ACE-Step Studio 3.8.4

2026-11-05 — 3.8.4

Fixed

• The later sections no longer come out as sound-alike nonsense on newer NVIDIA cards: the 3.8.0 move to a newer ggml brought a CUDA flash-attention kernel that is numerically wrong on Blackwell-generation cards (compute capability 10.0 and up, such as the RTX 5060 Ti), so a song started right and drifted into "sounds like singing but means nothing" later. The studio now turns flash attention off automatically on these cards; other cards are untouched and their speed does not change.
• A finished song is no longer lost when the system temp folder cannot be written: the FLAC was written through a temporary file, and a temp folder gutted, blocked or named in an awkward code page lost the song at the last step. It is now encoded in memory.
• Karaoke times original songs whose voice is buried under the mix: when nothing is heard in the mix and Mel-Band RoFormer is installed, the voice is lifted out and timed on its own, so a freely generated original is timed where the mix alone reported nothing - now with cloud recognition too.
• Your chosen maximum duration is kept when a new Create loads a request: reuse, an example or the writing assistant no longer writes its own shorter duration over the one you saved.
• The autogenerated lyrics fall back to your language instead of Russian: under "Automatic" the lyrics follow the studio's own language when the request does not settle one.
• Each changed component on the update screen has its own reason: the audio.cpp engine's CUDA libraries no longer share the karaoke description; they get their own line.