Bad audio can sink an otherwise great video. Whether your guest recorded on a laptop mic, you were stuck near an AC unit, or background music crept into your footage, the fix used to take real technical skill. Now, three AI tools can handle most of it in seconds.
Coming to you from Mike Russell, this practical video walks through three noise removal tools side by side using real-world problem audio. Russell tests each one against the same challenging clips: a guest recorded on an internal laptop mic, a recording made outdoors with bird noise, audio captured in front of a running AC unit, and a clip with background music that could trigger a copyright strike. The first tool he covers is Riverside, where he uses the Magic Audio feature inside the Riverside editor. The results on the laptop mic clip alone are striking enough to make you rethink how much you've been over-complicating noise reduction.
What makes Riverside's Magic Audio stand out is the adjustable intensity. Russell shows that cranking it to 100% works well for some clips, like the AC unit recording, but can sound slightly unnatural on others. Pulling it down to 30% or 60% often hits the sweet spot, keeping a little room ambiance while still cleaning up the voice significantly. For the outdoor recording, dialing it back to 35% removed the harshness from the iPhone mic while letting the birds stay in the background. Russell also mentions that Riverside is offering 20% off with his coupon code MRC, which he includes in the video comments.
The second tool Russell covers is Adobe Enhance Speech, and this is where things get more interesting for complex audio problems. Unlike Riverside's single slider, Adobe Enhance Speech gives you individual faders for speech, music, background noise, and more. The feature that stands out most is stem separation: you can download your speech as one clean track and the background music as a completely separate track, both isolated. Russell demonstrates this by importing the separated stems into Audition and showing the waveforms side by side. If you've ever needed to pull usable audio out of a chaotic recording environment, this level of control is hard to overstate. He also notes that in some cases, particularly the AC unit clip, Riverside's Magic Audio actually held up better than Adobe Enhance Speech when it came to keeping the voice sounding natural without robotic artifacts.
The third tool, Aonic, gets an honorable mention. It's free for up to 2 hours of audio per month, and while Russell finds it more limited than the other two, it does something neither of the others does as directly: it can match your audio to specific loudness standards in a single click. The full breakdown of how each tool handles every test clip, including where Adobe Enhance Speech pulls ahead and where it falls short, is worth seeing in full. Check out the video above for the full rundown from Russell.
1 Comment
I have used Adobe Enhance Speech a few times for enhancing and cleaning up old video files that I made at the start of my time interviewing artists, some while working in empty gallery spaces, others while using on camera mics (either because that was all I had at the time OR that I forgot to turn external mics on) In nearly all cases the results were quite good...a few had some artifacts but most of those words or segments could be used with a mix of the original file to get OK results. I'm hoping they don't start charging credits (or $$$) for it as I have dropped down to the Standard plan now I'm not doing paid work