Obsidian Neural AI SAMPLER VST futur integration
Hi! I was about to integrate Foundation-1 into my VST Obsidian Neural (https://github.com/innermost47/ai-dj) because the musical structure and control are exactly what I've been looking for. It's truly a fantastic model.
However, the lack of drum generation is currently the only thing holding me back from fully switching to it. Do you have plans to include percussion in future updates? This model is SO close to being the perfect tool for my workflow. I'm really hyped for what’s coming next!
Update: I went ahead and integrated Foundation-1 anyway - it's now live on the OBSIDIAN Neural provider network as of today!
For percussion and rhythmic content, I implemented a simple routing system: the LLM layer detects whether the user is asking for drums/beats and automatically routes to Stable Audio Open instead of Foundation-1 for those requests. For everything else (synths, pads, bass, strings, brass, etc.) Foundation-1 takes over. Works surprisingly well.
Really happy with how it turned out - the musical coherence and prompt control of Foundation-1 is a massive step up for melodic content. Thanks for building this.