iZotope RX 9 introduces Dynamic De-hum, a new way to remove complicated hum and buzz from dialogue and other recordings. Available in RX 9 Standard and Advanced, the tool can handle noise with many harmonics while keeping speech mostly intact.
It is a useful addition for filmmakers, video producers, and anyone cleaning difficult production audio.
Key takeaways
- Dynamic De-hum is available in RX 9 Standard and Advanced, but not the Elements version.
- It is designed for complex hums with many frequency lines and harmonics.
- The tool learns a noise profile from a selected section of audio.
- A longer, cleaner sample usually gives better results.
- Multiple short selections can be combined when one long sample is not available.
- Start with the default settings, then adjust sensitivity, bands, and filter width as needed.
Why complex hum is difficult to remove
A simple hum often has one main frequency and a few harmonics above it. RX’s traditional Static De-hum tool can work well in that situation. It identifies those problem frequencies and applies filters to reduce them.
The trouble starts when the recording contains a more complicated noise pattern. A fan, electrical device, recorder, compressor, or other equipment may create a large group of hum and buzz frequencies. In a spectrogram, these appear as horizontal lines stretching across the recording.
In the example, the noise reached well into the high frequencies. Using the older adaptive or static approach did not fully remove the hum. It also introduced a ringing sound around the dialogue.
That is not a great trade-off. Removing one problem while adding another can make the recording harder to use.
How Dynamic De-hum works in RX 9
Dynamic De-hum uses a selected piece of audio to learn the noise profile. The best selection contains the hum but does not include dialogue, music, or other sounds you want to preserve.
The basic workflow is simple:
- Open the De-hum module in iZotope RX 9.
- Choose Dynamic instead of Classic or Static mode.
- Select a clean section where the hum is present by itself.
- Click Learn.
- Preview the result and adjust the settings if needed.
- Render the repair when the dialogue sounds clean.
The spectrogram is especially helpful here. It lets you see the noise before and after processing. Once Dynamic De-hum is applied, many of the horizontal lines should disappear, while the speech remains largely unchanged.
Understanding the main settings
Dynamic De-hum does not have an overwhelming number of controls, which makes it fairly approachable. Still, each setting affects the result.
Sensitivity controls the strength of the reduction. Higher values remove more noise, but they may also increase artifacts. Starting at five gives you a reasonable middle ground.
Bands refers to the number of filters RX can use across the frequency range. If the hum extends into higher frequencies, increasing the number of bands may help. It makes sense to begin at 128 and add more only when necessary.
Filter Q changes how wide or narrow the filters are. A higher value creates narrower filters, while a lower value makes them wider. The default value of 1,000 is a useful place to start.
The high-pass and low-pass controls limit the frequency range being processed. Unless there is a specific reason to narrow the range, leave them at the outer edges so RX can address the full noise profile.
How much audio should you use for learning?
A longer sample generally gives Dynamic De-hum more information. In the demonstration, a five-second selection produced the strongest result. It removed much of the hum and buzz without making major changes to the dialogue.
However, you will not always have five seconds of clean noise. Production recordings often contain only brief gaps between words. A one-second sample can still work surprisingly well, although the result may not be quite as clean.
RX also lets you select multiple areas. Hold the Shift key while selecting additional sections, then click Learn. This is useful when the available noise-only moments are short or spread throughout the clip.
A practical approach is to gather several small sections rather than relying on one extremely short sample. The more representative the selections are, the better the learned profile is likely to be.
What to expect from the final repair
Dynamic De-hum is focused on hum and buzz. It will not necessarily remove every sound in the recording. Fan noise, mechanical noise, or compressor noise may remain after processing, and those problems may require other RX modules.
There may also be a small amount of artifacting, especially when sensitivity is pushed too high. Always preview the dialogue carefully. Look at the spectrogram, but trust your ears too. A recording with slightly visible noise can sound more natural than one processed so heavily that speech becomes metallic or hollow.
For podcast and video workflows, this makes Dynamic De-hum a useful repair step before equalization, compression, or loudness processing. It can be particularly helpful when a location recording has a complicated electrical or mechanical noise problem that traditional De-hum cannot handle cleanly.
Staff Report
Final thoughts
iZotope RX 9’s Dynamic De-hum is a practical upgrade for difficult audio repairs. Its main advantage is the ability to learn a complex hum profile instead of relying only on a simple set of fixed frequencies.
Start with a clean sample, use the default settings, and make small adjustments. When possible, provide RX with several seconds of noise-only audio. If that is not available, multiple short selections can still produce a solid result. For creators working with dialogue, film sound, podcasts, or live video, it is a straightforward tool worth adding to the cleanup process.
