How to Improve Virtual Meeting Audio Recordings for Post-Production
With more and more content being produced, since the pandemic, using remote meeting software there has been an increase in the need for solutions to improve the audio captured from this software.
Whether you’re using Zoom, Google Meets, Microsoft Teams, or Skype, the default settings will always use your built-in computer microphone and compress it for sending across the internet with minimal lag. This results in a thin and lifeless sounding recording when you hear it back in the edit.
So, what can you do to improve this audio to a point where it would be acceptable to be used in an edit?
Well, I’ve got something that may just surprise you.
How do I know what I’m talking about? Head to the DigiProTips Experience and Background page to find out how I’ve built up my knowledge over a career spanning feature film, broadcast TV and digital content production.
DISCLAIMER: This post may contain affiliate links. We make a small commission if you buy the products from these links (at no extra cost to you). As an Amazon Associate, Skimlink & Descript Affiliate I earn from qualifying purchases. But we only recommend products we would use ourselves. For more information, click here to see our disclosures.
Recently a program called Descript, which I have talked about in a previous post for auto-captioning in Premiere Pro, has introduced a tool called ‘Studio Sound’.
Start a new project in Descript, add your media and let it process it.
Once processed, head over to the ‘Clip Inspector’ on the right-hand side of the program and select ‘Add Effect’.
From here choose to turn on ’Studio Sound’.
Depending on the size of your file, it can take a few minutes for it to process the ‘Studio Sound’ effect.
Once it has finished processing it will add the effect instantly. Playback and hear your newly improved audio.
Use the dial to change the amount the effect is applied to your audio.
Watch the Tutorial
Recording Productions Using Remote Meeting Software
Since the pandemic started many productions moved to conducting video interviews, podcasts and whole events via software such as Zoom.
This posed many challenges but one of the biggest, aside from the visual elements, is how to get great quality audio from these sessions.
The BEST way to mitigate against bad virtual meeting audio is to record separately using proper mics or even a headset and your mobile phone, basically anything other than the in-built audio device used by the meeting software.
Learn more about budget audio setups here:
Then send the files from each person in the meeting to your editor. This will always be better than trying to fix the audio in post when all you have is a limited frequency range recording from Zoom or Teams.
I recently had the exact same frustration after a couple of my guests didn’t use external recording devices and so the only audio I had to rely on was the audio compressed by the Zoom software. (By the way, always record your virtual meeting as a backup for these exact reasons).
The reason why remote software audio is bad to work with in post-production is that these meeting software programs compress and reduce the frequency range to better enable the software to send and receive that audio across the internet in real-time.
This means you are missing a big chunk of frequencies when it comes to what is left behind from the software.
The human ear naturally picks up frequencies from around 20-20,000Hz, the more of this range that is in a recording the more ‘full’ the audio will sound and in some cases ‘feel’.
So when it comes to using audio from virtual meeting software in your edit/content, you are trying to make a very thin-sounding audio recording sound much fuller than it is.
That’s a hard task! And it was a hard task I needed to solve myself.
So What Can You Do to Make Your Audio Sound Richer?
To be honest…. Not much.
It’s like filming, if you’re over or underexposed then the information that is clipped is lost forever. You can’t get that back.
The next best thing you can do is to use the power of tech to synthesize the sound that was thrown away and what that may have sounded like.
Effects in Pro Tools, Logic Pro, Adobe Audition, RX Advanced or DaVinci Resolve will have some ability to make your audio sound ‘fuller’. By adding some bass and treble (frequently the frequencies that remote software throws away when compressing audio).
You could also add some compression and try some double phasing to try and trick the ear into thinking there is more depth and information in the recording than there actually is.
Note. Find out more about Izotope RX here:
Unfortunately, all these effects are doing is exactly that – tricking your ear. In some cases, this may be enough and you can carry on your way. However, for most circumstances, this will still sound thin and tinny or it may even sound worse.
I needed to find another solution that would add those frequencies, that were so desperately missing from the recording, back in.
Up Steps AI to Save The Day
We all knew that eventually computers would overtake us and be used for everything and anything.
Well, they’ve done it once again.
AI is finding uses in every crevice of our daily lives. Content production is no different.
It was when I discovered Descript’s new ‘Studio Sound’ tool that I knew AI had saved my production.
What does it do?
It does exactly what I mentioned above. It uses computer learning to analyze your recording and synthesize what it thinks the recording would have originally sounded like before being compressed.
It does this in real-time on your recording and adds those frequencies back into your audio.
Take a breath of relief!
Is it as Good as it Claims to Be?
So with these things, there’s always a promise of something great and the reality can sometimes be a little disheartening.
What was my experience with Studio Sound from Descript?
Well, like with all AI tools, it works but it’s not perfect.
Now, having worked with Descript before I know they have spent many years and iterations of their main transcription tool improving what their AI can do. Let me tell you, it has improved significantly since I first started using it around 2018/9.
I believe the same will be the case for Studio Sound. The more it learns the better it will.
That’s not to say what it does now isn’t impressive, it is, but I think it will get even more impressive as time goes on.
The biggest thing to bear in mind with the Studio Sound effect is that it’s not a fix-all tool. It will highly depend on the quality of audio you give it to begin with as to how much it can improve it.
If you have recorded someone with the gain turned right down, they’re in an echoey room and there are lots of other noises going on at the same time then it’s going to struggle.
However, if you give the tool something that is definitely audible but not what you’d expect for a production-quality piece of content (such as using Zoom/Teams/Meets audio) then it will produce much more realistic and usable results.
The main thing I have noticed with Descript’s Studio Sound tool is that there is a significant noise gate when a person finishes a sentence. It sounds like the person’s words are being clipped/tapered down.
So with my recording, whenever there was a pause or a definitive end to someone talking you could hear the AI tool turning off the effect.
How to Counteract the Noisegate?
We should probably understand what the AI Studio Sound tool is actually doing to improve your audio in the first place.
There isn’t much that Descript gives away about this but it’s fairly easy to hear that the tool is applying a re or de-reverb effect (depending on if your recording is echoey in the first place) to make the audio crisper, it is then adding a compression effect to bring the frequencies that we, as humans, are accustomed to hearing to the fore and then it is probably using some phasing to fill in any missing information in the waveform/spectrogram.
Now, that is guesswork on my part but if what I’ve stated above is true then there is quite a bit going on that the software is doing to make your audio sound more ‘life-like’.
So when there is nobody speaking it doesn’t have anything to work with and so shuts the effects off. This is what can be heard when someone finishes a sentence without going straight into another.
The software is doing this on its own and there aren’t controls in the tool to manually adjust where and when it turns on and off the effect but what it does have is an ‘intensity’ slider.
This slider will essentially tone done the effect and mix it back in with the original audio recording. The tool automatically sets the effect to 100% so mixing it back down may make the main body of your audio less ‘full’ compared to 100% but it will make the noise gate less noticeable.
The other, more manual, task you can do is to export the Studio Sound effect on your recording as a .wav and then bring both that and your original audio into your editing software and do a mix yourself on the more noticeable noise gate parts.
However, if you have an hour-long recording this would probably be an inefficient use of your time and the whole idea of this is to save time, working smarter not harder.
So How Did My Audio Sound Afterwards?
To be honest, it sounded great for what I needed it for.
It was being used on a 60 second piece of social media content that would have a musicbed and sound effects mixed in as well. So the overall impact of Studio Sound on my recording was fantastic.
It sounded much richer and fuller than the original tinny and thin recording I had from Zoom.
I would definitely recommend using it for shorter form content that is being used in conjunction with other creative media in your timeline.
For podcasts and long-form content… I’d say give it a go and see what you get. The results may just surprise you!
What’s the Catch?
I hear you ask.
And you’d be right to ask it too because Descript isn’t giving this away for free (well it is, kind of, but not for very long). It is part of the whole software and after your first 3 free trial hours of transcription, you have to pay a subscription to access the software.
I use the tool regularly for all sorts of reasons, transcription and subtitles creation being the main one but also because having access to side tools like this that other audio software doesn’t have the capacity to compete with yet is just an added bonus!
There are different pricing tiers, starting from $12/month which will give you 10 hours worth of transcription a month or $24 will give you 30 hours.
The Take Home
I highly recommend downloading and signing up for your first 3 free hours and using the tool.
There’s nothing to lose and it may just end up saving your audio and your content production as a whole.
Save time, work smarter, not harder. That’s what we’re all about here at DigiProTips.
For more information on the uses of Descript check out my guide to auto-captions here: