Following on from my last post about tape in a tapeless workflow, I also had some fun planning some timelapse shooting.
The requirement was to demonstrate the amazing transformation from empty exhibition hall to finished ‘environment’ – ditto for the main presentation theatre. The build would happen over 2-3 days, and I reckon that the timelapse sequences would last around 5-7 seconds each in the final edit.
Weapon of choice would be my EX1 – it has great timelapse, slomo and interval recording modes. Trouble was, client wasn’t going to pay for me to bring one, rent another, and hang around whilst the shoot goes ahead. I know somebody who got to spend a week doing a single timelapse shoot somewhere hot and sunny. And the plants died, so had to be reshot – what a job: filming grass growing…)
Alternatively, a camera like the Nikon D200 has a built-in intervalometer. Providing you have enough power (big power grip or a mains outlet nearby), you can shoot in extra-high definition with a really wide lens, convert the raw images into a QuickTime movie and then pan around the final movie from within FCP or motion. But the client wasn’t going to pay for a photographer to bring his kit and set it all up…
So a couple of PD150s were hired to do it. The Sony PD150 is a venerable little DV camcorder from the tape age and whilst tape isn’t known for its strengths in time lapse, the PD150 does have an ‘interval’ mode.
Tape isn’t good for timelapse because even when the tape isn’t moving, the head is rubbing over the tape all the time it’s laced up. Rubbing the same little bit of tape for minutes at a time could lead to the head clogging up or worse still, scoring through the oxide so it’s one big drop-out and candidate for the point at which the tape will snap. Furthermore, if you disengage the tape from the head after each interval, you’re wearing down the transport mechanism, dramatically reducing its life span. It also means that taking a shot every 1-10 seconds is well nigh impossible because it’s long enough to clog your heads but too quick to unthread and rethread the tape.
So the PD150 timelapse mode records ‘about half a second’ (15-17 frames) of video every XX seconds. It’s not good for shooting things like clouds, and still leaves you with some editing after you’ve sucked in the video. Because you’re only getting half a second of real time video every so often, the footage still looks ‘narcoleptic’ – like a security camera. To get a smooth result, you need to speed up the footage by a factor of x15 or so, so you’re only getting one frame out of each burst of 15.
Aeons ago, there was a bit of software that we got with FireWire PCMCIA cards (so we could grab DV video with Apple’s Wall Street PowerBooks) which enabled you to grab every ‘n’th frame of the incoming DV signal. I hear that Adobe Premiere can still do this, but Final Cut Pro does not. So the clip is sped up in the timeline. Tip: remove the audio, as you’ll have no use for it and FCP will spend processor cycles trying to speed that up too.
So let’s consider a bit of maths:
The camera will record half a second of video every, let’s say, two minutes. This will eventually equate to 1 frame (1/25th of a second) every two minutes.
A 60 minute tape has 3600 seconds of video, therefore 7200 half second intervals. With me? Those intervals occur every 2 minutes, so each hour tape will last 60 hours (2.5 days) in-camera, and 288 seconds (not quite 5 minutes) in the final edit. But then you can chop out the overnight bits and the periods where not much is happening.
With a little jiggling of the numbers, you can work out if you can get away with doing interval recorded tape. For anything faster (clouds, crowds), then you’re better off recording real time and speeding up in post.
Or get a file based device.