism. Her appearance was criticized for being unmemorable, her expressions unnatural, her lip movements inconsistent, and her stiff demeanor in every scene. The song's content and melody were considered unremarkable amidst the vast array of songs on the market.
After a year and a half of listening to public feedback, the team made many upgrades and adjustments to Ann. Therefore, this comeback, from her vocals to the visual effects and expressions, is more natural and diverse. At the press conference for the MV launch, Ann even appeared as a real person to perform the song. Unlike the two virtual singers, My Chau and Dam San, who were introduced at the Ho Do Music Festival 2022, Ann's voice wasn't dubbed by a real, anonymous singer, but rather a combination of AI algorithms and real sounds to create the desired vocal tone. The MV "Cry" is considered a new step forward in music produced by AI technology.
Recently, some music videos by real singers have also incorporated AI as an experiment, bringing a fresh perspective to audiences. In July, singer Dan Truong attracted attention when he applied 100% artificial intelligence to the visuals of his music video "Em Oi Vi Dau". The Vietnamese countryside, images of hardworking farmers, and even Dan Truong's portrait while singing were all handled by AI with realistic movements. The support of this technology saved him and his team the trouble of going to a studio or filming on location.
Dan Truong shared: “The AI team used many of my images to create the most accurate representation of the character. On average, to create a 4-second clip, we needed to use 4 to 16 images. To complete this MV, over 600 images were used with various AI tools.” Because this was the first time Dan Truong's team experimented with AI, despite their efforts, the MV's visuals still had many limitations. In some scenes, the countryside scenes looked stiff and artificial. The AI version's lip-syncing and facial expressions didn't match his vocals. Although criticized, Dan Truong said that this initial experiment helped him and his team learn valuable lessons to create better products in the future.
Also using AI to create a majestic historical music video, "mobile phone filmmaking mogul" Pham Vinh Khuong stunned many with his music video "The Picture of Dai Viet," illustrating the song "Nam Quoc Son Ha" (composed by DTAP, performed by singers Phuong My Chi and Erik). Various AI tools were utilized to their fullest potential to create a five-minute music video praising Vietnam's landscapes, culture, and proud four-thousand-year history. Pham Vinh Khuong stated that thanks to AI simplifying many processes, the cost of producing the music video was only three million dong. In contrast, with conventional production methods (i.e., actors, setting up historical scenes, and filming on location from North to South), the minimum cost for a music video would be three billion dong.
Online, simply searching for "how to compose songs using AI" quickly yields numerous results. Users only need to follow a few technical steps on available apps or software to create a song. The most prominent figure in building the "AI songwriting" model is musician and engineer Nguyen Hoang Bao Dai.
A young man in his 20s created an AI model that can compose 10 Vietnamese songs in just one second. The mechanism is quite simple. Users only need to provide the machine with a few musical notes or a short melody and press a button; the AI will automatically complete the song, generating many unique, complete versions. Users can then choose their favorite song or continue to have the AI adjust it until they are satisfied. This AI model is so intelligent that even if the musician has no ideas or provides no suggestions, the AI can still compose music according to its own wishes. Furthermore, Bảo Đại's model is even ready to produce music videos based on lyrical suggestions.
Great potential, many concerns.
At the "Application of Artificial Intelligence in Contemporary Music Composition" workshop held in Ho Chi Minh City in June, musicians and experts acknowledged that although still in its infancy, the future explosion of AI applications in music is an inevitable trend because it helps people create music more intelligently, conveniently, and with greater variety.
According to musician Nong Xuan Hieu, the advantage of artificial intelligence is that it can compose complete musical pieces with minimal human intervention. It suggests many new creative directions for musicians. AI can automatically create music, create virtual instruments, analyze music, and assist in harmonization and arrangement. Not only limited to production, AI also participates in live performances, interacts with audiences, and collects and analyzes audience feedback from various online platforms and media channels. “AI tools will continue to develop, becoming smarter and easier to use, enabling everyone, even those with little musical knowledge, to compose and produce high-quality music. Understanding and properly utilizing AI technologies will help the music industry thrive, providing more creative opportunities and improving the listening experience,” musician Nong Xuan Hieu stated.
The virtual singer model alone is a lucrative market in countries with developed music industries like Japan, South Korea, and China. Many virtual singers have become idols for young people, bringing in huge profits for their creators. For example, Hatsune Miku and Apoki generate tens of millions of dollars annually for the Japanese entertainment industry. In China, Luo Tianyi is the most popular virtual singer with numerous solo performances. In Vietnam, although the virtual singer model is still in its infancy, it has shown a keen ability to grasp the general trend.
BoBo Dang, the "creator" of the virtual singer Ann, revealed that his protégé is constantly being upgraded to be ready to appear on music stages and interact with fans like a real person. "One of our upcoming goals is for Ann to be able to interact like a real person and collaborate with other artists. We've had very promising negotiations with the managers of other singers. Certainly, in the very near future, Ann will be able to collaborate with real singers. Most show organizers have optimal support plans if Ann wants to perform on stage," he said.
Despite opening up vast horizons, virtual music still raises concerns among professionals. First and foremost is the issue of emotion and personal identity in musical products. Dr. Nguyen Bach Mai from Nguyen Tat Thanh University points out that AI still has limitations in terms of emotion, primarily relying on templates and algorithms to create music, thus often lacking authenticity and subtlety, and cannot compare to the emotions derived from real-life human experiences. Furthermore, AI's ability to empathize is limited and not as natural as humans. In particular, AI often struggles with improvisational and unconventional creation, something that musicians usually excel at. The lack of emotion, authenticity, and personal identity leads to AI-generated music being considered non-artistic.
Thanks to AI, even someone who doesn't know a single musical note can create a new song. This easily leads to the risk of complacency among composers, who rely entirely on AI. They may become complacent and believe they don't need much learning or training to proudly call themselves musicians. This is especially true when the creator of the "AI-written music" model asserts: "In the past, when listening to music composed by machines, people could immediately recognize the difference. But now, with the ever-increasing amount of data and increasingly intelligent algorithms, the difference between music written by AI and music written by humans is no longer clear."
The next concern is copyright. According to musician Nong Xuan Hieu, when AI learns from existing musical works to create new ones, the line between creation and copying becomes blurred. This raises questions about intellectual property rights. Sharing this view, Dr. Nguyen Bach Mai argues that determining intellectual property rights for AI-generated musical works is a complex issue. “Who will own the copyright—the musician, the AI developer, or both? The use of AI in music composition could be considered unethical if it completely replaces the role of the musician, diminishing the value of human creativity.”
Mr. Dinh Trung Can, Director of the Vietnam Music Copyright Protection Center, stated: Currently, copyright laws in Vietnam have not kept pace with the development of AI. Because the law lacks clear regulations on recognizing copyright and related rights for products created by AI, the Center is left out of the picture.
With the virtual singer model, there have been numerous counter-arguments from world- famous stars about their images and voices being stolen to create music products powered by artificial intelligence. When Ann first debuted, she also received a lot of negative feedback when her voice was said to be very similar to a famous local singer. Facing audience backlash, in her comeback with the music video "Cry," Ann's vocal style changed significantly to escape the accusations of imitation.
Due to ethical and copyright concerns, experts recommend that AI users should only view it as a tool to assist in the production and performance of music, not as a complete replacement for humans. A harmonious collaboration between musicians and AI will make music production more convenient, saving time, money, and effort. AI can suggest creative ideas, automate tasks such as arranging and mixing, and gauge market feedback. Musicians will have more time to focus on areas requiring intellectual investment, such as lyric writing and melody composition. Viewing AI as a tool that ultimately holds artists accountable for their work increases production efficiency without diminishing the artist's individuality and identity.
Source
Comment (0)