McWannabe wrote:
Are you saying that we should move the 'region of interest' so that we have 30 frames of clean video, and then we have the actor move into the scene and do a T-pose and start his motions?
Oh, no, not at all. There is a separate 'ROI' type range selector specifically for the clean video (I think this one is simply called 'Background'.) It's about 30 frames long by default and located at the beginning of the timeline, but you can expand or shorten it as desired. You need to move this range selector to where your clean video is. In my case, my clean video typically starts about 5 to 10 seconds in because my computer is in frame and I need time to move out of frame. (I hope to eliminate that 'dead' time when I get the voice command thing working.) The ROI of course, is the second, much longer range selector.
Quote:
I've been moving the region of interest to where the T-pose started - and the end of the region of interest to where I no longer wanted to capture motion, which I guess is all wrong?
I thought this was how it worked too but in fact the ROI can start where your action starts. Your T-Pose frame is basically used to 'calibrate' the rig, actor, and tracker (obviously this is not to be confused with scene calibration.) Then you move to the frame where you want to start tracking and press 'i' (for 'in'.) Here, you can adjust the pose of your actor to roughly match the video and then click Refit Pose a couple of times. You should probably hit Analyze Actor just in case some colors have shifted but this is up to you. In my experience, I've noticed that colors may shift throughout the video depending on how the lighting affects the performer, so clicking Analyze Actor from time to time seems to help the tracker. After clicking Analyze Actor, you may want to click Refit Pose again to see if it improves any. Before you start tracking, go to the last frame and press 'o' (for 'out'.) When you're tracking, if you stop to correct a pose and click Refit Pose, you should of course track backwards and forwards from that frame for a few frames to smooth out the transition. (FYI, I usually let it track completely through on its own first, and then go back and fix 'trouble spots.')
Now, having said that, there is nothing wrong with starting the tracking process from the T-Pose. In fact, this saves you from posing the rig to fit the 'first frame' of the actual mocap, which may be tricky sometimes. But the downside is that you are spending time tracking a lot of frames you don't need. I've done it both ways, so it's a judgment call you'll have to make.
One way to speed things up if you decide to start tracking from the T-Pose is to use low quality tracking (and maybe disable feet and shoulder tracking,) stop the track when you get to the first real frame of action (and enable feet and shoulder tracking if you disabled these features,) Refit Pose, and then start the high quality track. Before clicking the Track Forward button, press 'i' to set a new beginning for your ROI.
And don't forget we now have that useful Takes track just below the timeline. You can use this to make 'bookmarks' for multiple ROIs in a performance. Also, when you right-click over a 'Take' range, you have several options to choose from, including a 'Set as Region of Interest' command and the various export 'Take' options.
Oh, one more thing, use the '<' and '>' keys (which are actually the ',' and '.' keys) to step through the frames and to accurately set the 'i' and 'o' frames. It's MUCH easier than trying to drag the ends of the range selector to a specific frame.
Quote:
I just noticed that the red triangle be dragged to make it longer (the onion skinning). I have absolutely no idea what that is for - or how to best set the onion skinning interval.
It's a way to visualize motion from a still frame, like the view motion path option. It's somewhat useful if you're trying to smooth out certain motions, but you probably won't use it often. (I can't say I've actually needed it yet.)
Quote:
I've got 6 cameras now that work at 50fps (NEC chipset on my 4 port USB card, bummer) - so if I can ever get over being sick, I'm looking forward to testing the 6 camera setup. I wonder if Michael has determined the ideal layout for the different camera heights with 6 cameras?
I just ordered two more cameras too. Michael said that if I don't have the space, I can keep them in my current semi-circle configuration, and they should help improve the jitter problems I'm seeing. I'm guessing I'll have to reduce my framerate too.
I'll post my experience after I do my next session for the film.
G.
