MiniMax b15-49 variant significantly improves voice cloning quality
Foreforks · reddit · 2026-08-17
While creating adult animated comedies, the author noticed poor audio quality with the default bf16 R2V model during voice cloning. Switching to the b15-49 variant resulted in a major improvement in audio fidelity, with no noticeable difference in reference image handling. Users experiencing audio issues are recommended to explore different variants.
More from Multimodal
- ImageNet should retire: DiT method rankings negatively correlate between ImageNet and text-to-image — YouJiacheng · 2026-10-16
- Open-source anatomy-reveal skill turns Claude into an exploded-view motion design studio — msharmas · 2026-10-16
- Flux 3 Klein vs Flux 2 Klein hands-on: the new base model 'barely usable' for T2I — Both-Rub5248 · 2026-10-16
- USTC paper uses speculative sampling to cut latency in interactive video generation — ustc · 2026-10-16
- 800,000 Gaussian splats run in WebGPU; Casado sees near-infinite scaling — generativist · 2026-10-16
- Nano Banana 2.1 image generation now live in Perplexity Computer for all users — ChrisUniverse · 2026-10-16