> ## Documentation Index > Fetch the complete documentation index at: https://help.soundverse.ai/llms.txt > Use this file to discover all available pages before exploring further. # DNA - FAQs ## 1. What is DNA? **DNA** is your artist-trained AI music model on Soundverse. * Each DNA is built from the music or vocals of a real creator. * It learns that creator's **sound** – production style, harmonies, groove, textures, and (for voice models) vocal tone. * You can then generate new songs, instrumentals, or vocals that stay true to that DNA's vibe, while still being original. There are two main types: 1. **Full DNA** * Trained on full songs / instrumentals / acapellas. * Can generate **Song, Instrumental,** and sometimes **Singing** outputs depending on how it's trained. 2. **Voice DNA (Voice Model)** * Trained only on a singer's voice. * Used to generate **vocals** or to **replace your own voice** with theirs in a new performance. ## 2. What's the difference between a Full DNA and a Voice Model? * **Full DNA** * Trained on your music catalog (songs, instrumentals, acapellas). * Can generate songs, instrumentals, and sometimes vocals. * Also exposes Track presets (Song/Vocal/Instrumental Styles). * **Voice Model (Voice DNA)** * Trained on your **voice only**. * Used to generate or replace **vocals** (Singing output). * Does **not** include your production / instrumental style. ## 3. What are "Song", "Instrumental", and "Singing" outputs exactly? * **Song** * Full track: **music + vocals** (where supported). * **Instrumental** * Music only, **no lead vocals**. * **Singing** * Vocals only, typically using a **Voice Model** or vocal-enabled DNA. Availability depends on: * DNA type (Full vs Voice). * Track type (Song Style vs Instrumental Style vs Vocal Style). ## 4. Why can't I select all three outputs sometimes? Because the output options respect **what data the DNA/track was trained on**: * If you pick a **Song Style** track → All three outputs enabled. * If you pick an **Instrumental Style** track → Only **Instrumental** enabled. * If you pick a **Vocal Style** track → Only **Singing** enabled. * If you're using a pure **Voice Model** → Focused on **Singing**. This prevents you from asking for outputs the model doesn't have proper data for. ## 5. How many files should I upload to build a good Full DNA? * **Minimum:** 3 tracks. * But for better quality: * Aim for at least **5–10 solid, representative tracks** if you have them. * Up to **hundreds** if you have a large catalog – just note training will take longer. Always prioritise **quality and consistency** over sheer quantity. ## 6. What kind of audio should I upload for a Voice DNA? * Clean recordings of **one single speaker/singer**. * No background music, crowd, or heavy FX. * At least **2 different files** (under 5 minutes each), ideally more: * Different phrases.
Different dynamics. * Mix of natural speech + singing if appropriate. ## 7. What does DNA Sensitivity actually do? **DNA Sensitivity controls how many AI sub-models we create from your catalog.** If your catalog spans multiple genres (say **EDM, pop, rock**), DNA Sensitivity tells the system how finely to slice those styles: * **Higher sensitivity** * We detect **more nuance and sub-genres**. * Example: instead of a single "EDM" DNA, you might get separate DNAs like **House, Festival Techno, Synthwave, Future Bass**, etc. * The same applies to your Pop and Rock material – they can be split into more detailed sub-styles. * **Lower sensitivity** * We keep things at a **broader level**. * Example: you might just get **EDM, Pop, Rock** as three main DNAs, without breaking them into sub-genres. If you’re unsure what to choose, **leave it at the recommended value** – it’s a balanced setting that works well for most creators. ## 8. What is "Model Quality" in Voice DNA? For **Voice Models**: * "Model quality" adjusts how much compute/detail goes into capturing your voice. * Higher quality = more expressive and accurate voice modelling, at the cost of **longer training**. * Start with the recommended level and adjust based on results. ## 9. Do I have to stay on the page while DNA is training? No. * After uploading audio and starting training: * **Tagging and categorisation are automatic**. * You can **close the tab** or **move away**. * You'll get an **email** when your DNA is ready for review. ## 10. How do I get paid for my DNA? 1. Go to the **Earn** tab. 2. **Connect Stripe** (Stripe Connect account). 3. Build and **publish** your DNA as Public. 4. Ensure **Become Partner** is set to **Yes** (if available in your region). 5. When other users generate content using your DNA, and export it, your share is: * Calculated according to the **license and price** you chose. * Shown inside the **Earn** dashboard. * Paid out via Stripe's payout schedule. ## 11. Can I change my DNA price or visibility later? In general: * **Price** – can typically be adjusted later (future uses follow the new price). * **Visibility** – you can switch between **Public** and **Private**, which: * Controls whether your DNA is discoverable in **Discover** and usable by others. Use **My DNA** and the per-DNA options to manage this. ## 12. What rights do users get when they generate with my DNA? * This depends on the license you selected (e.g., Royalty-free, Distribution). * In general: * A **Royalty-free** license supports broad usage but under non-exclusive terms. * A **Distribution** license focuses on release/monetisation rights via DSPs, etc. * Users should always check the license label on a DNA before using it commercially. * As a creator, always choose the license that matches how comfortable you are with people using your sound. ## 13. Can I build multiple DNAs under the same creator profile? Yes. * The **Creator Name** is your umbrella identity. * Under that profile, you can create: * Multiple **Full DNAs** (e.g., AfroDesi, Piano House). * Multiple **Voice DNAs** (e.g., different personas, languages, or timbres). Each DNA can have its own: * Description. * DNA Picture. * Price. * License. * Visibility. * DNA and Track metadata. ## 14. What if I want to test a DNA privately before making it public? * During **Profile Creation**, set **DNA Visibility = Private**. * Train and test the DNA using **My DNA** only. * Once happy, update it to **Public** so it can join Discover and generate revenue. ## 15. Can I use my own DNA to make songs for release? Yes – that's a core use case. * Use your own DNA in **Model Based** or **Track Based** mode. * Generate songs that reflect your signature sound, then: * Optionally edit / mix / master them in your DAW. * Release under your artist name, respecting the **Soundverse license** attached to that DNA (which you control). ## 16. What happens after I publish a DNA? Once you hit **Finalise / Publish**, your DNA doesn't go straight to the store instantly. It goes through a short **review and moderation** process by the Soundverse team: 1. **Internal Review** * The team checks: * Audio quality and consistency. * That the content follows Soundverse's ethical and legal guidelines (no obvious copyright abuse, hate content, etc.). * That tags, description, and pricing look reasonable. 2. **If there's an issue** * The Soundverse team will **email you** at your account email. * They'll explain what needs to be changed (e.g., problematic tracks, unclear rights, inappropriate metadata) and what you can do to fix it. * You can then adjust the DNA (or your uploads) and resubmit if needed. 3. **If everything is fine** * Your DNA is **approved** and goes **live in the store**: * It appears under **My DNA**. * If set to Public, it can show up in **Discover** (Top Creators / Top DNAs) and be used by other creators in the DNA generation panel. * From that point, usage of your DNA can start generating **royalties** for you (as per your license, price, and Stripe setup).