Fix, Restore, and Enhance Audio Using iZotope RX 9 Advanced Audio EditoriZotope RX 9 Advanced is an industry-standard audio restoration suite used by audio engineers, post‑production professionals, music producers, podcasters, and forensic analysts. It combines powerful spectral-editing tools, machine-learning–driven modules, and precision metering to repair recordings that would otherwise be unusable. This article covers the core concepts, practical workflows, key modules, tips to get reliable results, and examples of real-world problems RX 9 can solve.
Why RX 9 matters
Many recordings are compromised by noise, clipping, hum, clicks, or problematic room tone. RX 9’s toolkit is designed to solve these issues non-destructively and with high fidelity. Where traditional editing treats noise as something to cut away globally (often causing artifacts), RX 9 gives surgical control: you can visualize sounds in the time–frequency domain and selectively remove problems while preserving the natural character of the source.
Key takeaways
- Spectral editing lets you see and selectively remove unwanted sounds.
- Machine-learning modules automate common repairs with intelligent choices.
- Non‑destructive processing and clip-based modules give precise control and recallable settings.
Getting started: preparation and session setup
-
File management
- Work from a copy of the original file. Always preserve the original.
- Convert lossy formats (MP3) to WAV/AIFF if doing serious repair; start with the best source available.
-
Session setup
- Set the project sample rate to match your source; avoid resampling until necessary.
- Use an appropriate bit depth (24-bit preferred for headroom during restoration).
- Create a session note with the problems you hear (noise, clicks, reverb, hum, distortion).
-
Listen critically
- Use good monitors or quality headphones in a quiet room.
- Identify problematic regions and mark time selections for focused work.
The RX 9 interface and key views
- Waveform view: traditional time-domain editing.
- Spectrogram view: shows frequency content over time — the primary workspace for precision fixes.
- Spectral Recovery view: specialized for frequency-band reconstruction.
- Module Chain and Batch Processor: apply multiple modules and process many files consistently.
Use the spectrogram with an appropriate color map and resolution: higher resolution shows more frequency detail but may be slower. Zoom into problem areas for surgical edits.
Core modules and when to use them
Below are the most commonly used RX 9 modules with practical usage notes.
-
De-noise (Spectral De-noise / Dialogue De-noise)
- Use for steady broadband noise like room hiss, air conditioning, or camera noise.
- Dialogue De-noise is optimized for speech; Spectral De-noise is more general.
- Prefer adaptive mode for variable noise; use learn or manual profiles for consistent background noise.
-
De-reverb
- Reduces room reflections and excessive ambience.
- Use conservatively: overuse makes audio sound processed or “swirly.” Work in passes with subtle reduction and compare bypassed sections.
-
De-click / De-crackle
- Removes transient clicks, pops, and vinyl crackle.
- Adjust sensitivity and click width; use low sensitivity first to avoid removing consonants in speech or transients in music.
-
De-hum
- Removes tonal hums and mains interference (⁄60 Hz and harmonics).
- Use automatic hum detection or manually specify the fundamental; remove only necessary harmonics.
-
Repair Assistant
- Runs automated analysis and suggests module chains and settings for common scenarios (music, dialogue, broadband noise).
- Good starting point, then refine manually.
-
Spectral Repair
- Excise or attenuate isolated noises (sneezes, chair creaks, camera pings, footsteps).
- Use the “Attenuate” or “Replace” modes depending on context; “Replace” can paint over gaps using neighboring spectral content.
-
Mouth De-click and Breath Control
- Specifically target mouth noises, lip smacks, and excessive breaths in dialogue recordings.
- Breath Control can automate breath reduction without chopping phrases.
-
Gain, EQ Match, and De-bleed
- Gain for level adjustments; EQ Match to match timbre between takes.
- De-bleed helps remove spill between microphones (useful in multi-mic sessions).
-
Spectral Recovery
- Reconstructs missing high-frequency content from degraded files (e.g., low-bitrate audio).
- Useful when restoring archival or heavily compressed audio.
-
Ambience Match
- Recreates consistent background ambience across edits and ADR to maintain continuity between cuts.
-
Composite View and Loudness modules
- Composite View lets you view and edit multiple files simultaneously.
- Loudness and Level modules help deliver compliant broadcast-level audio.
Practical workflows: common problems and step-by-step fixes
Workflow 1 — Clean noisy dialogue (location recording)
- Open file and listen through to mark problem regions.
- Run De-hum if there is noticeable mains hum; remove only harmonics you detect.
- Use Spectral De-noise or Dialogue De-noise:
- Learn a noise profile from a section of silence, or use Adaptive mode for fluctuating noise.
- Reduce conservatively; preview bypassed result frequently.
- Remove clicks/pops with De-click.
- Use Mouth De-click and Breath Control for artifacts from speech.
- Apply De-reverb if the room is too reflective; start light, compare.
- Final EQ and level adjustments; check in context with other tracks.
Workflow 2 — Fix music with transient noises (e.g., a camera shutter)
- Use Spectral Repair to isolate the transient and choose Attenuate or Replace.
- If the transient overlaps important material, use multiple narrow time–frequency selections and small-step processing.
- Use De-click for residual artifacts.
- Restore spectral balance with EQ Match if necessary.
Workflow 3 — Restore archival/low-bitrate audio
- Use Spectral Recovery to rebuild upper harmonics lost to compression.
- Run RX’s De-noise modules to control broadband noise.
- Use De-crackle for surface noise (if from vinyl).
- Final limiting and minor EQ for presentation.
Tips for preserving audio quality and avoiding artifacts
- Work incrementally with small reductions rather than one heavy pass.
- Frequently toggle bypass to compare processed vs original.
- Use spectral selection feathering and smaller FFT sizes for transient accuracy; larger FFT sizes for smoother noise reduction at lower frequencies.
- When in doubt, use less. Over-processing causes pumping, phasing, or “vocal plasticity.”
- Make non-destructive saves and keep module chains with notes for repeatability.
- Use high-quality monitoring and check outputs on multiple reproduction systems (headphones, monitors, laptop speakers).
Batch processing and automation
- Use the Module Chain to create standardized processing chains (e.g., De-hum → Dialogue De-noise → De-click → De-reverb → Gain).
- Batch Processor runs the chain across many files consistently — ideal for dialog editors working on multiple takes or episodes.
- Use Batch with caution: run on a subset first to verify settings.
Integrating RX 9 into your DAW/post pipeline
- RX 9 can be used as a standalone app for surgical spectral editing, then exported to your DAW.
- Many modules are available as plugins (VST/AU/AAX). Use plugin versions for real-time prep, but do heavy spectral repairs in the standalone for better tools and visual feedback.
- For dialog editing in video: use RX’s video import feature to work with picture, or export cleaned audio back to the NLE/DAW.
Example before/after scenarios
- Location interview with constant HVAC noise: Dialogue De-noise (learned profile) reduced hiss by 10–18 dB while preserving intelligibility; low-level De-reverb added clarity.
- Film clip with camera motor whine at 120 Hz: De-hum removed the 60 Hz fundamental and harmonics; Spectral Repair cleaned occasional camera pings.
- Podcast recorded in untreated room: Adaptive De-noise for broadband room tone, Breath Control to reduce breath prominence, and a light high-shelf to add clarity.
Common mistakes and how to avoid them
- Applying maximum reduction settings immediately — instead, use conservative settings and multiple passes.
- Ignoring the spectral view; many problems are visible and easier to fix visually.
- Relying solely on automatic assistants without manual verification.
- Processing only a short sample and assuming it generalizes; always test across the whole file.
Final checks and delivery
- Listen in mono to ensure no phase issues were introduced.
- Check levels against target loudness (LUFS) for the intended platform.
- Export high-quality masters (WAV/AIFF) and keep an edited, mixed, and raw-original version for archive.
- Document the chain and settings for future recall and consistency.
iZotope RX 9 Advanced is a deep, capable suite that rewards disciplined listening and conservative, iterative processing. With practice, it turns recordings that once seemed unusable into professional deliverables while preserving the natural character of voices and instruments.
Leave a Reply