GoKawiilHacker News user asks why audio-plus-text to audio AI tools remain scarce
A poster on Hacker News describes searching for a commercial AI model that accepts both a reference sound and a text prompt to generate new audio, similar to how image models use reference images plus prompts. They report trying ElevenLabs' text-to-SFX tool and Stable Audio 3, saying neither handled reference audio well or produced accurate results for their sound-effect generation use case.
GoKawiil's interpretation of the reporting above, not reported fact.
The post highlights a gap between the maturity of multimodal image generation tools and the comparatively underdeveloped state of audio generation guided by reference sound, suggesting this could be an underserved niche for AI audio startups. It also serves as informal market research, inviting the developer community to share whether such tools exist or confirm the need is real but unmet.
- Reference-plus-text to audio generation appears far less mature than equivalent image generation tools.
- The user found existing tools like ElevenLabs SFX and Stable Audio 3 insufficient for accurately recreating reference-guided sounds.
- The post functions as a crowdsourced request for alternatives, pointing to a potential unmet demand in AI audio tooling.
Focusrite Scarlett 2i2 Audio Interface — If you're working hands-on with sound effects and reference audio, having a solid audio interface for capturing clean source recordings can make a real difference in the quality of material you feed into any generation pipeline. The Scarlett 2i2 is a popular choice for hobbyists and professionals building custom sound libraries. Better raw recordings mean better reference material regardless of which AI tool eventually handles the audio-to-audio generation step.
See Focusrite Scarlett 2i2 Audio Interface on Amazon → Affiliate link — we may earn a commission on purchases, at no extra cost to you. Product picked by AI based on this article; it is not a tested recommendation.Source: news.ycombinator.com, 2026-10-05
Published there as: “Ask HN: Are there AI models for generating sounds based on a text and reference?”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.