OpenAI is preparing to revolutionize the creation of its OpenAI music generator tool. According to a report from The Information, the company is working on a new tool that makes original music from simple text and audio prompts. People could add full soundtracks to their videos or instrumentals to vocal tracks that are already there with this technology.
This project is a big step forward from what OpenAI has done with audio in the past. It is based on older models like MuseNet and Jukebox. The new OpenAI Music generator tool should give users a lot more power over how a song sounds, feels, and moves.
Some features could be the ability to make duets or quintets with more than one vocal track and AI-assisted mixing and mastering. This lets independent musicians and content creators make things that sound like they were made in a recording studio.
According to reports, the development process includes a unique partnership with students from the renowned Juilliard School. These students are helping to improve the training data by adding notes to musical scores.
The goal of this partnership is to teach the AI how to understand complicated musical patterns and show feelings. The goal is to get results that sound more like people and make musical sense.
The new OpenAI music generator tool has to get past many technical problems. It’s difficult to make music as raw audio instead of MIDI files, which are easier to work with. To get the full depth of music, raw audio needs to model very long sequences.
Jukebox introduced OpenAI’s method, which compresses audio into a separate space before generating it. The table below compares OpenAI’s new tool to its older versions and a major competitor.
Effects on the industry and problems with the OpenAI music generator tool
The AI music market is now even more competitive because of OpenAI’s entry. This puts the company on the same level as well-known companies like Google and new startups like Suno. This contest makes creators come up with new ideas quickly, which gives them more ways to show off their musical skills.
These tools are very useful for both artists and people who sell things. They can quickly write music for videos, podcasts, or posts on social media. This makes it easier for everyone to get good music by cutting down on the time and money needed to make it.
There are complicated questions that arise from the new capabilities of the OpenAI music generator tool. The music business is already dealing with issues of copyright and ethics. German collecting society GEMA is suing OpenAI for using copyrighted song lyrics to train its models without permission.
This is one of the first cases like this in the world, and it will likely change how AI companies handle copyrighted material in the future. In a similar case, GEMA has also sued the AI music company Suno.
The rules are changing. Other suggestions include systems that automatically check AI-generated derivative works for approval using a machine learning model that takes into account the content owner’s preferences. Then, a digital watermark could be put on approved works, and an authorization server would make sure they were used correctly. It’s still not clear what the law says about music made by AI, which could be a problem for people who make it.
People are talking about it, but it’s not clear when everyone will be able to use the OpenAI music generator tool. It doesn’t have a set release date yet, and it’s not clear if it will be its own product or part of other platforms like ChatGPT or the video app Sora. OpenAI’s move is definitely a new step in AI-assisted creativity that could positively transform the way music is created.
