A trained neural model predicts the front, center and surround content in your stereo audio. You get different tracks for different speakers, ready to mix or download.
Yes. The AI audio upmixer returns separate channel stems you can preview and download in Studio. Choose a model, import your stereo file or audio link, then run Multichannel upmix in Stem Split. Front stereo contains the left and right front channels; sides stereo contains the two surround channels; the center stem carries the center channel. The 5.1 model also returns sub frequencies. Jasper and Gillian return 5.0 without a separate low-frequency stem. For an Atmos master, review the spatial routing and export Combined Stems as Atmos (ADM BW64).
AI upmixing estimates a surround mix from stereo; it cannot guarantee recovery of the original surround channels or Dolby Atmos object positions. A stereo mix combines information from multiple sources, and that process loses spatial detail. Use the stereo to surround upmixer to create channel stems, then listen for bleed or artifacts and adjust the levels in Studio. If you have the original multichannel recordings or Atmos master, use those for a faithful reconstruction.
Use the cleanest available stereo source for a surround upmix. A lossless WAV preserves more detail than an MP3 encoded from the same recording. The AI audio upmixer accepts audio imports and returns separate surround channel stems. Converting an existing MP3 to WAV does not restore information already lost to compression. If your source has obvious noise or compression damage, you can try audio restoration before upmixing, then compare the results before exporting.
Choose the upmix model for your film or music, import your audio, and shape the channel stems in Studio.
Upload audio
Drag & drop or click to browse (multiple files supported)