The walnut cake movie is the first movie I’ve made in a while. It was made without any planning or forethought. Every time I make a movie, I screw it up in new and interesting ways. Here’s what I learned from my screwups this time:
- I’m deeply unhappy with the voiceover. Without a real mic, I was reduced into shouting into the lousy condenser mic in my laptop, which sounds about as bad as you’d expect.
- FCP’s voiceover tool is simultaneously nice and lame. It’s nice in that it seamlessly supports multiple takes and gives sound cues and lets you synchronize voice to video. However, for me, this led to a stilted and unnatural reading. If I could do it all over again, I’d sit down with the script and make someone else do the actual recording, do 10 takes without syncing to the video at all, and then do any necessary sync completely in post. In other words: I think I’d have been better off not using the voiceover tool at all.
- I wanted to use the audio mixer to set audio keyframes so that the background noise faded in and out depending on whether narration was occurring or not. In order to set audio keyframes, you have to click the completely unlabeled “record audio keyframes” button. The button has two states, and you can’t tell by looking at it whether it is on or off. You just have to try moving the sliders and then go back to your clip and see if any keyframes magically appeared.
- Even with the Canon Optura Pi’s optical image stabilization, the picture was unacceptably shaky. A tripod wasn’t practical under the circumstances, but I still wish I had had one.
- When I first visited the shop, the cashier was a personable, photogenic young lady who was excited to talk in enthusiastic detail about the product, its history, its manufacture, the customers, and so on. When I came back the next day with the camcorder, the only person available to talk was the polite but much less communicative owner. This is the standard photography lesson of “If you don’t have your camera, you can’t make the shot.” There’s really nothing you can do about this if you’re not willing to have a camera with you at all times, but I still feel like complaining about it.
- Maybe I just haven’t learned the right shortcuts yet, but I somewhat loathe the seemingly random elements of Final Cut Pro’s user interface. Here’s one example: Line up two clips next to each other. Grab an audio crossfade transition and try to layer it across adjacent audio clips. Sometimes the transition straddles both clips (which is almost always what you want). Sometimes it will only ‘fit’ on one or the other clip. In particularly pathological cases, it will straddle neither, but will drop itself between the two clips as a 0-length element, mocking you. I’m sure there’s some sophisticated explanation for this behavior, but the UI sure doesn’t make it apparent why these different behaviors occur.
- The “zoom” effect you get by stretching or compressing the scroll bar in the sequence window (which gives you a finer or coarser view of the sequence by stretching or compressing time) is the single worst UI element I’ve seen on a Mac. I have never once succesfully managed to reach my target ‘view’ without having to stop, readjust the scroll bar to recenter, and start over again. And of course, I use it anyway because you pretty much need to to do fine-grained editing.
- I want a hotkey shortcut that says “take this clip I’ve selected and move it left in my sequence’s timeline until it bumps up against something, and then stop.” That’s one of the single most common tasks I currently do by hand (and it generally involves wrestling with the awful time-dilation scroll bar) Does that exist? I haven’t found it yet.
Although lots went wrong here, I’m optimistic about the format. I’m a big believer in tiny yet good things. Perhaps there will be more 2 minute movies about food in my future.
Never use a computer’s microphone or the camera’s on board microphone. They’re lousy at best. While you could put a decent mike on your mac, usb or camera…trying to do a VO near your mac has too much ambient noise.
When you have the (A)rrow tool, hold down the option key to get the (P)en tool. This will allow you to create keyframes on your audio directly, if you have keyframe overlays turned on (this is in the bottom left of your timeline.
Never use camera stabilizers. Zoom out all the way, hold the camera near your body. If you need to get closer…get closer.
I believe opt-cmd-T puts an audio crossfade.
Use Shift Z to zoom out to see the entire timeline; command (or option) plus and minus to zoom in and out. It takes time to learn the lollipop zoom bars.
The T key selects the (T)rack tool. You can pull everything to the left (or the right) with it.
Make sure s(N)apping is on.
Email me directly if you have other questions.
Another way to move everything to the left is to select the empty space between the clips, right-click and choose Close Gap.
And if the audio fade is only applied to one track, it’s because the channels aren’t linked. Select them both and press opt-cmd-L (I think, that’s right.) I have never quite figured out why sometimes the audio isn’t linked during capture, but sometimes it’s not.
Jeff, Kristen, thanks for the comments! The “close gap” command is exactly what I’m looking for, and I never knew it existed. Jeff, your comments about voiceovers and in-camera sound match my beliefs (I basically absorbed my sound philosophy straight from Jay Rose’s “Producing Great Sound for Digital Video”). But as T.S. Eliot once said, between the idea and the reality, between the motion and the act, lies the fact that I’m fundamentally cheap and lazy and don’t typically carry around an off-camera mic. My work suffers for it, and I’m the first to tell someone _else_ that they should get a wireless lav setup, and I admit it is a fundamental flaw in the stuff I’ve done to date.
I think the mixing issues go beyond the ability to create audio keyframes. Fundamentally, a mix is about taking multiple tracks and adjusting them on the fly. Even using the “looks like a mixer” mode in FCP, you’re still limited by the mouse pointer’s ability to move a single slider. That’s not how you do a mix in real life; in real life you listen to the mix in real time with a bunch of fingers on a bunch of different sliders, and you move them all at once based on what your ears tell you (if you’re mixing more than about four to six tracks, depending on how dextrous you are, you have more than one person working the sliders.)
The “FCP way” to achieve this seems to be to listen to the track over and over again, setting keyframes as needed on each track (or linked pair of stereo tracks), eventually zeroing in on the right sound, hopefully. That’ll work, but it’s a very different workflow than I’m used to when I have the luxury of using someone else’s very expensive equipment.
The zero-length crossfade and other seemingly unpredictable effects puzzled me at first, too. I think they relate to how much source material lies past the out point of the outgoing clip and how much lies before the in point of the incoming clip. In other words, the transition needs more footage to work with than just a straight cut. You could add silence at the beginning or end of the offending clip(s), but I think this would end up sounding like a fade out – fade in rather than a crossfade. I guess you could loop some room tone in there instead of silence. Caveat: I never did an experiment to test this, and I only have a few months amateur experience with this stuff. Feel free to call me a bozo.