This is part 2 on a post chronicling my trailer-making experiments. For part 1 see here.

Here’s the Harbinger trailer again, so you can see what I’m talking about:

Step 4: Get Music
This has always been the trickiest part for me. Basically, you need a soundtrack you can cut up and modify (I’ll come back to syncing the music and the images later), and I had a lot of trouble finding those. A quick survey of people doing their own books trailers showed them either using public domain stuff (like old interpretations of classical music), or having musician friends/acquaintances who could provide them with slightly cheaper alternatives to mainstream music (I don’t even want to know how much the majors charge for using bits of song, given how bad they are at authorising authors to quote lyrics for a reasonable sum of money).

So, instead, I went trawling the net. There’s a couple royalty-free songs out there, but so far the only useful source I’ve found is ccmixter. A lot of the songs here are mixes and loops, which I don’t like (I think they lack a sense of progression, which is problematic when you have to hold the viewer’s attention for 1 minute). But you can find lots of really good things, too: for the Harbinger trailer, I downloaded two instrumentals with a melody (you can also find voices if you prefer), after making sure that the license allowed me to remix them.

Step 5: Put Everything Together
Aka the step where everything went south. Very time-consuming (not as time consuming as step 6 below, though). I had the images/videos, the script and the music; now it was time to mix everything together. Ie, put everything into a video editing software and tinker to my heart’s content.

In my case, the video editing software is straightforward: imovie ships with the mac. I don’t know what Windows offers, though (there used to be a Windows Movie Maker, but it got significantly stripped down in Windows 7…).

Substep 1 was making the movie without sound: I took the images (which had been resized to compatible sizes by that time, if they were too large or too small), superimposed the text, and worked out transitions. Superimposing the text was relatively easy: it had to be readable (ie, not stuck on a colourful patch), to stay onscreen long enough to be read, and to be broken up in chunks if it was too long. Most of the longer sentences in the script were broken up: they remained on the same image, but didn’t appear all at once (it meant the image stayed on screen longer).

Working out transitions was an important bit of this substage: one of the things they do is keep up the rhythm of the trailer, by replacing one image by another, and also by affecting the way text appears on-screen. At this stage, having an idea of the music was good, since it allowed me to go for slow transitions (the musics I’d picked were doom-laden and atmospheric). I kept most transitions between images the same, except for specific effects: when the star-demons are mentioned, you’ll notice that I used more brutal transitions than the rest of the trailer.

Finally, still images are boring–but fortunately, there’s a trick, which is known as a Ken Burns effect (basically panning around the image, giving the impression the camera is moving). Imovie has some options which allow you to zoom or pan, and can make up some neat effects. The gradual close-up on Tenochtitlan at the beginning is a classic, but I’m particularly proud of the effect on the Aztec statue, which combines a slow zoom with a panning up that finally reveals the face of the statue. Little tricks like that made the trailer seem alive even though it wasn’t conceived as a movie.

Substep 2 was syncing everything with the music. I’m not a very good mixer and my musical ear is terrible, so I went for the simple option of having large sections of the songs accompany the trailer (the first 20 seconds, for instance, are pretty much one continuous bit of music with no background). I used bits of both songs as appropriate: the more sombre one went at the end, and since it didn’t sound spiffy enough, I added harmonics of the first one in the background, as a reminder. Imovie is pretty flexible here, allowing you to mix stuff in the main window, and to set up the transitions from one bit of music to another. One important thing on this step was making sure, as much as possible, that the transitions from one image to another corresponded to musical beats. It wasn’t always possible to do that and keep the text readable, but I did my best to make it seem as though the music and the images were on the same page, so to speak.

imovie has the advantage of being free, but it does have a few flaws. By far the most annoying was that while I could place text on the images, I had a very limited amount of control of where to place the said text (which is a problem when the images are very colourful and detailed); and, having satisfactorily placed the text, a very limited amount of control on how the text arrived (fast, slow, blending in…) I did my best, but there are still placements I’m not satisfied with. Similarly, the Ken Burns effects could be annoying to set up; and the duration of effects couldn’t be tweaked below a certain resolution, making it hard to sync the music as finely as I’d have wanted to.

Of course, this was the part where I put the cap in: particularly, the book cover, the availability date, and the credits. The credits went between the unveiling of the cover and the release date, because it’s smarter to end the movie on a reminder of when you can buy my book; but I didn’t want to skimp on attributions for the stuff I’d downloaded.

Step 6: Handle Video Formats
It might seem like I was done at this stage: after all, I had the video with the images, the music and everything, right?

Or not. One major headache was dealing with video formats. First off, tweaking the compression settings was more complicated than I thought: I wanted a movie file that wasn’t too large, but whenever I lowered the compression rate, the text would end up looking really funky. Second, there was the question of readability: the best export quality I could get was in quicktime format (which obviously isn’t readable if all you have is Windows Media Player). So I attempted to convert quicktime to WMV or avi. Ouch. It’s kind of harder than it seems, because there are some many variables involved: not everyone has the same codecs, and figuring this out took me the better part of a day. I had the luck to have two Windows machines and a mac, which allowed me to make some tests before uploading everything to my website/youtube.

I couldn’t find anything decent in the way of video conversion: the only software I found (courtesy of the fabulous Jeff Spock) was SUPER, which is basically a graphical interface for a number of freeware conversion libraries. It’s only for Windows though (Mac basically sucks at video conversion, not unless you’re ready to pay quite a bit more than I was), so I installed it on another machine and used it to convert my MOV into a WMV. I had to set the video quality mainly through trial and error, which is where things slowed down: my other Windows machine is a netbook, ie possibly the worst platform ever to do this, and it took me a good 20 minutes per conversion.

When I finally had the format I wanted, I was all set–I showed the trailer to my agent and my editor, factored in the feedback, and they later uploaded it to youtube for the world to see. Et voilà!