With the number of videos and podcasts published on the rise, the need for sound engineering is becoming more prominent and less a business of specialised audio engineers. Recording an interview on the street using state of the art equipment and a soundman usually results in comprehensible audio, although background noise can still make a recording sound bad. The noise is then often removed by a specialised engineer. In contrast, a one-man operation recording with ‘prosumer’ material will often get less ideal output that screams “repair”. It seems that for both pros and prosumers, iZotope RX 3 and RX 3 Advanced are your easiest and most surefire way on the road to audio perfection.
Table of Contents
- iZotope RX 3 no longer only for professionals
- …But pros aren’t left in the cold
- Perfecting your audio with RX 3
- What I couldn’t try out
iZotope has been a leading supplier of sound enhancement and editing technologies for many years. The company’s products, such as its flagship MBIT+ dithering algorithm, are found in many high-end software and hardware products. RX and RX Advanced have been the two most powerful and effective sound repair applications available on the market. I have been reviewing RX Advanced since version 1 and found it unbeatable in terms of sheer power and capabilities. Versions 1 and 2 were less user-friendly than the current new version 3, though. And since last year, iZotope also got competing Sony SpectraLayers Pro breathing down its neck. After reviewing SpectraLayers Pro 2 earlier this year, though, I don’t think the folks at iZotope should worry too much.
I received a test license for RX 3 Advanced three weeks ago and have been trying out the new features since then. Under the hood, the Digital Signal Processing (DSP) has been further improved. In addition, RX 3 and RX 3 Advanced have a brand new interface that is much more user-friendly than the previous versions. It also has a large number of new and improved features – I counted 26 very obvious ones, but with the “under the hood” and not-so-obvious ones (such as better Audiosuite integration), I can easily count to 100.
Many of these make it possible for even the unskilled to quickly remove noise and artefacts from a sound recording. The most spectacular feature in RX 3 Advanced is Dialogue Denoiser. Other new features include Dereverb, Asymmetric Declip, Center channel extraction, Pitch Contour, Radius time and pitch stretching. In RX 3 as well as in RX 3 Advanced you’ll now find multi-file capabilities, multicore processing, the ability to save edits as RX documents, multiple noise selections, phase rotation, etc. RX 3 also has three features that previously were only available in RX Advanced: Plug-in hosting, MBIT+ dithering and SRC sample rate conversion.
One major new feature of RX 3 and RX 3 Advanced is the ability to record audio straight into the application. Together with the new capability to save RX 3 edits into RX documents, this enables you to use RX 3 as a full-blown sound editor. Combined with the Radius time/pitch stretching tool it can even be used to exactly synchronize sound recorded through an external audio recorder with a video recording, provided both are recorded continuously (if you’ve started and stopped during recording, you’ll want to use PluralEyes 3).
One of the advantages of recording straight into RX 3 is that you can mix a mono signal — usually you will use one microphone when recording a screencast or a movie clip in the field — into a perfect stereo channel and immediately start editing the file.
For Indie movie creators, documentary makers and those of us who aren’t sound engineers by education, RX 3 Advanced has an incredibly effective noise remover that doesn’t require you to make decisions about method, fine-tuning or whatever else: the Dialogue Denoiser. The Dialogue Denoiser is a one-click noise removing algorithm that does all the heavy lifting for you. It has been developed with podcasters and video producers in mind. It’s specifically tuned to dialogue and voice recordings. I tried it with a recording of a street interview and the results were astonishing. In environments where you voice really gets drowned in background noises like sirens and other people speaking loudly, there are other modules that will do a better job then Dialogue Denoiser, but an interview in an office or a presentation at a crowded seminar or event will need little else than getting this Filter applied.
A step up in terms of control is the RX 3 Denoise module, which is much more powerful as you can literally muffle noise from even the noisiest recording you can imagine. One situation that regularly occurs is that you don’t have enough noise to create what is known as a ‘noise pattern’. RX 3 (and noise removers in other apps) needs a “noise fingerprint” in order to learn what noise is in your recording. In previous versions, you were limited to one selection of continuous noise. In controlled environments, you can ask people to be quiet for a good 10 to 20 seconds, but more often than not, this is impossible.
I had the pleasure once to shoot an interview where asking this was out of the question, while I had to record with a background of industrial inkjet printers making a lot of hard, squealing sounds that literally interspersed my interviewee’s words. Back then I did what I could to make the whole interview sound good, using RX 2. But as I could only offer the app a noise pattern of five seconds, it wasn’t perfect. With RX 3, I could select multiple non-consecutive areas containing noise to serve as the ‘fingerprint’.
But it gets even better. The Spectral Denoiser can “denoise from anything”, as the iZotope representative told me.
This means you don't need to make a full bandwidth, time based selection of noise to create the fingerprint from anymore. You can actually select elements of the noise profile that appear in-between harmonics, using the selection tools and learn the noise profile from that. There’s even an 'auto complete' feature (the app asks you if you want to complete the profile if it hasn’t got enough information) to fill in the gaps.
I tried this and was expecting synthetic sounding results at best, but in reality, this turned out to be a very powerful audio repairing feature that enables you to remove noise from recordings that I would have found impossible to clean properly without.
Customisable noise removal, however, is not the only professional feature of RX 3. There’s also Spectral Repair, in which the way selections are made is more accurate and has been refined. The resulting processing is more accurate too in the new version.
Professionals will also appreciate iZotope’s Insight now included for free with RX 3. Insight gives you all the metering and analysis tools you will ever need — I don’t think there is a spectrogram, loudness or any other metering tool left out of this package.
Declick and Decrackle, the module that allows you to remove short impulse noises like digital errors, mouth noises, cell phone interference, etc., now offers RX 3 users Declick presets. RX 3 Advanced users also get two advanced parameters: Click Widening and Frequency Skew. The latter tunes the frequency response of the click detector, while the former lets you increase the processing region around each detected click. The result is better handling of low frequency clicks and cleaner results.
Sometimes an audio recording will contain clipped sound. Especially when you’re recording music you may run into the clipping problem. Clipping is to sound what overexposure is to a photo, so it should be avoided if possible. However, if it happens, RX 3 has an enhanced declipping algorithm with the declip areas showing directly in the waveform. This makes it much easier to see to what you’re doing and how much of the signal will be corrected. RX 3 Advanced also has asymmetrical declipping, which further increases accuracy by allowing you to independently set signal top and bottom thresholds.
Another major new feature of RX 3 Advanced that looks like it’s usable in a lot more circumstances than you imagine at first sight is the Dereverb module. Dereverb removes resonance resulting from the room’s sound character. Sound recorded in a church will have much more reverb than when it’s recorded in a small room. However, I found that even in my small office room I experienced some reverb as a result of sound bouncing off of bare walls and windows even. The Dereverb tool let me remove that bit of reverb and gave me a much tighter sound, much more in tune with the business video I was recording. Dereverb can also be used to introduce more reverb, so it’s a module that cuts both ways.
Often, when recording a presentation or screencast with a voice-over recorded afterwards, you will misjudge the timing of what you’re saying with what is happening on the screen. Instead of having to start all over again, RX 3 Advanced allows you to speed up or slow down your sound. The excellent iZotope time and pitch stretch module Radius lets you change timing – and pitch if you need it – without changing the overall sound characteristic/quality (the formant). Time stretching can result in your voice sounding Frankensteinish or helium-inhaled when the formant changes too. As far as I know, there are only two applications/plug-ins that keep the formant intact when time stretching: Celemony’s Melodyne and iZotope’s Radius.
Of course, if you want to change your voice into Donald Duck’s, this tool is your friend too.
Phase rotation is another tool that can turn a good audio file into a perfect one. Contrary to Phase tools in other sound editors that I know, the RX 3 tool allows you to rotate freely. Rotating the phase of a signal changes its peak values but leaves everything else intact. Rotating the audio channels so they become more symmetrical, adds to the signal’s headroom. Except for static rotation, you can have it continuously adapted too.
There were a number of new features I wasn’t able to try out with RX 3 Advanced. Azimuth adjustment, for example, adjusts for channel speed differences, but I couldn’t get my recordings to be out of phase this way. Pitch contour, which corrects for pitch drift over time, seems to target tape or vinyl recordings, and I lack those.
p>RX 3 Advanced is the best sound repair and enhancement application on the market. Even excellent products like Sony’s SpectraLayers Pro 2 can’t compete. RX 3 is faster, offers more powerful tools and in most cases is easier to use. RX used to be a specialist-only tool, but while it still offers everything a professional sound engineer needs, it now offers user-friendly tools for more occasional users. I would recommend RX 3 Advanced for any job that requires pristine sound quality from not-so-pristine recordings.