By Claudio Santos
Sound editing has many different faces. It is part of big-budget blockbuster movies and also an integral part of small hobby podcasting projects. Every project has its own sound needs. Some edit thousands upon thousands of sound effects. Others have to edit hundreds of hours of interviews. What most projects have in common, though, is that they circle around dialogue, whether in the form of character lines, interviews, narrators or any other format by which the spoken word guides the experience.
Now let’s be honest, dialogue is not always well recorded. Archival footage needs to be understood, even if the original recording was made with a microphone that was 20 feet away from the speaker in a basement full of machines. Interviews are often quickly recorded in the five minutes an artist has between two events while driving from point A to point B. And until electric cars are the norm, the engine sound will always be married to that recording.
The fact is, recordings are sometimes a little bit noisier than ideal, and it falls upon the sound editor to make it a little bit clearer.
To help with that endeavor, Audionamix has come out with the newest version of their IDC (Instant Dialogue Cleaner). I have been testing it on different kinds of material and must say that overall I’m very impressed with it.Let’s first get the awkward parts of this conversation out of the way. First, let’s see what the IDC is not.
– It is not a full-featured restoration workstation, such as Izotope RX.
– It does not depend on the cloud like other Audionamix plugins.
– It is not magic.
Honestly, all that is fine because what it does do, it does very well and in a very straightforward manner.
IDC aims to keep it simple. You get three controls plus output level and bypass. This makes trying out the plugin on different samples of audio a very quick task, which means you don’t waste time on clips that are beyond salvation.
The three controls you get are:
– Strength: The aggressiveness of the algorithm
– Background: Level of the separated background noise
– Speech: Level of the separated speech
Like all digital processing tools, things sound a bit techno glitchy toward the extremes of the scales, but within reasonable parameters the plugin makes a very good job of reducing background levels without gargling up the speech too noticeably. I personally had fairly good results with strengths around 40% to 60%, and background reductions of up to -24 dB. Anything more radical than that sounded heavily processed.
Now, it’s important to make a note that not all noise is the same. In fact, there are entirely different kinds of audio muck that obscures dialogue, and the IDC is more effective against some than others.
– Constant broadband background noise (air conditioners, waterfalls, freezers): Here the IDC does fairly well. I couldn’t notice a lot of pumping at the beginning and end of phrases, and the background didn’t sound crippled either.
– Varying broadband background noise (distant cars passing, engines from inside cars): Here again, the IDC does a good job of increasing the dialogue/background ratio. It does introduce artifacts when the background noise spikes or varies very abruptly, but if the goal is to increase intelligibility then it is definitely a success in that area.
– Wind: On this kind of noise the IDC needs a little helping hand from other processes. I tried to clean up some heavily winded dialogue, and even though the wind was indeed lowered significantly so was the speech under it, resulting in a pumping clip that went up and down following the shadow of the removed wind. I believe with some pre-processing using high pass filters and a little bit of limiting the results could have been better, but if you are emergency buying this to clean up bad wind audio I’d definitely keep that in mind. It does work well on light wind reduction, but in those cases as well it seems it benefits from some pre-processing.
I am happily impressed by the plugin. It does not work miracles, but no one should really expect any tool to do so. It is great at improving the signal-to-noise ratio of your sound and does so in a very easy-to-use interface, which allows you to quickly decide whether you like the results or not. That alone is a plus that should be kept in consideration.
Claudio Santos is a sound mixer and tech aficionado who works at Silver Sound in NYC. He has worked on a wide range of sound projects ranging from traditional shows like I Was Prey for the Animal Planet and VR experiences like The Mile-Long Opera.