Suno the AI Music Revolution

(Edit: This blog has been updated to include the latest news 04/04/2024)

How we interact with creative mediums is rapidly evolving. AI-powered tools are assisting and even undertaking the creative parts of many jobs, leaving everyone wondering what the future holds. At AI Heroes, we know this is just the beginning of a world run by accessible AI. A world that will revolutionise how you approach creative marketing – and Suno, the AI music generator taking the world by storm, is just the next piece of this transformation.

We’ll explore good tips and tricks for using Suno effectively, and discuss how it has managed to penetrate the music generation space. Suno is just the next piece of this transformative journey, and we’re excited to guide you through it.

That audio clip you just heard was fully AI-generated – from lyrics to melody and rhythm. While it may not top the Billboard charts yet, it’s more than enough to work with. This is all thanks to Suno’s cutting-edge v3 model, and a quick browse through their website reveals a treasure trove of user-generated content already.

AI with Music Theory

How did Suno break into music generation so quickly, outshining attempts by tech giants like Google, Meta, and Stability AI? The secret likely lies in Suno’s use of multiple models working together. While official details are not available, it’s safe to assume multiple models are at work. The true artistry, however, is in the seamless mixing of these elements, resulting in a product that doesn’t just resemble music but is genuinely creative.

Suno’s generations are more than just looping melodies, they’re complex with multiple instruments, choruses, breakdowns – it’s as if the AI has learnt music theory to orchestrate its creations. It’s a sight to behold, and it makes this technology not just interesting but truly valuable.

Demolishing the Competition

Suno demolishes its competition, and that’s why you should care. The only real competition is Google’s MusicFX, Meta’s AudioCraft, and Stability AI’s Stable Audio. While technically impressive, these models are capable of generating sounds, not fully realised music.

LoFi Beat – Generated by Stability AI’s “Stable Audio”

Funk – Generated by Google’s “MusicFX”

It’s clear that Suno stands out for the high quality of its generated music compared to other AI models. Its ability to create full songs with realistic vocals, lyrics, instrumentation, and song structure in a wide variety of genres is unmatched.

In contrast, Google’s MusicFX and Meta’s AudioCraft tend to generate more artificial-sounding instrumental music without vocals, while Stable Audio, though promising, still lags behind Suno in overall quality and realism.

Stable Audio v2 – How Does it Compare?

Barely a week after Suno shook the music AI world, Stability AI fired back with their Stable Audio v2. In this rapidly evolving space, what does Stability AI have to offer? Stable Audio offers some interesting perks, including a more customisable experience, pre-built prompt libraries, and most interestingly, a music-to-music option. With Stable Audio, you’ can ‘re able actually mix your creations and have more ways to interact with them. These are features that we believe should become standard for any music generation service. However, they mean little if the baseline model isn’t up to standards.

Sadly, we would argue Stability Audio v2 is still lagging behind what Suno can do, regardless of the fact it won’t generate vocals – it’s just not very good at generating instrumental audio. It’s very clearly looping and lacks the same anility to create exciting movements in the songs, choruses and breakdowns are something it does very poorly. This might not be a big issue if your optimal use case is to just use samples of audio but as an over all product its still lacking.

How to Use Suno

Suno has a very easy to use UI. Just have a quick look below:

However if you want the best quality results we recommend you giving the following a go.

1. Suno uses ChatGPT 3.5 for their lyric generation. We recommend using Claude for free instead, or GPT-4.0 if you have access to it. Lyric quality becomes much better, especially if you also develop what genre you want it to be.

2. The voice might sound a little too echo-y and have too much reverb. Look into doing some basic audio mastering if you have the capability to make it sound crisper.

3. Don’t be fooled by the lyrics input section. It has much more utility than just that! For example, you can guide how many voices, what instruments, and what type of singing style (acapella, white voice, etc.) you want in each section.

Useful Examples with Suno Prompts

Suno’s main appeal is for artists looking to incorporate generative AI into their music. But what does this mean for you?

AI-generated music’s true gift is the great amount of options it opens up. Many modern genres heavily rely on sampling and remixing existing songs, often forcing DJs to navigate complex copyright laws to make their work legally viable. Similarly, businesses frequently have to invest in purchasing licenses just to have background music playing.

However, these issues disappear when dealing with AI-generated samples, allowing for greater creative freedom and cost-effectiveness. Here are just three examples of why music generation could be useful for you, regardless of who you are:

1. Quickly Creating Copyright-Free Content: Suno won’t replace music entirely, but it serves as a great entry point for those who need music. Copyright infringement, a lack of royalty-free samples, and the difficulty of producing songs mean independent creators and small businesses often miss out. With Suno’s premium accounts, you can use AI-generated songs anywhere without worrying about licenses. Whether you’re a small business needing background music for promotional material or a DJ without access to expensive licenses, Suno has you covered.

Instrumental track,ambient,relaxing,atmospheric vibe,non-distracting from narration,royalty-free

2. Altering Existing In-Built Music: For many businesses, acquiring music licenses is just the beginning. As promotional material evolves and plans change, the original intent for the music may shift. Getting music altered or creating more in the same style is an expensive endeavour, risking wasted resources. Music generation empowers businesses to alter and fine-tune music in-house, saving costs while allowing for greater flexibility and precision in how music is utilised. For now this could be easily achieved by feeding your music through chat-GPT, and then through Suno:

[Lyrics Generation In chat-GPT]

Prompt: Convert the following song lyrics into a similar poetic style while maintaining the original meaning and emotion:

[Insert Song Lyrics Here]

3. Studying Mnemonics: A fun example of how being creative with AI can help you in ways you least expect. Studying and generative AI go hand-in-hand, not by making learning easier but by giving students control. Some learn better visually, others aurally. Being able to create songs with custom lyrics allows anyone to craft catchy mnemonics for studying, even in languages with smaller speaker populations. This could be revolutionary for preserving endangered languages.

Create a catchy and memorable song with lyrics that can serve as a mnemonic device for helping me study the following material:

[Insert Study Material Here]

The song should be in the style of [specify genre/artist if desired], and the lyrics should creatively incorporate the key concepts, facts, or information I need to remember from the study material. The goal is to make the content more engaging and easier to recall through a musical mnemonic.

These are just three small ways you could use Suno. We’re excited to see what later versions have to offer. Simple prompting is a powerful way to interface with the model but we’re excited for music inputs and the ability to fine tune more with the model, similar to what Stable Diffusion is able to do.

Accessibility

Suno is truly the only true consumer service out there. Its web-based interface is intuitive, allowing users to generate complete songs from simple text prompts. The ability to specify details like genre, mood, instrumentation, etc. provides control without requiring deep musical knowledge.

Other AI music tools like MusicLM and AudioCraft are more research-oriented and lack the user-friendly interface and features of Suno at this stage. They are not as well-suited for the average user or businesses looking to easily create songs.

Copyright and Ownership

An important consideration with all AI generation models is copyright. Suno allows users who subscribe to its paid plans to own full rights to the songs they create, and use them commercially.

MusicLM and AudioCraft have not clarified ownership rights and Stable Audio has not provided details on its data sourcing. More transparency is needed across the board since there are still valid concerns around copyright and proper compensation for artists whose work is used to train these AI models. As the technology advances, clearer legal and ethical frameworks will need to be established. But Suno has taken an early lead in the race to define the future of AI music creation.

Our Musical Creations:

AI Heroes Theme Song

Lyrics

AI Heroes, they’re here to save the day!
Using algorithms in the cleverest way
From chatbots to robotics, they do it all
With AI solutions that stand tall

But what’s that? A bug in the code?
The AI’s gone rogue, it’s out of control!
AI Heroes scramble to shut it down fast
Before the robot uprising is vast