Tip: How to get SUNO to generate the specific sounds you want
I see a lot of complaints that tracks generated by SUNO are "too generic," which generally boils down to frustration that it is not outputting what people want.
And the truth is that the more you leave the algorithm to guess what you want, the less satisfied you're going to be with what you get.
Short of learning to play an instrument, recording the bones of a track and inputting to SUNO -- which is not what many people want to hear, and always devolves into arguments -- the best solution is to learn to better describe what you want.
That also presents a problem, because SUNO won't allow you to use artist and song names the same way a lot of image generators won't let you name artists or characters.
So how do you get around this problem?
By using the vast amout of information about popular music available on the web.
I will use Michael Jackson's Thriller as an example, since everyone is familiar with the song.
You can look up information on the studio sessions for Thriller, or better yet, you can go to Youtube to find people recreating the song on the original hardware used to record it.
Here's one such video:
https://youtu.be/BH9LJaaokj4?si=L2GDdpis_0pIIzAV
From this, you can learn that the majority of the synth and drum machine work on Thriller was recorded on a Jupiter-8, a Minimoog, and the ARP-2600.
The Youtuber even breaks down the wave forms, oscillator, attack and reverb settings, etc.
Now when you go back to SUNO, you can tell the algorithm that you want a synth bass as generated by a Jupiter-8 saw wave, and give it explicit instructions on reverb, attack, delay, etc.
Try it. It works!
If you want a specific progression you will still need to upload a stem file because SUNO still can't follow simple and explicit instructions on chord progressions and notes, but you can combine the above information with specifics on genre, rhythm, BPM, etc.
When you put in the extra effort like this, entirely new possibilities open up as you are able to exert more control over the sound.
Is it the same as playing it yourself? No. You will still need to learn to do that if you want absolute control. But that is a fact of life, even in the age of algorithms, and it is the same thing you would have to do anyway if you were using, say, the software VST modeled versions of a Jupiter or a Minimoog by Native Instruments, Arturia, etc.
If people find this helpful, I will post more examples of this sort of thing, at least within my realm of experience and genre familiarity. But a lot of this stuff is universal anyway, or as close as it's possible to get in music.
And the good news is, unless you're trying to emulate the sounds of an extremely obscure artist, you should be able to expand your descriptive vocabulary and greatly increase your control over the sounds you get.
There is a universe of difference between doing it this way and simply writing a prompt like "hip hop, west coast, summer vibes, slow tempo" and hoping for the best.