THE STUDIO
Session help ↗Find your frequency.
Your audio is processed locally on your device by default. VOID Remastered is prerelease software.
EARLY ACCESS
VOID Remastered is currently in Early Access. Try one complimentary guest master without creating an account. Invited Early Access testers can sign in for unlimited local mastering during the testing period. VOID Remastered is prerelease software. Features and mastering results may change before commercial release. Your audio is processed locally on your device by default.
A starting point for your sound.
Auto builds a fresh chain from your source, renders it, then checks the result. Local measurement-based suggestions; no AI service or uploads. Undo restores your previous chain.
Standard retains the original processor behavior. Studio filters around the clipper to reduce aliasing and adds a small preview delay. The optional ceiling check applies whole-render attenuation after processing; it is not a certified true-peak limiter and is not applied to live Processed playback. Play Rendered auditions the checked output.
Choose a style, then listen to the result. Auto does not repair an already clipped recording.
Run Auto or export to prepare a measured master for comparison. Matching affects audition only.
A fresh master. Your track’s character.
Sign in as a WordPress administrator to use AI. Local Auto is available without signing in.
Generate a remaster from the track profile, then compare and fine-tune the processing chain. Leave the direction blank for a balanced, dynamics-preserving starting point.
WAV genre/date tags fill these fields when available. Correct or clear them before sharing.
The tonal profile is measured locally when you request AI. Missing tags are left unknown.
AI receives these genre/date fields, source bit depth/encoding, duration, channels, sample rate, loudness/range, peaks, crest factor, correlation, over-full-scale percentage and sampled tonal-band energy. No audio, waveform, filename or other tags are sent. It has not listened to the recording.
AI remaster enables Studio processing and a checked output ceiling of -1 dBTP (or your stricter ceiling). It renders and checks before applying; Undo restores your previous settings. Manual controls remain available below.
Ready for one track at a time. Local Auto above works without an AI connection.
Loudness & true peak
Source
Rendered output
Render the chain to measure its output.
History: grey = source momentary, blue = source short-term, orange = rendered short-term. M requires 400 ms; S and LRA require 3 s. LRA is provisional below 60 s.
K-weighted, gated programme loudness. True peak is estimated using 4× interpolation; the limiter still controls sample peaks only. These measurements do not automatically change gain.
Processing chain
Adjust controls during Original or Processed playback. Bypass is per module. Reset Chain restores neutral settings.
Clip uses 4× oversampling in Studio mode, with filtering near Nyquist; filter ringing can exceed the clip threshold. Limit is a linked 5 ms lookahead sample-peak limiter. The optional final ceiling check measures estimated true peak and attenuates the rendered output. EQ frequencies are capped at 45% of the processing rate. Width changes require mono/stereo.
Preview meters idle.
Live readings follow what you hear, including A/B and parameter changes. They restart with playback. Whole-track readings are above.
Live spectrum
Stereo scope
Spectrum is a windowed FFT display, not a loudness measurement. Scope is available for stereo only.
No processed render yet.
Original / Processed switches at the current position with a short crossfade. Live preview requires HTTPS and AudioWorklet. Preview uses the device rate; export retains the source rate and channels. Use Original / Master with Match loudness for level-matched comparison. Matching changes listening gain only, never exports.
Export applies the current chain at its actual level; comparison loudness matching is never exported. Studio WAV preserves internal samples. PCM WAV uses optional dither; PCM and MP3 reject out-of-range encoder input. MP3 is lossy and may produce different decoded peaks. 32-bit PCM input may lose low-order precision in Float32.
MP3 encoding uses LAME through the bundled lamejs encoder. No audio is uploaded.