An artist name is an adjective with nothing under it

Warm. Dry. Epic. Intimate. Ethereal. Every one of those is a real word about music, and not one of them tells Suno to do anything.

That is the same reason the band name you typed did not work, and it is the part the guides skip when they hand you a table of adjectives to use instead.

The name is blocked before anything is generated

This is documented, not folklore. Suno's moderation article lists what will stop a song being made, and the first item is "the names of well known artists or people", alongside copyrighted or trademarked terms. The consequences it lists run from the generation being refused to material being removed without warning.

So the advice to describe the sound instead is correct. It is just incomplete, because nobody says what a description has to contain to beat the name it replaced.

The best descriptor list in the research was written by an engineer

The largest vocal vocabulary in our research sits on a thread from August 2025. The list in the post is a screenshot, which is why the readable part is a comment underneath it, and that comment is the interesting one anyway. It gives ten metal vocal types with a specification for each. Fifty-six descriptors in total.

Here is one entry, unedited:

Blackened Screech (Male)
High-pitched rasp, thin-band EQ, nasal resonance, airy fry layer,
plate reverb, slight pre-delay, icy top boost

Read those fifty-six descriptors and a pattern shows up that survives every entry. The names of the voices are feelings: Wraith Whisper, Epic Power Tenor, Pig Squeal, Blackened Screech. The specifications underneath them are engineering: fry, chest resonance, EQ bands, multiband compression, plate and hall and cathedral reverb, pre-delay, doubling, stereo width, mic proximity.

Not one of the fifty-six is a feeling word standing on its own. Every adjective in the list is attached to something physical. Icy attaches to a top boost. Warm attaches to mids. Soaring attaches to a tenor belt. Dark attaches to an EQ.

Warm what?

That is the whole rule, and you can run it on any style prompt you have ever written. After each adjective, ask what it is an adjective of.

Run it on somebody else's work and it stays useful. A June 2026 guide on getting an artist's sound without the name offers this as a worked example: "1980s synth-pop, mid-tempo, analog synth arpeggios, gated reverb drums, melancholic and nostalgic, smooth emotive male vocal." Three of those name something somebody did, and the arpeggios and the gated reverb are why the prompt works at all. Melancholic and nostalgic are riding along. Cut them and the generation barely changes, which is a cheap thing to try on a prompt you are already unhappy with.

An artist name is the extreme case. It is the most compressed feeling word available, carrying an era, an instrumentation, a voice and a production style at once, and it names none of them. Two people who both write the same name get different songs and learn nothing from the difference, because neither of them said anything a model could act on.

Four questions that produce the nouns

The reason people reach for a name is not laziness. It is that the nouns are genuinely hard to hear until somebody tells you what to listen for. These four get most of the way there, and you can answer them while the reference track is playing.

How close is the voice? Close enough to hear breath, or set back in a room. That is mic proximity, and it is most of the difference between an intimate record and a big one.

Is it one voice or more than one? Doubling is everywhere and almost nobody names it. A doubled lead sounds confident, a single lead sounds exposed.

What room is it in? No room at all, a small dead one, a plate, a hall. Reverb type does more for era than any decade tag.

What is holding the low end? Upright bass, a synth sub, a kick with no bass at all. This decides the weight of the whole thing.

Answer those four and the prompt writes itself:

early 1990s indie rock, jangly twelve-string, dry drums with no room,
close-mic male vocal doubled on the chorus, upright bass, tape saturation

Six descriptors, every one of them a thing somebody did, and every one a dial you can turn on the next generation to hear what it was holding.

The band name would have been faster. It would also have been a guess, made by you, about music you had already heard, and handed to a model that has no way to check it.