Suno is one of the leading artificial intelligence music generators, capable of creating an entire song — with vocals — from nothing but a text prompt.
It has improved rapidly in the year since it first launched, extending the duration of songs, adding new community features and even allowing you to create a singer.
With version 4, things will move to a new level. The next-generation model is set to be released "soon." Still, we've seen several examples of its output on social media from Suno employees and people given early access — and it's impressive.
Suno has an iPhone app that allows you to create a song from an image, and this is likely to include v4 when it launches. If you want to get started with Suno we have a guide to using the AI music generator.
What makes Suno v4 so different?
I haven't personally put Suno v4 to the test, so I'm going purely on the videos I've seen shared on social media but the vocal quality and overall production tone seem leaps and bounds above anything we've seen before.
It is possible v4 also comes with other new features but for now, based purely on the sound, its going to be an upgrade everyone will want to put to the test. Many of the sample tracks I've heard made me double-take over whether it really was AI.
One area Suno has always done well is country, but the vocals always felt a little like someone had overdone the autotune. Not with v4. This sample from imolivercom sounds like something I'd find in a top Spotify playlist.
This is #SUNOV4 in Action!Country: pic.twitter.com/hj1HBknRcENovember 12, 2024
The combination of music and video is going to be significant, especially now we have good lip synching. Here's an example from AI creator DAViD on X that shows just how good the voice parts are from Suno v4.
It starts with apes and monkeys talking about Suno generally, before they break out into song with the music produced using v4.
Suno v4 is 🔥@sunomusic Video made with @runwayml, which is also 🔥Sound on 🔊 pic.twitter.com/hq0JQ551JyNovember 12, 2024
Jamain Gerard went more pop with his Suno v4 sample. He animated with Viggle and wrote: "The advancements in song generation and vocals with V4 are miles ahead of V3. This is only the beginning!"
So far most of the sample clips have only been a few seconds long but that is more than enough to see the marked improvement in vocal tone.
🚀 My First @sunomusic V4 Early Access Creation: Ignite the Night! 🎶The advancements in song generation and vocals with V4 are miles ahead of V3. This is only the beginning! 🎹🔥 Coming Soon 🔥💃✨ Dancing and animation by @ViggleAI#SunoV4 #AImusic pic.twitter.com/QodtzMU45QNovember 12, 2024
The Suno team shared their own v4 example with a real soulful tune giving off John Legend vibes (without sounding like him at all). Suno puts a lot of effort in to ensure users can't just replicate the sound of real artists.
This sample is what first got me excited for the new model as it demonstrates a level of vocal fidelity even human singers struggle with.
v4 is coming soon 🎧 pic.twitter.com/UIke42RNncNovember 11, 2024
Finally we go back to imOliver who responded to questions from his followers about how well it handles heavier songs. This is a nu-metal track made in v4 called Stone.
I don't think its as good as some of the other examples, but I'm also not a big metal, nuMetal or heavy music fan so it may just be my personal bias. I think you could get something close to this with v3.5, which was always good at heavier tracks.
I see alot of people are asking for Heavier Examples of #SUNOV4 so here's a nu-metal version of 'stone' created with V4 pic.twitter.com/MLcH1EK0CBNovember 13, 2024
AI generated music is coming on so quickly that we've gone, in a year, from not being able to generated vocals at all to creating custom singers across multiple tracks that sound like a human.
Not everyone is going to be a fan of AI generated music. There are ongoing court cases around the use of commercial music in training data and musicians are concerned for the impact it will have on their income.
I think both worlds can exist and in fact human musicians using AI will allow us to enter a new golden era of music with new genres, mashups and styles coming faster than ever before as experimentation just got a whole lot cheaper.
