Three things I know to be true: Nobody is perfect, I am an editor first and videographer by necessity, and when working with a new client you need to absolutely nail the first project.
Earlier this year, I was lucky enough to work with The Nature Conservancy of Rhode Island on a video highlighting the recovery of a protected forest that had recently suffered its largest fire in 60 years. I was feeling great. It was my first project since leaving my in-house job, it was only 5 business days since my last day at that job, it was a project near and dear to my heart. This forest is exactly 4 minutes by car away from my childhood home, so I was thrilled to be giving back to my community.
You’ll imagine my momentary panic when at the end of the day, copying my footage down I realized the audio was quiet.
I’d been too excited. We were outside. The batteries were low. I had no extra set of hands. The mic was too far away because the lens was too long. These are my excuses, and this intro is my mea culpa. But at the end of the day it never became an issue because, trite as the expression may be, I knew how to fix it in post.
In this blog post I’ll be going over three tricks for getting fantastic audio that I use on almost every project, and a new one that could well become one of my biggest time savers.
Use Compressors, Use Expanders
At its most basic, a compressor is an audio tool that takes the loud parts and makes them quieter. If you picture the waveform of your audio as a literal wave, a compressor compresses the peaks of that wave to be shorter. This makes it so that when someone suddenly begins speaking very loud, their audio doesn’t clip.
An expander is the exact opposite. When there’s audio beneath a certain threshold of volume, the expander increases that volume. So whereas a compressor lowers the peaks of the waves, an expander raises the valleys.
To become an expert in compression and expansion you’d need to learn all about attack (how quickly the effect engages), release (how quickly it disengages), threshold (when it kicks in), ratio, limiters, gates, and so on. But for my purposes I can typically get everything I need out of Premiere Pro’s “Dynamics Processing” effect.
Dynamics Processing simplifies the process because it gives you a graph where you can control just about everything. The x axis is how loud the incoming or “raw” audio is, and the y axis is how loud you want it to be. By creating and dragging points on the graph, you tell Premiere how much and where you want your compressor and/or expander to operate.
Noise Reduction
I won’t pretend to know exactly how Adobe’s “Denoise” filter works at a mathematical level, but I simply could not be a bigger fan. You can apply it flat across all frequencies, but it has settings to target the noise reduction on the lows, highs, lows and highs, or mids. This effect has saved me so many times on windy shoots or talking heads when there was a fan in the background. In the case of my video for The Nature Conservancy, my weak audio signal meant that in order to get intelligible dialogue, I needed to crank my expander past what you’re “supposed to” use it for and to the point where it was introducing additional hiss. After a subsequent application of “Denoise,” that hiss went away leaving me with now-usable audio.
Let me be clear. By no means do I advocate using these tools together as a crutch for not learning how to get good levels. But, stuff happens, and the editor being the last creative in the pipeline, sometimes you need to just make it work.
EQ the Dialog against the Music
Getting dialog to sound right is like getting skin tones to look right. We all spend all day listening to the human voice, so the audience will be able to tell when it's been tampered with too much – even if they don’t know why.
Vowel sounds produced by the human voice typically fall somewhere in the range of 250 to 2000 hertz, with unvoiced consonant sounds (think S’s and T’s) much higher but carrying less of the volume. For men these values can be a little lower, for women they can be a little higher. I can hear those of you who didn’t just click away accosting me now. “Kyle, this is boring, I came here for video wisdom not physics class!” Well, you can use this knowledge to your advantage especially when you have a music track, by creating a hole in the mix. Here’s why that’s a cool thing:
If you know which frequencies the audio you want to hear is and you have another track to drown out everything else, you don’t need to go heavy on your noise reduction in the first place.
On your next video, try applying an EQ to the dialog track and lower the volume for frequencies outside the range of the speaker’s voice. Once you have an idea of where in the mix your speaker’s dialog sits, take a look at your music. If you EQ the music bed so that it's louder in frequencies outside of where there’s any dialog, the music will compete with any noise you’re trying to hide in those frequencies. This also allows you to raise the volume of the music higher than you’d be able to otherwise without it competing with the dialog. Your audience will spend more time listening to the track, and less time thinking about whatever rumble or hiss you’re trying to hide.
Fix your Video’s Audio with AI
This final trick is one of the wildest things I’ve ever heard and this video was my first chance to try it. If you have a creative cloud subscription, you can use Adobe’s online speech enhancement tool to do much of this work for you – at the expense of the control you get with any of the above methods. Simply export your problem soundbite as an mp3 or wav file, upload it, and in seconds, you have a clean piece of dialog. Mind you, it doesn’t work on every clip, but when it does…wow.
Conclusion
Hopefully, this will help you fix your videos’ audio too. It's the age of we-all-always-have-our-phones-on-silent, and we mostly consume our media as little intermissions in the neverending scroll. So understandably there’s a lot of push for making every video work without audio at all. But when it matters, it matters.
There is such a thing as unworkable audio, but this shoot was most certainly not an example. I may not have gotten it perfect on location but it was well within the bounds of something I could work with, and I wanted it to be as close to perfect as possible.
I was raised in a musical family, and audio was in a lot of ways my gateway into video in the first place. For me, I want to watch the video, not read the subtitles. And that means that it's important to me personally that my content is listenable. Good audio is just as important to good storytelling as is lighting or framing or color. So it’s important to me personally, that editors know how to make the sound better in the edit, too.
Comments