Download VoiceProcessor-1.1.0-win-x64.zip (936 KB), unzip, run VoiceProcessor.exe.
Needs the .NET 10 Desktop Runtime.
The self-contained zip bundles the runtime instead — 58 MB, nothing to install.
New in 1.1.0
Batch processing. Point at a folder, get a queue, process one file after another. In the
app that is the Folder... button; on the command line:
voiceproc batch --input-dir .\clips --output-dir .\out --voice Robot --suffix "-robot"
Outputs are named after their input plus the suffix. The exact total time is shown before
it starts — rendering is 1:1 with real time, so a folder's combined length is the real
cost. A file that fails does not discard the ones already done.
Voicemod's monitoring is no longer touched. 1.0.0 switched "hear my voice" off before
every render. It now stays exactly as you set it, so you can hear what is being processed.
Voice filters by name. --voice Robot instead of a GUID.
MP3 presets lead with VBR, each stating its approximate bitrate. VBR 2 — high (~190 kbps) is the new default: transparent for speech and appreciably smaller than the
320 kbps CBR it replaces.
The client key is masked in the app, so it no longer appears in screenshots.
The download is 68x smaller — 936 KB against 61 MB. Most of that was the .NET runtime,
plus 13 folders of translated framework exception strings this app never surfaces.
Requirements
Windows, the Voicemod desktop app running locally, a Voicemod Control API key, and
VB-Audio Virtual Cable. QUICKSTART.txt covers the setup —
two settings inside Voicemod, once.
ffmpeg is optional and only needed for Ogg, Opus, FLAC, and AAC output.
Still not exercised
Ogg / Opus / FLAC / AAC output, latency calibration (manual — see docs/audio-routing.md),
and long batches under load.