Mirelo SFX 1.6 Brings Iterative Audio Editing to Generative AI

Summary

Mirelo has released SFX 1.6, a major update to its sound effects generation model that introduces iterative editing capabilities for the first time. The new version enables audio inpainting—the ability to regenerate specific sections of audio clips while preserving the rest—addressing a fundamental gap in how generative models have operated. Rather than requiring full re-generation when one detail is wrong, creators can now target and fix specific problems.

The update brings four major new capabilities: long-form generation up to 60 seconds (six times longer than before), audio extension that seamlessly continues existing clips, looping for ambient soundscapes, and inpainting for cleaning unwanted sounds from existing recordings. These features are available immediately across Mirelo Studio, Adobe Premiere and DaVinci Resolve plugins, and the Mirelo API.

The release represents a philosophical shift in how generative audio models approach creation. Rather than celebrating raw generation speed, Mirelo argues that the real benchmark for creative tools is iteration—the process by which craftspeople refine their work. By enabling local editing and targeted regeneration, the company is positioning its model as a tool that matches how professional sound designers and film editors actually work, not just how fast a model can produce something from scratch.

• Available immediately across Studio, Adobe Premiere, DaVinci Resolve plugins, and API

Editorial Opinion

This release signals an important maturation in generative audio tools. The shift from 'generate and retry' to actual iterative editing acknowledges a fundamental truth: great creative work isn't about speed, it's about refinement. For a medium where sound carries 50% of a movie's emotional impact, the ability to edit locally rather than regenerate wholesale could genuinely change how creators use AI in their workflows.