DiffSinger vs SightSinger

DiffSinger is a free, open-source singing voice synthesis system built around a shallow diffusion mechanism. The maintained version adapts generated audio to 44.1 kHz and incorporates updated acoustic models plus faster diffusion sampling methods. Variance models and parameters provide control over pitch, energy, breathiness, and other vocal characteristics. The linked MakeDiffSinger project supplies dataset-building pipelines, including recording-slicing and labeling tools. Its guide describes preprocessing data, training models, and running inference on DS files with variance or acoustic models. Installation requires Python 3.10 or later, with PyTorch 2.4.0 or later recommended. ONNX is the deployment format. The repository links OpenUtau and DiffScope as production or deployment projects; DiffScope is an editor powered by DiffSinger and remains under development. Users prepare their own data and assets. The project prohibits generating another person’s voice without consent and includes a utility for removing speaker embeddings from checkpoints before distribution. Support channels include tutorials, GitHub issues and discussions, QQ, and Discord.

SightSinger turns MusicXML sheet music into singing audio demos through a natural-language chat workflow. It is designed for score-led rehearsal and demo work, not writing songs from text prompts or full studio production. Choir and worship leaders can make section references and part-learning demos, including split SATB parts; music teachers, songwriters, arrangers, and independent singers are also named users. The site recommends.xml or.mxl files exported from MuseScore, Finale, Sibelius, Dorico, or Logic Pro. Its multi-track player supports score highlighting, repeats and endings, per-track mute, solo and volume, instrument sound changes, and mix downloads. Users can add movable-do or fixed-do solfege, including la-based or do-based minor options. SightSinger says it detects lyric language and supports Mandarin Chinese, Cantonese Chinese, Japanese, Thai, English, Spanish, French, Italian, Portuguese, and European Portuguese. Listed voices have voice-specific language and style support. Vocal audio costs one credit per 30 seconds for each vocal part each time it is rendered; MIDI accompaniment uses one credit per two minutes per uploaded score. The web service offers a free plan with eight monthly credits and paid plans from $6 per month.