Artificial intelligence is also generating music
In previous articles on the blog, I shared with you updates on how artificial intelligence is changing our professional lives. I discussed LLM (Large Language Models) like ChatGPT, creating images using prompts with MidJourney or DALL-E, and generating videos from text using Sora AI. Now it’s time for… music!
In April 2024, the internet buzzed about new tools for creating music, such as Suno AI and especially Udio.
Suno officially operates in Cambridge, USA, and on March 21, the company publicly released the third version of its AI tool. Udio is a bit more mysterious, with its beta version being released on April 10, 2024. On both platforms, you can already listen to songs generated by internet users using artificial intelligence: udio.com and suno.com. In both cases, the services are available for free.
Udio
How does Udio work in the beta version? It’s quite simple, as you just enter the song or music you want to create. The Udio website prompts: “A song about…” encouraging “creation.” Then you can specify the genre of music, such as rock, folk, or pop, and add musical instruments to be used. You can request the creation of lyrics or add your own. For better results, Udio suggests providing your song lyrics, not longer than 250-300 characters (for now).
In three simple steps, from text to music. In the free version, you can generate 30 seconds of music. After a minute, two music files are ready to listen to. Then, using the Extend command, you can extend the song by another 30 seconds and finally, step by step, generate a song lasting several minutes.
Suno AI
Similarly, in Suno AI, the key feature is the ability to generate music based on a text prompt, simple or more elaborate. By providing a description of the desired genre or theme, Suno’s advanced algorithms can quickly produce a complete musical composition. In both cases, you cannot specify a specific artist as a reference for the created music.
I’ve tested AI music generation in various configurations, and I must admit that already, you can hear surprisingly high-quality musical productions. Testing both Suno AI and Udio, I can say that Suno seems easier to use for generating longer tracks and is better in terms of musical and vocal composition and maintaining coherence in each track. However, it’s definitely noticeable that the overall sound quality of Udio is better, despite the fact that it’s still in beta. The sound from Udio sounds less distorted compared to Suno. This is particularly noticeable in generated classical music, which on Suno may sound like a low-quality mp3 file, while Udio sounds almost like CD quality.
On Spotify, it would already be difficult to recognize which song was created by an artist and which by the mentioned AI services.
Noisee AI – genetsting video clips
What other tools can be used in this music fun? For example, Noisee creates animations for video clips. The company behind it advertises with the slogan “Seeing the unseen,” and indeed, in simple steps after clicking Imagine and providing a few prompts, you can quickly generate a video clip for the given music (link to the Suno file or uploaded music file). The entire process of generating the clip takes less than 10 minutes. While it’s not at the level of Sora AI, it looks good. . Link to Noisee AI.
Fascination and Fear
It’s fascinating and terrifying at the same time. I understand artists and musicians creating musical works for commercial purposes, for music banks, stock music, and even creators of film music or game music. And this is just the beginning of this musical revolution. According to some opinions, tools like Udio and Suno AI could completely change the stock music market within a year or two, eliminating people working in this profession. This is the pessimistic view.
The Future of AI Music
Those who know me know that while I am an enthusiast of new technologies, I also remain a realist and contemplate the negative aspects of such solutions. I intentionally don’t use the word created but generated, because as with images and videos, Suno and Udio had to learn based on songs created by humans and fed into the system by the founders of these ventures.
And again, there’s the question of copyright for music creators whether it has been violated. Certainly, legal issues will need to be clarified as soon as possible because they are not clear and currently do not protect creators and their musical works. IT companies have entered the territory of powerful record labels, and therefore, I think this will happen much faster.
I leave you with this update from the world of AI and pose open questions on how to leverage this in marketing and tourism.
