LALAL.AI: An AI-Powered Stem Splitter for the Music Industry

With the rising prevalence of AI in today’s society, there is no shortage of interesting things that folks are developing the tech to be used for. One of the most interesting uses of AI technology in relation to the music industry is being developed by a company called LALAL.ai.

LALAL.AI is an AI-powered vocal remover and stem splitter that can separate instrumental tracks, vocal tracks, or even isolated singular instruments from a master audio recording. Despite being a relatively new technology, this is a powerful and impressive piece of tech that opens a world of possibilities to music industry professionals and musicians alike.

For this post, Matt Zutell of Charleston, SC recording studio Coast Records shares his experience with the software, what he discerns that it can be used for, as well as its effectiveness at its intended tasks. Read on to learn all about this exciting new technology.

How to Use LALAL.AI

Before we get started with possible use cases for this AI-powered vocal remover and stem splitter, I wanted to give a general overview of how to use LALAL.AI.

The app can currently be used both in the browser via their website and remotely via their API, and will soon also be available via VST plugin and mobile app.

To use LALAL.AI, follow these steps:

1. Open LALAL.AI in your browser.
2. Select what you would like to extract via the dropdown menu.
3. Click the “Select Files” button to upload your audio or video file.
4. Wait for the processing to complete, which usually takes a couple of seconds.
5. Click the play icon (‎▶) within the “Preview” section to check if you like the result.

After the initial processing, but prior to downloading, you can also adjust and fine-tune the intensity of audio processing by following these steps:

1. Click Back in the upper left corner of the window.
2. Select the desired intensity level at the top of the file upload section. The default is the “Normal” level, with options for “Mild” and “Aggressive” processing as well.
3. Click “Select Files” again, and pick the song you previously uploaded.
4. Playback the new vocal track preview. If you want to try another filter, click “Back” in the upper left corner of the window.

Once you are happy with the sound of the preview, you can complete the use of LALAL.AI by completing the following steps:

1. Click the “Process the Entire File” button.
2. Finally, click the download icon (⤓) to download the processed audio file.

Use Cases for LALAL.AI

There are many possible things that LALAL.AI can be used for, which we’ve broken down below based on Zutell’s personal use of the software in his audio production and engineering business, and some other ideas that came up during our conversation.

Instrumental / Acapella Versions of Songs

The single most effective way to use LALAL.AI is as a vocal remover, separating the instrumentals from the music entirely to create either an instrumental or acapella version of the track. From there, you can get creative with your uses.

“This thing is most effective at separating the vocals from the music,” Zutell explains. ” You can pull a vocal out and put an instrumental by itself and you basically can’t hear the vocals at all anymore, which is pretty amazing. In that regard there are a lot of uses for it.”

Isolating specific instruments

As Zutell explains above, the technology works best for isolating vocals from instrumentals, but can also be used to isolate specific instruments from the mix.

“It has settings where you can isolate the drums from the rest of the music, the bass from the rest of the music, the guitars, the synths, like the acoustic guitars. Some of those settings — it’s amazing how effective it is, but it’s not perfect yet,” Zutell says. “I only imagine that as AI gets better it will continue to get stronger and more clean.”

While this aspect of the software is still developing, it can be a very effective use for a musician who is trying to learn how to play a certain song, or may simply be curious about what the drummer is doing in a recording, for example.

“That’s another effective use, like if you were a musician trying to hear what your instrument’s part was doing by itself,” Zutell explains. “To at least be able to isolate that instrument and be able to hear it cleanly is a really cool feature. It really depends on the song itself and how the actual recording songs, and how glued together all the sounds are, but for more sparse recordings. Like if you had a more minimal song, where you can really hear each instrument cleanly, it’ll separate em better than if you had a really dense mix with a lot of aggressive stuff going on at the same time.”

Content Creation for Social Media

LALAL.AI can be a very effective piece of software for those in the content creation industry, specifically on Youtube or TikTok. This tool gives you the power to get very creative with your content ideas, creating unique instrumental or acapella versions of songs that can then be used in content creation.

“It creates so many more opportunities to use songs or get creative with popular things that are out there,” Zutell explains. “Like if you wanna take specific elements from it and do your own thing with it. I think the content creation thing is a big part.”

Creating Backing Tracks

Creating a backing track for a live performance was actually the first use case that drew Zutell to this software. This is especially useful for solo artists, or smaller groups who have more complex recordings than they are able to reproduce in the live setting.

“The reason I pulled it up was I was helping my brother do a song for a show he was doing for his acroyoga-trapeze thing,” Zutell says. “We were thinking about making a backing track to a song, because we did a performance of him and I, just on drums and guitar, but we were like ‘maybe we can make a track to support the extra stuff.’ So I went on there to try and pull out the bass, synthesizers and stuff.”

“So, you could use it to make like, live backing tracks, and have the different elements separated out. And then you can mix together the parts you needed from the song. You know, you could mute or not use the stuff that you were actually performing with, and then supplement the various things that you don’t have being played on stage.”

Karaoke Versions

The ai-powered vocal remover is also very effective as a karaoke maker, as it excels at separating vocals from instrumental versions of songs. This is useful for karaoke as often times the available karaoke versions of songs are very poorly mixed and separated, and don’t sound nearly as good as the actual song. LALAL.AI is one way to remedy that.

“A lot of times the Karaoke versions of songs are kinda lame sounding, and they sound kinda fake and purposely karaoke,” Zutell explains. So this would be a way to do it with the actual song itself, as the bed of music.

Sync Pitches and Licensing Opportunities

One opportunity for musicians to use LALAL.AI is for sync pitches and licensing opportunities, when you’d need an instrumental version of a song to be used in a TV show, commercial, or movie. Often times, artists will not have access to the original stems or recording files, so they would have to go through an expensive and time-consuming process in order to get that. LALAL.AI makes it easy to create these things.

“Let’s say you’re wanting to do a sync pitch, and have one of your songs put on a TV show, or a commercial or something, or a movie, and all you have is the master recording,” Zutell says. “you could create an instrumental version that has the vocals pulled out to put in the background of a commercial or a show or something. That’s super effective, saves you a ton of time and money and it gives you opportunity. Like if you didn’t have an instrumental, you can create one…it allows you opportunities to control your music.”

Remixes, Mashups, and Sampling

Another huge use case for LALAL.AI is for producers, DJs, or remix artists who are looking to extract certain elements from music to create their own unique content. This includes vocal samples, beat samples, and more.

“Let’s say you were a producer, or a remix artist, or a DJ or something, and you wanted just the vocals from a song to put on a beat or remix that you made, that’s an amazing thing,” Zutell says. “Especially if you don’t have access to all the stems from the actual producer, or the engineer which most people don’t. And even people with their own music don’t always have access to old stuff. You know, if it’s a song that you recorded 10 years ago, and you never made an instrumental, or you don’t have the actual stems from the song, there’s a chance that those files are long gone.”

Note that you will still need to go through the copyrighting and licensing process if you would like to publish your original creation using stems taken from another song that is not yours.

Extracting Voice from a Video or Movie

One final use case for LALAL.AI that we were able to come up with is to extract voice from a video or movie. This allows content creators to get quite inventive, as they can take a famous scene from a movie, that perhaps famously had music playing, and replace it with whatever music they would like.

Additionally, this can be used to extract vocal samples and voiceovers from videos or movies to be used in music production.

Conclusion

LALAL.AI is an impressive piece of software that points to the future of technological advances in the music industry. The ability to extract vocals from instrumentals alone is a step forward in terms of expanding creative possibilities for musicians and producers, and as the tech continues to advance it will only become more streamlined and effective — though it is already quite useful as it stands.

For pricing and more details, visit LALAL.AI.