MusiCraft Stem Splitter vs Open-Unmix

MusiCraft Stem Splitter is a browser-based tool for separating audio into stems. Its 66 separation modes cover vocals, drums, bass, instruments, choir and sound effects, with options including vocal-and-instrumental, 4-stem, 6-stem and full-band splits. The listed stem types also include guitar, piano, strings, brass, woodwind, saxophone, speech, music and other instruments. Users can upload MP3, WAV, M4A or FLAC files, or select a track from My Music. The in-browser player lets them preview stems and solo or mute them before downloading. Stem splitter mode pages limit uploads to 100 MB and eight minutes. The Free plan includes two editing runs per month and preview playback, but downloads are locked. Paid plans start at $7.49/mo. In Speech / Music / SFX mode, processing costs five credits per minute and output is 320 kbps MP3; 16-bit WAV is available on Studio and Business. The privacy page says uploaded audio is processed securely, not shared with third parties, and stems are deleted after download. The privacy policy also says uploads and generated content may be processed by AI, storage and infrastructure providers.

Open-Unmix is a PyTorch reference implementation for music source separation, aimed at researchers, audio engineers, and artists. Its ready-to-use models split pop music into vocals, drums, bass, and other instruments; a separate model, umxse, is for speech enhancement and was trained on the 28-speaker Voicebank+DEMAND corpus. Pre-trained choices include umxl, umxhq, umx, and umxse, with spectrogram variants also available. Each source model uses a three-layer bidirectional LSTM, a design the project says cannot run online or in real time. The code supports reproducing results, developing architectures, and adding data, while training is documented separately from the package. Install through pip or use the Docker container; pre-trained models can be loaded in Python or through torch.hub, and a Colab notebook supports online experimentation without local setup. The command line handles WAV, FLAC, and OGG, not MP3; stempeg can extend formats, including STEMS export. The software is MIT-licensed; umxl weights are non-commercial under CC BY-NC-SA 4.0.