One thing I didn’t think about when starting this dictation/transcription writing journey was how noisy my car actually is. My brain is more than capable of tuning out the sound of the road, but the voice recorder picks up everything– even the sound of other engines. While my audio recordings are intelligible to the human ear, it confuses the heck out of the transcription software.
I spent several frustrating weeks having to listen to every minute of every recording, carefully correcting every mangled word and even adding in entire sentences that the software just gave up on. From the various dictation groups I’ve been following, this is the point when a lot of people give up on dictation and go back to typing. Unfortunately, until they have a road full of fully automated cars, I can’t do that. I pressed on, mostly because even if I spent an hour correcting an hour’s worth of transcription, I got more words down in that time than I would have otherwise.
Now, the obvious answer would be to use a headset with a unidirectional microphone. Unfortunately, the recorder I have doesn’t have a Mic jack, so I am stuck with the omni-directional built-in microphones until I bite the bullet and buy a new recorder (but this one is only a couple of months old!).
I looked on-line and found that some people recommended a wind screen or “dead cat” muff to reduce the sound of breathing and wind, but there were two problems with these: First, It’s engine noise, not wind, that is the problem and I didn’t know how effective those would be. Second, I’m cheap and didn’t feel like ordering a muff and trying to size it for a voice recorder only to find it doesn’t do what I need. I also saw some suggestions for replacing the weatherstripping on my car doors and otherwise “soundproofing” a car. That was also not likely to happen. I don’t have time to write and I don’t have time to weatherstrip my car… especially when people then pointed out that it wouldn’t completely solve the problem.
Then, a digitization project at work introduced me to Audacity. This is a free audio editing software that has been a huge time-saver for me. Using Audacity, I can import my audio, reduce the road noise, then export the audio back to MP3 and run the new file through the transcription software.
This has reduced my errors back to very manageable levels. I’ve hit a point now, where I don’t bother editing the transcribed file. There will be mistakes (Dragon confuses “seal” and “steel” all the time) but I’ve become familiar enough with what to expect that they aren’t a deal breaker anymore. I mean, I don’t get outright gibberish and I can infer what the correct word was by the topic I was dictating and the surrounding text.
There are things to note before diving into Audacity:
- It doesn’t export to MP3 “out of the box.” This is due to some proprietary something or other that forces Audacity to not package a specific file in the software, but it’s easy to add the functionality. The University of Wisconsin-Eau Claire explains it here. If that link breaks, let me know, but basically you could Google: Audacity export MP3 Lame Library (yes… lame…).
- Noise Reduction: Follow a tutorial to explain the steps to get the most out of reducing noise. You’ll select a few seconds of noise (not speech), go to “Effects” and choose “Noise Reduction.” Step one is to get the noise profile. Then you’ll select the entire track, go to “Effects” again, choose “Noise Reduction” again, and follow step 2 (“Reduce”). I leave the settings at their default. I usually only have to do this once, but you may find other areas in your file where a new weird sound needs to be reduced (Maybe a Semi revved its engine next to you for a few minutes, like happens to me). Then export the file and voilà.
- Finally, read the manual! Seriously, if you are not already an audio editing pro, then read through the manual, familiarize yourself with the buttons, etc. Don’t just wing it.
Noise Reduction is the only thing I do to my audio files to prepare them for transcription, so I don’t need to know how to make the software do back flips, but it’s good to know what I’m looking at.
So, if you start dictating in the car and your transcription software is spitting out garbage, then maybe you need to run your audio through an audio editor first.
Do you have any experience with other audio editors or ways to reduce the noise inside a car when recording?