What is VOXSPLIT PRO? +
VOXSPLIT PRO is a local-processing AI voice separation plugin available as AAX for Pro Tools, VST3 for other DAWs, and Standalone.
Is it a real-time plugin? +
VOXSPLIT PRO is designed around non-real-time local processing and editorial review rather than live insert-style use.
Does it work inside Pro Tools? +
Yes. The AAX version is built specifically around Pro Tools timeline capture, local processing, audition, and timeline return.
Is VOXSPLIT PRO available outside Pro Tools? +
Yes. In addition to the AAX plug-in for Pro Tools, VOXSPLIT PRO is available as VST3 for other compatible DAWs and as a Standalone application. The VST3 and Standalone versions use a file-based workflow: load an audio file, process it with the same AI engine and profiles, and export the result, rather than the Pro Tools timeline capture/render flow.
Who is it for? +
VOXSPLIT PRO is intended for dialogue editors, sound editors, mixers, and post-production professionals.
What do STANDARD, PRECISION, and RESCUE mean? +
STANDARD is the default everyday mode. PRECISION is a more careful alternative for ambiguous material. RESCUE is a fallback profile for harder problem cases.
Is it made for music stem separation? +
The product is primarily focused on dialogue and voice-isolation workflows in post-production.
How can I get VOXSPLIT PRO? +
Visit the
pricing page to see the current purchase and trial options.
How will updates work? +
Minor updates (bug fixes, x.x versions) are included at no extra cost. Major updates (new x.0 versions) may be subject to separate pricing. Updates are distributed directly by Limited Edition Audio.
How many computers can I use? +
A single perpetual licence allows use on one computer at a time. You can move the activation between enabled iLok USB and iLok Cloud locations through iLok License Manager.
Which version should I install, Apple Silicon or Intel? +
Apple Silicon with macOS 14 or later: choose Silicon. Apple Silicon with macOS 12.7 or 13.x: choose Silicon Legacy. Apple Silicon running your DAW under Rosetta 2: choose Rosetta. Intel Mac: choose Intel.
Is processing slower on Intel than on Apple Silicon? +
Yes. On Intel Macs, the AI engine runs on CPU only, without GPU acceleration, so all processing is noticeably slower than on Apple Silicon, whichever profile is used. Apple Silicon, particularly M2 and later, is recommended for the best performance.
What are the system requirements? +
Minimum: macOS 12 Monterey, Apple Silicon or Intel, 8 GB RAM, 4 GB storage, Pro Tools 2022.x, a VST3-compatible host, or Standalone. Recommended: macOS 14 Sonoma, Apple Silicon M2+, 16 GB RAM, 8 GB storage, Pro Tools Ultimate 2023.
Does it require an internet connection? +
Audio processing runs locally and your audio is not uploaded. Internet may be required for licence activation or validation, and an active connection is required while using iLok Cloud.
Why is the first processing run slow? +
On first launch after installation, the AI engine initialises its model, which takes longer than usual. Subsequent processing runs are significantly faster.
Can it separate more than two voices? +
The AI engine is designed and optimised for separating two distinct voices. Beyond two simultaneous voices, results are not guaranteed: separation will be imperfect or unusable depending on content density.
The plugin does not appear in AudioSuite. What should I do? +
Make sure the installation completed successfully and that Pro Tools has been fully relaunched. Verify that the installed version (Intel or Apple Silicon) matches your configuration.
The source does not load after clicking CAPTURE. +
Make sure an active timeline selection exists in Pro Tools before clicking CAPTURE. The selection must be longer than 1 second. Click CAPTURE again after verifying the selection.
There is no audio in Preview. +
Make sure a source is selected with SELECT in the Timeline Return panel. Then start Pro Tools Preview after clicking SELECT.
The rendered file does not appear in the timeline. +
Make sure a track and time range are selected in Pro Tools before clicking RENDER VOICE A or RENDER VOICE B in the plugin.
Processing is very slow. +
Use the STANDARD profile. PRECISION and RESCUE are slower by design. Verify the machine has sufficient resources (RAM, CPU). Very long selections (over 60 seconds) will take longer to process.
The result contains artefacts. +
Try a different profile. Verify the selection is not too short. On extremely complex content, some artefacts are inherent to current AI technology. Start with STANDARD, then try PRECISION, then RESCUE.
How can I contact you? +