AI audio processing vs mixing it yourself – which is better for your videos?
There’s been a lot spoken about AI when it comes to photographs and video. Deepfakes are rife, and the situation isn’t looking great. There’s also images being generated based off the work of people who didn’t provide permission for their work to be referenced. But what about audio? What if you’re using AI to tweak something you’ve already created? What if it’s simply modifying or enhancing what you’ve already done?
Both Premiere Pro and DaVinci Resolve have added a number of AI-powered tools for enhancing audio in your video edits. But are they any good? Do they really compare to mixing it yourself? Is it better than doing it the old fashioned way? Should you even bother learning to mix? The team at Syrp Lab attempts to answer these questions and more in their new video, with a guest appearance by professional audio engineer Alex Knickerbocker.
I took to audio pretty early on in my video creation path, investing in Sennheiser and Rode microphones, a Tascam field recorder and Beyerdynamic studio headphones. AI-powered tools for pretty much anything, let alone audio, didn’t exist yet, so I figured out how to do it all myself as best as I could. Ultimately, the videos I was producing were mostly going online or being displayed on screens in controlled conditions. I didn’t need to confirm to any kind of TV or Hollywood standard. It just needed to sound “good”. But is “good” still worth the effort?
As mentioned by Alex in the video above, “good” is subjective. What sounds good to one person might sound completely terrible to another, and vice versa. And this is really the crux of the big debate with manually mixing things yourself vs using AI-powered tools. Beyond clicking a “Fix my audio!” button, you don’t really have much control over these AI-powered audio processing tools. You can’t tell them that you prefer a little more bass or presences, or example, than it’s applied. Or you can’t tell it to dial things back a bit if you think it’s applied too much. Some people may think the result is subjectively “good” while others think it’s pretty terrible.
For many people, particularly absolute beginners with no idea where to begin, the AI-powered tools found in applications like Premiere Pro and DaVinci Resolve can be quite handy. They can make bad audio sound passable (or even “good” – subjectively speaking) and they can make already good recordings sound pretty fantastic.
But the state things are in right now, you’re still going to want to learn how to tweak manually to get things sounding just the way you want them. AI might do its job fairly well, but it’s not a mind reader. It never will be. It’s going to require some input from the user in order to be able to give it at least some direction in how you want it to sound. Or, as mentioned, you have to tweak it manually – assuming the AI gives you a tweakable result.
It’s great to see how far AI audio tools have come, though, particularly when it comes to things like background noise removal. Being able to take out those distracting background elements like traffic or people while still having your voice clearly heard over the top of them can be difficult manually. AI certainly seems to make the job a little easier sometimes – even if it’s sometimes too strong, making locations sound like they were recorded in a vacuum with zero background noise. That actually sounds a little creepy.
It will be interesting to see how these AI-powered audio tools look in a few more years.
John Aldred is a photographer with over 20 years of experience in the portrait and commercial worlds. He is based in Scotland and has been an early adopter – and occasional beta tester – of almost every digital imaging technology in that time. As well as his creative visual work, John uses 3D printing, electronics and programming to create his own photography and filmmaking tools and consults for a number of brands across the industry.