10/03 3D Rugby

There was a reasonably widely-publicised live 3D broadcast of the rugby match between England and Scotland last Saturday, and there's a fairly non-technical writeup on the BBC News website. There's one particularly interesting quote:

"One of the first shots of the crowd showed a fan waving a large flag back and forth. It seemed to come right into the room and I had to resist the urge to reach out and touch it."

It's interesting, because I was present at the screening, and I swear that virtually all the shots were framed such that the action was (correctly) placed behind the screen in 3D space. Including the "flag-waving" one that the reporter particularly mentioned. Indeed, one of the few criticisms I'd make of the shot framing was that some shots had unavoidable foreground material (eg the pitch, or people sitting just in front of the camera in the stand) that were placed in front of the screen and were hard to converge because of their extreme disparity and proximity to the edge of the frame. I believe that the 3D team were understandably limited to camera positions not required for the simultaneous international 2D broadcast, and so they might have chosen differently given free reign.

A colleague who was present at the screening mentioned that the cameras were set a bit too far apart, resulting in a "Subbuteo" effect, whereby the players appeared to be a couple of inches high. He had a point, although looking at the "making of" videos linked to from that BBC News writeup, it becomes clear why; using broadcast-quality cameras and lenses and simple (aka foolproof) stereo mounting rigs, it simply wouldn't have been possible to get the cameras' optical axes closer together.

Another noteworthy point is instanced in the second of the three "making of" videos: lots of assumptions and working practices that apply to 2D television and film craft need rethinking when you're shooting stereo. The example here was rain - an annoyance when shooting sport in 2D, but entirely liveable with. In stereo 3D, any kind of blemish that is present only on one of the two camera images, such as raindrops on the lenses, is entirely unacceptable. And that's on top of ensuring that zoom, iris, focus, gain and all other camera parameters are kept synchronised on each camera in a pair - something that's far from trivial with current equipment.

Overall, it was a triumph for the Outside Broadcast team from BBC Resources, all the more so since they discovered to whom the BBC was selling their department on the day before the experiment: the sell-off being a process which is inevitably concerning them at present. I believe that their efforts have unambiguously demonstrated the feasibility of stereo 3D coverage of live sporting events, and hope that they (and we) can go on to develop a craft of 3D sporting coverage - perhaps one day even bringing it to the living room.

03/03 Lenny speaks!

I don't like blog posts that are just links to other people's blog posts, but this is going to be one. Lenny Lipton (in addition to writing the lyrics for "Puff the Magic Dragon") has become probably the world's foremost authority on stereoscopic cinema, and his blog is worth reading in its entirety if you're interested in the field. His latest post is of particular interest to me and my colleagues - because it reminds us, at a time when we're starting our involvement in a 30-month EC-funded collaborative engineering project to try and make 3D television practical, that television is an art form, that good artists will always push the boundaries of any technical limitations imposed by engineering decisions, and there are no guarantees that a system that is modelled on the human visual system (eg two cameras, 9cm apart) will be optimal, no matter how obvious it seems.