Generative AI tools took the world by storm in 2023, culminating in an unexpected end-of-year partnership between Microsoft and text-to-song company Suno AI.
In March 2024, the company beta tested and released their latest v3 model. Celebrated by many as the best music generator to date, the web app has achieved relatively high fidelity across and incredibly broad spectrum of subgenres.
Vocal AI music ethicist Ed Newton Rex of Fairly Trained has proceeded to grill the company publicly, both on Twitter and in major publications like this article from Music Business Worldwide. Several other prominent voices have taken to social media to protest the lack of transparency from Suno and their most recent competitor, Udio.
In April 2024, a member of the US House of Representatives proposed a bill that would require companies like Suno and Udio to disclose the music in their training data. There's currently no indication of imminent penalties for training, though it would seem to be implied. This would be the first step in disclosure.
Table of Contents
Getting started with Suno AI
When you sign up for Suno, you get a certain amount of generation credits for free. They've been experimenting with pricing models, however in some cases they have required users to pay for a subscription access the new V3 model.
Below is a screenshot of what the interface looks like from the Create view. As you can see, there's a custom mode that supports the option to enter lyrics or themes. You can use AI to rewrite the lyric prompt and build out full verse and chorus sections.
Below the lyric area is an instrument switch, which disables lyrics entirely. Users indicate the style of music they want using any combination of mood or genres that come to mind. There's a model selector dropdown at the bottom of that panel, followed by the create button.
Fast generation times: Where most AI music generators take minutes to create a single song, Suno has developed a unique streaming technology that begins playing the track before it's completely rendered. This seems to indicate that they're using a transformer model under the hood.
Low cost: The current pricing structure at Suno is extremely competitive, offering 500 2-minute songs (1000 minutes total) for just $10/month. By comparison, other services we audited charge the same amount for 20 minutes of music generation.
Playlists: Every time you submit a prompt and hit create, two new tracks are generated. Suno spins up some custom art and displays your lyrics in the right hand panel, along with a voting mechanism to indicate if you like or dislike it.
Download and share: The Suno community has been extremely active, long before this web app was available. You can share into the Explore page, post direct links on social media platforms like Twitter or Meta, and even download the file in an mp3 format directly to your computer.
Suno’s vision for the future of AI music creation
I met with two of Suno’s co-founders, Keenan Freyberg and Martin Camacho, to better understand the company’s longterm goals. We spoke at length about AI-powered music creation and the crowd that they’re aiming to serve in 2024.
Camacho explained they want to make songwriting fun and easy for all people, regardless of music training. With Suno, you would never be more than a text prompt away from turning your lyrics into original music. That goal has already been achieved with flying colors.
They currently have no plans to compete in the DAW plugin market, but that hasn’t stopped thousands of songwriters from using the app as a source of creative inspiration.
The Suno AI + Microsoft CoPilot partnership
To understand the significance of Microsoft’s partnership with Suno AI, it might be helpful to rewind briefly and see where each company was positioned during 2023.
Suno was the first to offer AI generated instrumental music with compatible vocal melodies or rapping on top. Tracks are generated end-to-end, with lyrics and instrumentation created simultaneously, making for lightning-fast delivery.
Microsoft, on the other hand, was falling behind in the AI music race. In August 2023, I had interviewed Microsoft’s Principal Software Engineer for Windows MIDI Services, Pete Brown, to learn more about their AI audio research team Muzik and their new MIDI generator MuseCoco.
The pseudo-classical compositions MuseCoco were underwhelming at best. The team offered no public interface where people could try it out and the demo page felt sterile.
On the other hand, Microsoft launched their Bing chatbot in February 2023 and enjoyed a considerable advantage over Google and OpenAI for most of the year. That upper hand likely inspired them to leverage the popularity of Copilot to introduce other AI tools and services.
So you can understand then, why a partnership with Suno AI was so valuable to Microsoft. Copilot added new Suno plugin around mid-December and now the whole internet has access to its text-to-song capabilities. Discord is still available but no longer a requirement.
Suno Discord: Accessing Chirp and Bark
For the best experience with Suno, we recommend that you use the web browser app instead of Discord. The interface is much cleaner and easier to navigate.
That being said, if you prefer to use Discord, get started with this invite link. You can run Discord from your web browser or download the standalone desktop app for free here.
Once you're in Suno's Discord server, read through the terms of service. If you agree then you can start creating music. On the left navigation menu, you'll see a Generate category with a number of channels labeled chirp-beta. There's a
There are close to ~250,000 users registered in the server currently. To reduce the clutter, they've created a few clones of the same chirp-beta channel. Pick any of them to get started. Each one features a long list of user-generated lyrics, music and art that you can freely explore.
At the bottom of the channel is a text area, where you can type the command "/chirp" and summon a special modal. This is where you'll type in your lyrics, or have ChatGPT make lyrics about a topic of your choice.
Once you've submitted your lyrics, be sure to check the time stamp. Chirp needs a few minutes to render the vocals and instrumental music. When it's finished, you'll have to scroll up to that time code to find it.
Here's an example of how one of our music producer friends, Simulation Beats, used Chirp in Discord to create original songs:
When a new track has been generated in the Chirp-beta channel, you'll find a set of controls directly below the lyrics. Here's what each of them does:
Recycle symbol: Generate two new songs using the same lyrics. The vocals and background music will change completely.
Heart symbol: Let the channel know that you liked something. A full-width feature of the song will appear to help make it more visible for everyone else. People can add emojis to celebrate it together.
Fast-forward (extend) symbol: Use the same vocal tone and instrumental arrangement to come up with new melodies and chord progressions, based on new lyrics that you provide.
Tweet: Share a Chirp song on Twitter so your audience can check it out
How to use "Continue" to expand on your Chirp songs
Of all these controls, the option to expand on your music is by far the most powerful option. Most AI music generators give you a single piece of content without the option to continue in that style. Suno lets you extend an existing piece in any genre, carrying over some qualities (like BPM) despite the shift in style.
To download your best ideas, hover over the song and click the download icon located in the top-right corner of the image as shown here:
The songs will be saved as .wav files, so musicians can import them into a DAW with the highest resolution.
This is where the Suno journey ends and you get to decide what to do. Will you chop it up, use a denoiser, stem separate, or perhaps run an audio-to-MIDI app like Samplab to get help figuring out what the chords and notes are?
I've personally transcribed a few Chirp songs to guitar and vocals already. The composing style is very free and as a result, I discovered new melodic ideas and chord progressions that I wouldn't normally use.
At this time, Suno doesn't seem to have any plans for rolling out text-to-MIDI. Music producers looking for inspiration in the DAW can check out AudioCipher VST.
A timeline for text prompts in AI generated songs
If you’re newer to this space, you might be wondering how the whole text-to-music wave began. During the summer of 2020, AudioCipher released the first ever text-to-MIDI VST, made for music producers working in a DAW.
AudioCipher was based on musical cryptograms instead of artificial intelligence.
The public’s reaction to AudioCipher was a mix of amusement and bewilderment. It was only when OpenAI released Dalle-2 that the value of text-to-music became fully apparent.
By December 2022, OpenAI had also released ChatGPT while simultaneously turning off access to their AI MIDI generator, MuseNet. A small tech startup called Riffusion released the world’s first ever AI text-to-music web app at the end of that year as well.
Suddenly anybody could go online and generate original music from a few text prompts. The audio quality was in some ways very bad, and yet the text-to-music experience was so gratifying that it didn’t matter. Riffusion went viral overnight and raised a $4M round in less than a year.
A chain reaction was set off, first at Google and later at Meta. Both had in-house teams working on an AI music generator. Google published their MusicLM research paper on Github less than 30 days after Riffusion’s release.
About six months later, MusicLM and Meta’s MusicGen became available online for free and public use. Each service offered its own aesthetic quirks, but also shared some basic properties in common. They generated short, low-fidelity instrumental clips in almost any style of music.
None of the apps I’ve mentioned were providing a service that turned lyrics or song ideas into finished songs. Yet that seemed to be exactly what people were craving. Influencers racked up millions of views showcasing DAW workflows with AI singing models and AI music generators.
One genai pop song, Heart on My Sleeve, broke records on the Spotify and TikTok charts. This forced an intervention and a number of challenging but important conversations around AI music and copyright laws.
At the end of 2023, we saw three major contenders in the AI text-to-song space, with Suno currently taking the lead over Splash Music and Riffusion in my opinion.
The music industry is working closely with companies like Rightsify to establish new, relevant copyright laws. Fair artist remunerations will be a critical part of this process, but it will take time for those standards to emerge.
Visit Suno's website directly or at copilot.microsoft.com to try it out!