There was a taikai today for high school students, so no practice. Instead I spent some time editing and watching a video of my own shooting. One of the teachers recorded it while we were at Byodoji, and I find it very instructive. The big difference between video and looking at a mirror is that you can slow the video down and see movements that take place in milliseconds. What I see quite clearly is that my body moves considerably: left shoulder and arm moving slightly upward and forward at hanare, then reversing course and moving down and to the right as yugaeri carries the bow around. At the same time, the right elbow and shoulder move down and back, then reverse course coming forward and up again, so by the time all of this settles down, both left and right are at the same height, if dropped somewhat. There’s no way I could see any of that if I were just relying on a mirror: by the time I’d glance over, I’d already be in zanshin, all the action would be over, and what I’d see is left and right level, if dropped, and (at least in this one shot) the bow tilted too far forward. It would be very deceptive.
Teachers do see the millisecond movements. Or at least sense them in a subliminal way, which is quite interesting in itself, but this is the first time I’ve been able to watch what’s going on. What to do about it is another matter, but one thing I’ve decided to look into is the orientation of my elbow. If you think of it as a hinge (which it is), then the forearm can really only move in one plane. If the joint is vertically oriented, that plane will be left and right horizontal. But in my case the normal position is at roughly a 45 degree angle, counterclockwise from vertical. So the initial movement up and forward, as well as the countering down and backward movements are completely consistent with the orientation of the elbow. At the next practice I’m going to straighten that out. I know I talked about this before, but since then I’d seen contradictory advice. Now I have evidence that I’ve seen with my own eyes. And a plan…
In fact I wonder if all of the movement is just compensation for this one joint orientation, but I think that would be too easy!
Nevertheless I recommend doing this kind of thing to anyone who wants to really know what is happening in their hanare. You don’t need a camera with slow-motion capabilities. What I did was take quickly start and stop the video during hanare, and then take frame captures at each point. Doing that allowed me to capture 16 frames (about 40ms apart for 2/3 of a second). Then I used other programs to scale and rotate the images, then animate them as a sequence with 100ms gaps between each frame. Do-it-yourself slow-motion! It’s pretty cool, plus with images you can later zoom in on particular places (say the left hand) to see what’s happening, something that’s not so easy with video.
On Linux what I did (for future reference) is convert Sony’s unusual MTS video format to AVI using ffmpeg, then play the video with xine in slow-motion mode, using the t command to capture the frames in PNG format. Then pngtopam to convert the images to PAM format, followed by pamscale and pamflip to scale and rotate them (the video was made with the camera rotated 90 degrees, so I had to compensate for that). Then finally the animate program to animate the PAM images with a reasonable gap between frames. It sounds cumbersome but once you script it all it’s very simple. No doubt other operating systems have similar tools. In any case, be prepared for some surprises! A lot happens in that split second.