podslashin'
I've been a little skeptical of the ![[info]](http://p-stat.livejournal.com/img/community.gif)
podslash phenomenon, even though I recorded one myself last week, because I'd way rather read a story than listen to one while I'm sitting at my computer. But this weekend I finally figured out how to use my husband's i-Thingy (it's not an iPod, it's an "iRiver") and loaded up some audiofic to amuse me while I did housework. And now? I'm totally a fan.
It's definitely time-consuming to create audiofic, though. I spent quite a bit of time working on a recording of "Briar Nose" last night, and I'm only about 3/4 done. (It's longer than I thought! It's going to run around 20 minutes, yikes.)
For those of you who are curious (or who would like to record their own stories and want some pointers) here's how I do it.
I use Audacity, which is a free cross-platform sound editor. In order to export mp3 format, you also need the Lame encoder. Audacity is fairly easy to use in a basic way - I recommend going to the tutorials and going through sections I and II. For hardware, I have a rather old stick-style microphone. I cover the end with some thin cloth, based on a suggestion by ![[info]](http://p-stat.livejournal.com/img/userinfo.gif)
cathexys, but I have actually no idea whether this does anything - I did a couple of tests and I thought it sounded better, but it might have been my imagination. I also use headphones for playback.
The way Audacity works is that every time you hit record (and end with stop), a new track is created. If you position the cursor at the end of the previous track, the new track will begin temporally just after the previous one, so when you mix the tracks the result is a seamless whole. This means that you can record in segments and then splice them together.
So what I do is read one or two paragraphs at a time, then stop and have a sip of water and relax a little before the next segment. If I make a reading mistake (a stutter, or misprononciation, or just voice something in a way I didn't intend) early, I delete the segment and start over; if I make one late, I stop just after I make it, then delete the sentence (i.e., everything after a long pause) containing the error. I read slowly and use a range of emotion to add color, but I don't give character voices distinct effects (much). I also try to keep the microphone at the same distance from my mouth during every segment. I record perhaps 4-8 segments (tracks) at a time before preliminary editing.
The first time I played a test piece back, I noticed a loud hum underneath my words. This is noise, and it's easy to remove it with Audacity: Select a few seconds containing only noise (i.e., a pause between sentences), choose Effect/Noise Removal, and click the "get noise profile" button. Then select the whole track and choose Effect/Noise Removal again, this time clicking "preview". You'll hear a few seconds of your track with the noise removed. I find that positioning the slider about 1/4 from the left end seems to produce the best results. When you're happy with the result, click "remove noise."
All segments recorded under the same conditions have pretty much the same noise profile, so I just go to each track in turn, select the whole track (by clicking in the left box - not the X, which deletes it, but anywhere else) and then "Repeat Noise Removal". Then I select all the tracks I've just recorded and choose Project/Quick Mix to make one long segment out of everything I've just made.
Then I grab scrap paper and a pen and listen to what I've just recorded. There's a slider at the top with time marks; every time I hear something that I don't like - a bad inflection, an error I missed, a background sound, a pop where I've joined segments, a too-long pause - I jot down the time. After I've listened to the whole section, I work backward from the end and fix each spot I've noted. If it's something that needs deletion, like a long pause or a joining discontinuity, I select the bit and delete - that's easy. If it's something more complicated, I may have to record a new bit -repeating a sentence or a phrase - and then delete the old and paste in the new. I always listen again to each bad spot and make sure it's the way I want it before moving on (backwards) to the next. (Working backwards doesn't displace the earlier time marks, which is why I do it that way.)
After I've fixed all the problems with this part of the story, I move on and record another 4-8 tracks, then repeat the process of noise removal, mixing into a single track, listen-testing, and fixing. When the second group is finished, I mix it with the first and check that they are seamless, then record another group of tracks. And so on, until I'm done.
ETA 9/21/06: I've been experimenting with reducing the quality on export, to make a smaller file, since voice shouldn't require the same bit rate that music does. A bit rate of 32 (rather than the default 128) reduces the file size to approximately 25% (32/128 = 1/4) and doesn't appreciably reduce the quality. A bit rate of 16 is unacceptable to my ear, though. The bit rate for mp3 export is set in Edit/Preferences/Quality.
ETA 10/25/06: After reading a few more stories, I've decided that it's easiest to fix verbal stumbles that I notice while reading by pausing for a few seconds, then re-reading the problematic sentence. That way I have a single track and I can simply cut out the bad parts. I frequently take long pauses between sentences to compose myself (or breathe, or drink water) and I cut those out as well. Also, I have a new microphone, and the quality is way improved - but I still remove the background noise. I also usually increase the overall volume a little.
ETA 10/8/07: More tips and tricks, posted 7/30/07
You can download my podfic from my audiofic page.
creative
