Note: the original post references audio files which no longer exist.
I’ve temporarily taken over the reigns of post-producing the Star Trek: Dimensions audio drama to give Waldo a break and let him focus on school. Unfortunately I discovered that one voice actor’s audio is just not up to snuff… so much so that I’m tempted to just purchase her a better microphone myself.
But because we’re supposed to release part one of the Before The Mirror mini-series today, I was left with some pretty low-quality audio and not a whole lot to stinkin’ do about it.
It’s time for some 4 AM audio heroism. Some of you may know that I do thePsychotic Resumes podcast. I’m the only one who records, so I can control the audio quality as much as I care to. But it’s not always bunny rabits and roses. You may not know that I was the 90.5 FM KCSU Production director for a year, and an understudy for a year before that. If I hadn’t been in radio for so long, I wouldn’t know about the silly stupid things that can save you from ripping your hair out over poor audio quality. If you’ve never dealt with this problem before, consider yourself lucky.
Most of the files that I received sound like they’re recorded over a bathroom fan. The audible portion of her voice was buried somewhere around -18 dB or lower. The audible noise (ambient room noise, computer fan or just a bad connection to the computer) was hovering around -25 dB. What’s worse is that it’s your run-of-the-mill high-to-mid-frequency hiss, that is, it intersects with the good audio all over the place. There’s not a whole lot of room to rescue the good audio (her voice) from the crappy noise.
If you look at the audio file, the wav actually resembles crab grass popping up from time to time whenever she is speaking. That there’s a field between those sprouts is not a good sign!
If you turn up your computer volume, you can hear a nice section of the hiss before she starts speaking.
This file is not that heavily affected, but it’s still not what I would consider broadcast quality. If you line up a higher-quality audio dialog track against this one, the difference is easily noticeable.
You can use a noise filter all you want on that, but it’ll only come out sounding like a big artif-cked mess (play below). Turn down your speakers for this one unless you want to sap your will to live. Note: this is what most people resort to, and they live with the artifacts. I don’t like hearing random ghost whistles while I’m trying to enjoy something, though, so…
This is audio that has been damaged. Sure, you won’t get the hiss, but people will wonder why her voice sounds so hollow. This is not an acceptable broadcast quality.
Instead, here’s what you need to do.
First: amplify your audio so that the peaks are at least maxed at -15 dB. You can use Audacity for this. Export from Audacity as a WAV file.
Second: use Levelator to automagically adjust the audio levels in the file. What levelator automagically does is apply normalizers, limiters and compression filters to smooth out the audio track. The best part is – it seems to know generally what is good audio and what is bad audio. That is, it was able to minimize the noise amplification while simultaneously bringing out Georgia’s lovely voice.
Third: NOW is the time to use those noise filters, but do it right! Gather a large sample of the noise without breaths, pops, clicks, or other oddities. Then apply your noise filter on either the lowest or second to lowest settings. Trust me – the Levelator does good work, you will not need much noise reduction and the automagic that it did will prevent most artifacts from being audible to anyone but dogs.
Check out the new waveform and audio:
You can hear the difference – this file is now suitable for broadcast. Now, it’s time for bed and a LOT of coffee in the morning… then later tonight, the premier of Star Trek: Dimensions – Before The Mirror – Part One!