Fast Fixes with Voxengo Deconvolver: Step-by-Step Cleaning Techniques
Voxengo Deconvolver is a specialized tool for dealing with unwanted reverb, room tone, and other impulse-response-based artifacts in audio recordings. This guide gives concise, practical steps to quickly clean audio using Deconvolver, plus tips to speed up your workflow and get reliable results.
1. Quick overview — when to use Deconvolver
- Best for: Removing or reducing unwanted room reverb, mic bleed with clear impulse characteristics, and repetitive echoes captured as impulse responses.
- Not ideal for: Removing broad-spectrum noise (use noise reduction plugins), or heavily distorted audio.
2. Prepare your audio
- Isolate a short noise/reverb sample: Find a section of the recording that captures the unwanted decay or impulse (a pause, a transient followed by reverb, or a test sweep/impulse if available). Aim for 0.5–3 seconds.
- Clean the sample: Trim silence at the start/end so the sample begins exactly at the impulse onset. Apply a high-pass at ~50–80 Hz if low-end rumble is present. Normalize to -1 dB to maximize resolution.
3. Create the deconvolution impulse
- Open Voxengo Deconvolver and load the isolated sample as the “impulse.”
- Set the FFT size: start with 32k or 64k for higher resolution on reverb tails; use 8k–16k for quicker results on short impulses. Larger FFT = better fidelity, slower processing.
- Choose deconvolution mode: use “Regularized” (or similar wording in the plugin) to avoid extreme EQ boosts — it preserves stability. If available, enable a small amount of spectral smoothing to prevent artifacts.
4. Generate and inspect the inverse
- Run the deconvolution to generate an inverse impulse.
- Listen to the inverse alone — it should sound like a short clicky/suppressed burst rather than musical content. If it sounds like noise or has weird tonal spikes, increase regularization or apply gentle EQ to the original impulse and regenerate.
5. Apply the inverse to the affected audio
- Load the full affected recording into Deconvolver’s processing window.
- Apply convolution with the generated inverse impulse. Use a bypass toggle to compare processed vs. original.
- If the result introduces phasey artifacts or metallic ringing, reduce the intensity (mix dry/wet) or reduce deconvolution strength/regularization parameters.
6. Fine-tune and rescue artifacts
- Blend Dry/Wet: Often a 60–80% processed/40–20% dry mix gives a natural result while still reducing reverb.
- Use multiband processing: If deconvolution overly affects highs or lows, apply the inverse only to mid frequencies (split bands) or follow with gentle multiband compression.
- Spectral smoothing: Re-run with slightly more smoothing if you hear tonal zippering.
- Manual repair: For stubborn spots, create shorter, targeted impulses from those sections and process them individually.
7. Speed tips for quick fixes
- Start with a medium FFT (16k) for a balance of speed and quality.
- Save inverse impulses as presets for similar rooms/mics to reuse later.
- Batch process multiple files by using the plugin in a DAW with offline rendering.
- Keep default regularization slightly higher to avoid repeated tweaking.
8. Common pitfalls and fixes
- Over-deconvolution (hollow/phasey sound): Lower strength, add regularization, or blend with dry signal.
- Ringing or metallic artifacts: Increase spectral smoothing, reduce FFT size, or apply light de-essing/high-frequency roll-off.
- Poor impulse capture: Re-record a proper test impulse (hand clap, starter pistol, sine sweep) if possible.
9. Final checks before export
- A/B test with original on multiple monitors/headphones.
- Check mono compatibility and phase correlation.
- Apply final gentle limiting or EQ only if needed for tonal balance.
Quick workflows like this let you remove or reduce unwanted reverb and room artifacts without heavy-handed processing. Save impulse presets, keep reasonable FFT/regularization defaults, and rely on dry/wet blending to preserve natural character while cleaning fast.
Leave a Reply