Kyle Gilman’s Camera Noise Productions

I’m back from Amsterdam and we’ve finished editing La Commedia. A few weeks ago I described the basic setup we were working with. The whole thing worked beautifully. We almost never rendered anything. There were times when playing back a freeze frame while 2 other video streams were running would cause dropped frames, but that just required a quick render on the 8-core Mac Pro we were working on. The finale of the show finally uses all 5 screens at once in a seizure-inducing extravaganza that also included some effects, so that needed rendering as well. But for the majority of the show I was able to edit up to 4 streams of resized HD video in realtime.

This is an example of the format we used during editing. Each picture represents one of the 5 screens that will be in various locations throughout the theater. The “C” screen is not in use in this example.

The real difficulty in this project was creative. The opera doesn’t really have a clear narrative like a traditional opera. It’s more fragments of ideas that all relate to a theme. Hal Hartley, the director of the show, came up with a separate story for the film that was inspired by the ideas in the opera. But while there’s a clear relationship between the two, the film is definitely not just an illustration of what the people are singing about on stage. And I usually had no idea what anyone was singing about anyway since we edited to MIDI recordings of the score.

As we started editing there were a lot more decisions to be made than usual. In a movie you can take for granted the fact that you’re going to have an image on screen most of the time. You might have a few seconds of black here and there, but in general movies tend to have something going on all the time. But with a stage production, you might not have any video at all for several minutes while the audience focuses on some activity on the stage. And if we do want to show some video, we have 5 different screens to choose from. And some portions of the audience can’t see some of the screens. Some of the audience sees the back side of some of the screens, so the image is in reverse and we can’t put text on those screens. It was all very tricky.

I made a lot of progress on my animation project recently. The biggest breakthrough came from Aharon Rabinowitz’s tutorial on Lip-Syncing for Character Animation over at Creative Cow. I highly recommend it. I’m using his time remap method for lip sync and eye movement. It’s really helped me understand how this project can be done as quickly as I had originally hoped. I should be able to create a whole scene from multiple angles and only do the lip sync once. I’ll just use the same frames for each mouth position for every angle (straight-on, profile, etc) and copy and paste keyframes to each one.

Here’s my first test. It’s a line of dialog from the first episode. I did an arbitrary camera move and put Jennie in front of a background so I can start figuring out how those things work. I finally figured out that increasing the aperture size of an After Effects camera is a quick way to ensure the low depth of field I was looking for.

There’s definitely an Uncanny Valley effect going on here. Since it’s so close to looking like a real person, I think it is guaranteed to look a little creepy no matter what I do. But I’m going to work with that, not against it.

I’m in Amsterdam for 5 weeks editing an opera + movie. It’s not a movie of an opera and it’s not a movie with opera music. It’s an opera that also has a movie component that will be projected during the performance. There will be performers on the stage who are also in the movie. The movie consists of one giant screen projected in 1080p25 and 4 smaller screens projected in SD PAL anamorphic. The whole extravaganza is directed by Hal Hartley and the music is written by Louis Andriessen.

The movie portion was shot before I arrived on a Sony HDW-750P which shoots 25psf; basically putting the same frame in both fields of a 50i video stream, which is functionally the same as 25p, but not technically the same. It also looks a lot like 24p, with none of the weird motion that I usually see when I watch PAL video. Although there are definite judder effects when people move too much from frame to frame in front of high contrast backgrounds. I didn’t see things like that in Fay Grim, but I’m not sure what accounts for the difference. It’s HDCAM, so it’s 1920×1080 8-bit video. For the first time in my career we’re actually editing at 1080p using Apple’s ProRes HQ in FCP 6. We captured from an HDCAM deck using a Kona LHe card. Everything goes to an Xsan which so far has been able to fairly reliably play back 5 streams of ProRes HQ.

That’s important because in order to simulate the effect of 5 screens, we’ve broken our canvas up into 5 sections. There’s one large picture in the middle for the big screen and one 25% sized picture in each corner. In the theater, some of the screens will actually be perpendicular to the proscenium, but this is a good-enough approximation until we get our 3D holographic monitors going. I’ve set up 10 tracks in our timeline, one for video and one for graphics and other overlays for each screen. I’ve set up 5 motion path favorites in FCP and assigned them to the numbers 1 through 5 on the keyboard. As soon as I cut a clip into the timeline I select it and type the number of the screen it’s assigned to and it moves to the appropriate position. We aren’t doing any rendering at all. This stuff is getting indistinguishable from magic.

A word about ProRes HQ. It looks really amazing. At first I couldn’t see any difference between it and the HDCAM originals, but every once in a while now I’m seeing a small amount of aliasing on high contrast diagonal lines. The kind of thing that is always really tough for digital video. I’m willing to let that slide though, because otherwise it’s fantastic. EDIT: Those problems are entirely an artifact of monitoring through the Kona card in 8-bit mode. In 10-bit, I don’t see any problems with the picture. We’re monitoring on a Sony LMD-2450W 24-inch LCD screen and everything looks incredibly sharp. I am now officially spoiled by HD.

The mystery is over. YouTube, in another in a long series of overdue moves, has revealed some real information about how people find your YouTube videos. In my case, it turns out that my most popular videos happen to be declared “related” to some other, more popular videos. I’ve never questioned the success of the “Bad Webcam Sex” video, which is naturally connected to all kinds of filth that people are mistakenly looking for on YouTube instead of the entire rest of the Internet. What really surprised me was the sudden and unexpected rise of viewers for Two Night Stand. I’ve now learned that nearly 50% of the traffic comes from being “related” to a video called “Fake Wife Swap” which was made for one of those 24-hour film festival challenges. 90% of Two Night Stand viewers come from related videos.

The other big change—which still hasn’t quite worked itself out yet—is the so-called “high quality” option for YouTube videos. On certain videos (the criteria isn’t at all clear to me) you can add &fmt=6 to the URL and get a significantly better video. Unfortunately I don’t know what you have to upload in order to get the higher quality. Is it a higher resolution, or a higher bitrate that I should be going for? Adding &fmt=6 to most videos gets you the usual blender-set-on-purée look. Some guidance from YouTube would be nice.

UPDATE: Brian Gary has an article at kenstone.net explaining the best settings to use to take advantage of YouTube’s higher quality options.

I read a lot of job postings on Craigslist and Mandy. I have RSS feeds for any “editor” jobs that pop up.  Amazingly enough, I’ve ended up with a handful of great contacts from jobs I got through Craigslist. For some reason, among the many misspellings (trailor) and inaccuracies, the one that bugs me the most is requests for editors who can work with “Final Cut Pro HD.” Version 4.5—and only version 4.5—of Final Cut Pro was known as Final Cut Pro HD. It was a mistake to call it that at the time, and it’s just led to confusion. Anybody who is working in Final Cut Pro HD should upgrade. FCP 6 is much better.

My other pet peeve is the people looking for free work who try to make me feel guilty for wanting money for my time and extremely specialized skills. They’re always variations on the theme of “don’t apply to this job if you just got into this business to make money.” Well, you know what, I’ve worked for free on movies when I thought it would be worth my time, and I’ve asked people to work for me for free as well, but anyone with that attitude is not going to be fun to work with. You know who goes into a business hoping to lose money? People with complicated tax schemes, that’s who. Whenever you ask for someone’s time and effort, you need to compensate them. It doesn’t have to be money, but credit and a copy of the finished film are not compensation, they are an obligation. A positive experience is compensation.

I use Azureus quite a bit when I want to download something quickly thanks to the bandwidth of thousands of my friends. In an effort to become legitimate, Azureus has launched a streaming service called Vuze that encourages distribution of HD content. I don’t have any HD content to share right now, but on the web, 640×480 is pretty high-res. So I decided to upload a few things and try it out. I started with Getting Laid Tonight. Everything happens very quickly. Upload is as fast as you can make it, and the download is too. But the audio was a few seconds out of sync. I’ve had trouble with Getting Laid Tonight in the past because the fully cropped video is a non-standard size. That might have been the problem. So I tried the Brad Wood music video. The sync was fine, but the compression destroyed the music quality. Everything sounds over-modulated. I’ve removed both videos from the “content network” but I’ll try out a few other things. I’m mainly interested in this for the distribution of my animated series. I’m going to create it in HD, so this could be a good way to distribute the HD video.

Ever since Media Composer was released as a software-only option (no longer requiring expensive Avid hardware) I’ve had a hard time understanding why it cost so much more than Xpress Pro, considering the complete compatibility between them, and the large overlap in features. I didn’t see much need to upgrade to Media Composer myself. Well, apparently Avid is done with Xpress Pro, and they’re slashing the price of Media Composer. I think this is a great move. It puts Media Composer much closer to the price of Final Cut Studio, and removes the vestigial Xpress Pro line. It’s getting so cheap, I might even buy a Mojo some day.

Every once in a while I go through the various unlabeled and ambiguously labeled DV tapes I’ve been collecting since college, mostly in the hope that I’ll find the long-lost rehearsal tape that I shot with Jennie Tarr before we started shooting Camera Noise. I remember it being quite good. But I also think I taped over it.

Luckily, I found this little gem. Uncle Gary Breaks The Burro is a 4-minute story that my uncle told me for a class called “Life Stories.” This was the warm-up for my later project The Life Story of Kyle Gilman as Told to Him by His Mother Mary.

Last week Jennie Tarr and I did a photo shoot with Jenny Woodward for what I hope will eventually be a very quick photo animation technique. Here’s a preview.

Film lists. Like film itself, they are on the way out. When we did Lake City we shot on super-16mm and transferred all of the negatives to HDCAM-SR with 10-bit log color. We edited with downconverted DVCAM tapes, then did an online using timecode. Since it was edited on an Avid, and I used ALEs from the lab to batch capture the DVCAMs, keycode was also tracked automatically. It was good to have around, in case for some reason we had to re-transfer something from s16, but in the end the HDCAM-SR tapes were our master tapes. And obviously if you originate in HD you don’t have any film to track. So believe it or not, in all the modestly budgeted films I’ve worked on, I’ve never had to deal with film lists before.

There are still some times when you need to track film. If you shoot 35mm, you might cut negative and make a print the old-fashioned way without a DI. In that case you’re going to be using keycode like crazy. You’ll probably conform a workprint before you cut your negatives too. Or if you shoot 35mm and do a DI, you’ll benefit from the increased resolution of 2K and 4K. At the moment it is not cost-effective to scan in all your dailies at 2K, and certainly not at 4K. So in that case your DI facility will have to scan the negatives from a film list that you generate.

But I’ve recently been working on a movie shot on s16 and transferred straight to DVCAM. It was then edited in FCP. I was not involved in the project until they brought me in at the hand-off to the sound editors when it became clear that there were problems. Final Cut Pro is a great piece of editing software, but it’s so flexible that if you don’t know exactly what you’re doing, you can get in trouble fast.

The first problem is working with film shot at 24 fps and telecined at 23.98, and trying to synchronize it with BWFs (Broadcast WAV files, which are WAV files with timecode) from a digital sound recorder. One way to avoid this hassle is to conform all the video to 24.0 fps using Cinema Tools. But it’s really best not to do that. Sound editors don’t like it, and you’ll have some trouble viewing on an external NTSC monitor or outputting to tape. It’s best to keep everything at “video speed.”