Controlla XYZ was one of the first AI singing voice generators to hit the market. While their peers focused on celebrity deepfakes and memes, Controlla took a different route and created ethically-trained AI voice models for music creators.
Their new text-based AI choir generator is easily one of their most innovative features to date. Users can enter lyrics and turn them into massive ensembles that sound like real choir music. Download those samples and load them into your DAW to create original music.
An AI voice changer is also available, for those who want maximum control over the notes and vocal performance. In this article, I'll guide you through both techniques, starting with the lyrics-to-choir feature. Then we'll going deeper with the stem-by-stem approach.
Before diving into Controlla, I want to take a moment to explain why a tool like this is an exciting resource for beat makers and composers across any genre.
Table of contents
Choir sections make your music sound huge
Choirs and vocal ensembles give any track a huge sound. Take Childish Gambino's hit song This is America or Anderson Paak's The Dreamer. The choir sections create a soulful mood that levels up the track and makes it feel larger than life. Other major artists have sampled religious choir music for the same effect, like Kanye West in Jesus Walks with Me.
Writing choir samples from scratch is hard work
Most producers don't feel confident writing vocal choir sections from scratch. We're more likely to grab royalty-free voices from a sample marketplace like Splice. This shortcut has its downsides though - other producers can make music with the same ingredient and our song is less unique as a result.
Those of us who do have the skill to record vocal samples on our own still find it difficult to imagine what that arrangement will sound like once it's handed to a group of live singers. Will our idea still sound good with their vocal characteristics?
Controlla solves this problem with lyric-to-choir generation and voice swapping AI tools. They make it easier for people to generate acapellas on their own and use those samples across a wide range of different genres.
Text-to-choir is the easiest place to get started, but I'll show you how advanced users are recording multiple voices in their DAW and applying the voice changer to blend it all together.
Easy mode: Turning Lyrics into an AI Choir Sample
Let's start with the text-to-audio AI choir generator first. This path allows you to select an existing vocal style preset or upload audio files from a real choir in your personal collection.
The demo below features their co-founder Jeff Nang creating a cover song using lyrics from Kendrick's recent "Not Like Us". Check out how he incorporates these kinds of choral music samples into his music production workflows.
So let's get started. First you'll need to register an account on their website and verify your email. Log in and use the top menu to access their style sampler. From there you can select the "choir" tab as shown in the screenshot below:
As you can see, the workflow is simple as entering lyrics and selecting a vocal style. When you click the "select audio" region, Controlla will present you with three options.
Community Styles - Choose from a variety of choir presets included as part of the service. Each of these AI singers have their own timbre and singing style.
My styles - This is the library of custom styles you've created over time. They are generated whenever you upload reference audio or record directly into your browser.
Upload or Record - This is the main entry point for creating a new style. Upload a wav, mp3 or m4a file to get started.
You'll need to upload an audio file that's at least 15 seconds long in order to generate your own styles. Controlla will pass your lyrics through the voice style of your choice to generate your outputs.
As you build up a library of generative outputs, the newest ones can be found to the right hand side of the page. To view the full list, click the "My Library" tab in the top navigation menu bar and select "My style samples". You'll be able to search by voice or file name, to access the list of AI singers more easily.
Click on a voice sample to expand the view and find the download icon to the right side of the container. Save it to a folder on your hard drive or if you have an audio file manager like the AudioCipher Vault, you can organize it there instead.
Here's a second demo from Jeff, this time with an epic choir section for a Travis Scott type beat.
Advanced: Using a voice changer for note-control
Controlla's text-to-speech audio model is a great entry point for beginners. However, it does have some limitations like the inability to dictate specific notes or vocal articulation. Advanced users can go for the voice changer to get full control over the song's composition.
Music producers can follow the YouTube link below for instructions on how to build AI choirs from scratch. In this demo, their cofounder Jeff writes an original song using his own musical ideas. The process begins by recording vocal tracks one at a time in the DAW and then swapping those stems into a new voice style.
With this technique, music producers are no longer limited to creating choirs from their own voice or hiring live singers to record the ideas. It becomes easier to achieve a wide range of vocal timbres, in a much shorter period of time.
Step 1: Recording your vocal stems in a DAW
Start by opening your DAW project file and recording up to four layers of low vocals and four layers of high vocals.
The vocal performance on each stem should sound distinct—experiment with different tones and delivery styles to create the illusion of multiple singers. Keep your recordings clean and dry (no reverb or effects), so they’re ready for processing.
Once those are recorded, export each vocal layer individually. Make sure to label them clearly so it's easy to know where they belong in the arrangement. A simple naming convention like "high 1", "high 2" ... "low 5", "low 6" will work. The numbers indicate their track position in the DAW, from top to bottom. This will help the next step go quickly.
Step 2: Using "Swap Voice" to change the timbre
Upload your vocal stems to Controller Voice and navigate to the Swap Voice tab. This is where the magic happens.
Assign different vocal layers to AI-generated voices. For example, swap your high vocals with a female voice like "Lola" and your low vocals with a male voice like "Scott."
Optionally, apply pitch shifting to create additional vocal variation.
Click Swap Voices and let Controller Voice process the transformation.
Once you're satisfied, download the new choral layers and bring them back into your DAW.
Step 3: Mixing and expanding the Choir
Now it’s time to turn those AI-generated vocals into a full, immersive choir.
Panning – Spread the voices across the stereo field for a wider sound. Slightly pan some takes left and others right.
Grouping – Create a summing track (bus) to process all the vocals together.
EQ & Distortion – Roll off unnecessary low frequencies, boost highs for clarity, and add a touch of saturation for warmth.
Compression – Apply light compression to glue the choir together.
Reverb & Space – Use concert hall and church-style reverbs to place the choir in a grand acoustic environment. Layering multiple reverbs can enhance the depth.
After processing, your simple vocal recordings will sound like a full-fledged cinematic choir, ready to be layered into your composition. Add instrumental backing, and you've got an epic choral arrangement—all from your own voice.
Controlla's advantage over other AI choir generators
There are currently two other choir AI generator tools on the market that I'm aware of. Kits AI and Ace Studio both offer text-to-voice models that can be used to build choirs one track at a time. Here's a close look at how these three products compare, so you can decide for yourself.
Controlla vs. Kits AI: Quality and Flexibility
While Kits AI offers a selection of AI choir voices, Controlla prioritizes quality over quantity. It also differentiates on the number of custom voices you can have.
Unlimited Add-on Voices: Controlla's creator plan provides unlimited voice styles, whereas Kits AI limits their paid tiers to 2, 5 and 12 slots depending on how much you pay. Their 12-voice option is $48/mo compared to Controlla's $10/mo plan.
Lyrics-to-Choirs feature: Controlla transforms lyrics directly into a full choir, whereas Kits AI uses text to speech with singing voices for their data.
Higher Quality, Royalty-Free Voices: Controlla focuses on meticulously crafted, ethically trained AI voices, resulting in a more polished and professional sound. The majority of Kits voices are user generated and quality control is an issue.
Superior Voice Blending: Blending multiple voices is seamless and of higher quality in Controlla, accommodating any number of voices for richer harmonies.
Granular Stem Splitting: Controlla offers more precise stem splitting capabilities, giving you greater control over individual vocal parts.
Controlla vs. ACE Studio: Accessibility and Affordability
ACE Studio is an excellent AI voice generation DAW. However, when it comes to choral music, Controlla's new feature still has the upper hand.
ACE doesn't offer text-to-choir: The DAW does support lyric-to-voice generation, so it can be used to create choirs one voice at a time. However, just like Kits AI, they support the advanced approach only. They don't have an equivalent to Controlla's text-to-choir feature at this time.
Affordable Custom Voice Clones: Controlla offers custom voice clones at a fraction of the cost of ACE Studio's, which reportedly charged around $398 up front or $18-22/mo for two years, compared to Controlla's $10/mo price point.
Accessible Browser Interface: Controlla's browser interface can be accessed at any time and from any device, without software installation. You can use it from your phone while you're on the go. ACE Studio is a DAW and runs on desktop machines only.
Simpler Learning Curve: Controlla's streamlined workflow makes it accessible to both beginners and experienced producers. ACE studio is a more complex tool and
The ethics of using an AI choir generator
Artists have expressed concern about AI generated music and its impact on the music industry. Instant song generators like Suno and Udio scraped audio from the internet without paying for licenses. This approach is unfair to creatives and it overlooks something incredibly important; the joy of making music as a recreational hobby.
Suno's CEO Mikey Schulman even went so far as to say that making music is "too hard" and that "most people don't enjoy making music". Social media erupted with criticism over that statement, showing that people fundamentally disagreed. The overwhelming consensus was that we love making music. That's why we do it!
Controlla's co-founders have precisely the opposite attitude. When I first discovered them back in 2023, they were already prioritizing ethical artificial intelligence. Other companies rushed to deliver celebrity deepfake models, but their singing voices were not great.
Controlla is committed to creating AI voices for music producers specifically. All of the AI Voice models have been trained with consent from the rights holders in their dataset. This makes it the most affordable, ethical choir AI generator in the music industry today. You can visit the Controlla website to learn more.