I've been working diligently away, as my free time permits, at the Boats of the Glen Carrig. It's a good-sized project -- if the average chapter size holds, it means it will eventually clock in at about 350 minutes of final audio, or almost seven hours. (By comparison, if I tried to record The Night Land in unabridged form, it would probably be about three times longer than that).
So far, I have raw recordings of chapters 1 through 7. The first three chapters are completed and released. Chapter 4 is mastered and mixed with music, and is nearly done.
In terms of chapter count, I'm not yet at the halfway mark, or even the quarter-mark. That's a little discouraging! However, since the process is becoming more standardized, I'm getting considerably faster at it. Even if this podcast never attracts very many listeners, I will have learned quite a bit and improved my podcasting skills considerably by the time I'm done.
Each chapter goes through roughly the following steps:
1. Raw recording of the chapter text. I can't record more than two chapters at a stretch without going hoarse; if I try to record too late at night, my concentration and my eyes start to fail me, making it hard to keep my place in the text. But the best time I have is after everyone else is asleep and there isn't much noise outside or inside the apartment, and that's usually between 1 a.m. and 4 a.m.!
2. Editing. This is sometimes quite time consuming. Hodgson wrote very long and complex sentences and they often trip me up. One example is a sentence where he used both the word "ladder" and the word "latter" a few words apart. That sentence took several takes to get right -- I have to unnaturally emphasize the words so that the listener will be able to hear the difference. Hodgson also loved commas. Sometimes I try very hard to respect his commas, because if you don't get the pauses in the right place, the sentence's meaning becomes unclear. At other times I run a little roughshod over them -- if a characters is being attacked by a giant crab, it ought to sound like the narrator is a bit breathless! Occasionally in the editing phase I'll have to re-record a phrase or sentence that just came out wrong, and then try to paste it in, which might involve very tedious tweaks to the levels and even pitch-shifting a bit to try to make it sound like it was part of the original take.
3. Mastering. This means running the edited vocal track through the master effects strip in DSP Quattro, configured with a compressor, de-esser, equalizer, and peak limiter. Each recording is slightly different, but I've settled on compressor effects that I'm mostly satisified with, so I no longer have to tweak and tweak. I don't seem to be able to apply the mastering effects at faster than real time, though -- the compressor doesn't sound right if I do that. It would be nice if I could improve on that.
4. Music selection and mix-down of the main chapter. This part seems to be the most time-consuming. I use iTunes to convert music and sound effects files from MP3 to AIFF. I open up all the music or sound files that I'm going to use in DSP Quattro, and set up an output file on a separate hard drive. Then I do the mix-down "live," adjusting levels, pausing and playing the vocal track when I feel like it needs to pause for atmosphere, and fading in and out the music tracks. Sometimes I get it in one, and sometimes I screw it up and have to do several takes. Because this is in real time, if I'm interrupted with some concern from the real world I have to start over. Then each chapter needs a separate, short music and sound effects credit recorded and mixed with audio.
5. Assembly. I make a playlist in DSP Quattro, and put in the pieces -- the standard intro, the chapter, the standard outro, and the per-chapter music credits. Then I set up fades between them, and tweak them until I like the results. Then I have DSP Quattro write this all to one file. The result is the final AIFF file.
6. Conversion, markup, and upload. The way I'm doing this now is to import the final AIFF file into iTunes, hand-convert it into an MP3 file, and fill out the info fields. Then I drag it into an upload directory, renaming it to the filename I want for the finished product on the server. I edit the XML file for the feed, adding the new entry, setting the date, file size in bytes, file length in minutes and seconds, and writing an entry description. I also add an entry to the index page (or, now, to this blog). The audio file and the feed file then gets synchronized with my FTP client. Uploading to my web host is often very slow.
When it's done, I test out the feed, updating my podcast description, and confirm that it looks OK. If there's some minor metadata correction I want to make, I re-upload the file and try again. On one occasion I had to replace the whole file because I accidentally had left five seconds of silence at the beginning. I'm a bit of a perfectionist, although sometimes I have to just grit my teeth and go ahead and finish the file, even though there is something I'd really like to fix, on the grounds that I just don't have endless free time. I'm not trying to do this in a sprint, but if I don't maintain the goal of finishing the whole novel, I'll never complete it.
There are some improvements I'd like to make. The testing could all be done with a local FTP and web server running on the same machine. There are workflow tools that would let me do step 6 more easily. I've contemplated writing some Ruby code to merge the metadata into the XML file. I certainly should learn how to use DSP Quattro better -- It has batch processing, keyboard shortcuts, and all kinds of other goodies. But for now, that's how I do it! How long does it take? Well, you can't really count the first few -- they would throw the curve completely out of whack. I think I will spend less than five hours on chapter 4. I expect to get that down to perhaps two and a half hours per chapter. Not too shabby! Of course, that doesn't count the time I spend blogging about it!