NEUTRINO vs SoulX-Singer

NEUTRINO is an AI singing voice generator that synthesizes vocals from sheet music. Users provide lyrics and melody; the system estimates vocal timing, pitch, voice quality, and breathiness, then a vocoder produces the singing. The project says its models draw vocal traits from recordings made with performers' permission or audio released by public institutions, rather than existing pretrained models. The FAQ describes Muon voice retakes and adjustable quality and processing speed, with advanced, standard, and elements modes. Advanced and standard output is 48 kHz, 16-bit mono WAV; elements output is 24 kHz, 16-bit mono WAV. It is available for Linux, macOS, web, and Windows. Users download the software and add a voice library. NEUTRINO is free for general music and audio production, but embedding the software or using generated audio inside an application requires separate permission. Character-specific terms also apply to ordinary audio-file uses and vary by character and asset type. Users cannot currently create their own voice libraries with a released framework, according to the FAQ.

SoulX-Singer is a singing voice synthesis model that generates voices for singers it has not been tuned on. Pitch, rhythm, and expression can be guided by melody-conditioned F0 contours or score-conditioned MIDI notes. Its voice-conversion system changes raw singing audio to a target singer’s voice while retaining melody, rhythm, and lyrics, without requiring lyric or MIDI transcriptions. It supports Mandarin Chinese, English, and Cantonese, plus singer-timbre cloning, cross-lingual synthesis, and lyric editing while retaining natural prosody. Preprocessing tools handle vocal separation, dereverberation, F0 extraction, voice activity detection, and lyric and note transcription. Users can edit MIDI metadata, including lyrics, phoneme alignment, pitches, and durations, before synthesis. Soul-AILab provides generation and conversion demos and a separate MIDI editor through Hugging Face Spaces; the repository also supports local inference. Code and model weights use the Apache 2.0 license. Maintainers caution that automatic preprocessing may misalign singing with lyrics and notes, so manual correction may be needed.