Sound Lexicon
Two libraries about the same stubborn problem: which words actually change what a generative music model produces, and which ones only sound like they should. One collects prompts that work. The other takes musical terminology apart and asks whether the model hears it at all.
Ready-to-use prompts for singing voices — grain, register, delivery and the production around them. Every entry is a usable artifact: copy the style string, paste it, get the sound. Searchable, mapped, and demonstrated.
What musical terms mean in plain language — instruments, tools, techniques and song theory — and whether the model actually distinguishes them. A card here is a lesson about one word, not a prompt. Searchable by the words a listener would really use.
Tools
Drag Suno or YouTube links in, put them in order, and play them straight through. Library demos name the entry that made them. Keep it to yourself, or share it.
A log of working sittings — what was made, in the order it was made, and the correction that made each round differ from the one before. Open a song for its world, its slots, the style it was fired with and the lyric as submitted. Heart the takes worth keeping.
Lo-fi cues for a comfy zombie-survival base-builder. Every create with the in-game inputs it was fired with: mood, energy, setting, weather, threat. Mark each take keep, maybe or cut, and see which inputs are landing.
Queues people have published, with a chat on each one. Paste a link in the chat and the owner can add it with a click. Yours are yours to delete.