3D Ball Watching

Michael Brennan experiences history with possibly the first live screening of an international sports event in 3-D HD via satellite. Is this a brave new world or a just a theme park ride effect looking to evolve?In the past century stereoscopic or 3-D presentation of movies has had numerous false dawns. Recent advances in resolution and brightness of  digital projection is now giving live High Definition TV it’s opportunity at  keeping the 3-D sun above the horizon.

In a collaboration with The 3DFirm and  BBC Sport a live 3-D HD transmission was  conducted on March the 8th to the Riverside Studios in London.

The 3DFirm, is a consortium comprising media communications firm, Can Communicate, 3-D specialist company, Inition and hire and post production house, Axis Films. 

The 3DFirm had testing 3-D live with BBC Resources in the preceding six months.

The 8 March event was apparently the first live screening of an international sports event in 3-D HDTV via satellite. 

The audience at the Riverside Theatre in London – March 8, 2008Live 3-D TV has been trailed in the US and following on from the Riverside Studios live 3-D BBC test, HD pioneer Mark Cuban conducted a live 3-D transmission on March 22nd of  a match  between  his NBA basketball team Dallas Mavericks and the Los Angeles Clippers, to one of his theatres in Dallas.

Clearly great minds are thinking alike and despite 3-D movies failing to reach orbit, live 3-D coverage of sports may just be the rocket fuel required to make the business plan work.

Riverside studios is no stranger to ambitious and innovative technology. Forty years ago the it was the location for the first transmission of colour TV. It took 25 years for widescreen to arrive another ten for HD. Now just five years later 3-D has reached a point where the technology is reliable enough to warrant a serious test. The pace of change is indeed increasing.

So on the 8 March 2008 BBC Sport used the RBS Six Nations rugby match between Scotland and England to test live 3-D HDTV to an invited audience.

To help get the studio audience into ‘match’ mood  beer and steak burgers  were provided.

This was enough to fill the studio and most importantly help create the right atmosphere that otherwise may have been missing.

The audience was interested until the actual sport bored them, and the rain didn't helpThe3DFirm and BBC Resources  produced a stereoscopic (2 x HDTV) signal from a 3x camera Outside Broadcast at Murrayfield stadium in Edinburgh and delivered the signal via satellite to the Riverside studios in London.

The 3-D feed was stand alone and not cut into BBC Sport’s live host broadcast on BBC One.  It was supported with BBC Radio 5’s commentary which further distanced the experience, in a positive way, from being a TV event.

David Wooster of The3DFirm commented: “Premium sporting events, where demand for tickets outstrips supply, lend themselves perfectly to live 3-D transmission”. 

According to Aashish Chandarana, BBC Sport Innovations Executive, the process is very much at a test stage, with the BBC dipping their toes into the water rather than taking the plunge. 

Just as well,  because from a technological standpoint the current  technology is only just adequate. This is no mean feat,  but I’m not about to say that  watching an entire  game in 3-D with three or four cameras at HD resolution on a large screen is the same as being in the crowd. It isn’t. HDTV simply does not have the resolution to match the detail that a spectator can experience. 

It needs help in the form of long lenses, multi-cameras and action replays, these additions will fill the gap, even if they are in 2D it is an acceptable fudge given that it is pointless trying to hook two 100x lenses into a 3D system.

A 2D multi-camera experience including close-ups with 100x zoom lenses is more compelling than a wide-shot in 3-D at HD resolution. 

Perhaps this was proven when a quarter of the Riverside audience retired to the  bar to watch the 2D coverage, some claiming they missed the close-ups. I tend to believe them although since the bar remained open I can’t be sure their motivation was to see the game through the end of a pint rather than Fujinon glass.

Interestingly, coverage of the game included a wide shot that showed the large TV screen at the stadium. In the absence of replays from the 3-D presentation truck, and with the 3-D camera pointed at the stadium screen, replays shown on the stadium screen were actually quite interesting to watch and it is a clever way to introduce 2D into the 3-D broadcast!

The Riverside 3-D picture looked a little soft compared to a typical 2D HD digital presentation, this is a subjective evaluation of course. The promo graphics which introduced the game appeared to be sharper than the images from the match cameras.

Bandwidth  shouldn’t have been an issue as the 43Mb/s BBC used, on paper, is enough to put HD out of the pall park.

Indeed there were no visible compression artifacts or blocking, a result of  months of testing. Whilst there wasn’t any objectionable noise in the blacks, highlights from the pairs of Sony f950 cameras looked a little clipped. In comparison I’d say the picture had less noise but was softer than my Sky HD picture at home.

Very poor weather on the day certainly had its part to play in degrading the picture resolution, lenses working at less than favourable apertures and mist or rain degraded the image. 

In particular rain spots that didn’t play ball by perfectly falling on both lenses simultaneously were  very troublesome!  Ideal camera positions chosen  the previous day were scrapped to keep the cameras under cover and one of the cameras was lost  for a while as it was repositioned. All good production experiences to have had in a test environment with a friendly audience.

Existing HD cameras and zoom lenses have numerous faults and issues that are bearable in 2D but when one image is laid on top of another, the combination can be unacceptable. All the cameras appeared to use HDTV zoom lenses rather than primes, once again the ideal technology for large screen is at odds with the demands of live presentation.

Convergence was set to just beyond the nearest touch line on the field.

The result was a pleasant 3-D experience that was easy on the eye, without the overly dramatic 3-D effects of objects protruding out of frame. In this respect it is very different from watching a 3-D movie where the lens convergence and subject placement are used to create action that moves forward of the screen and into the audience.

Some sports will benefit more from 3D HD. No doubt goal cams positioned at the back of a soccer net will use appropriate convergence to create effect of the ball flying into the theatre. In fact I’d argue that Rugby isn’t the ideal sport to showcase 3-D so the fact that the audience enjoyed the experience is a positive sign.

Back to the game where the action closer to the audience was interesting to watch as the depth characteristics revealed player movements and ball trajectory in a more involving way than 2D.

But when the action moved to the opposite side of the field the 3-D effect diminished.

Some of this was due to the camera needing to be repositioned being further back from the field and out of the rain. 

It was here that I thought that a flycam rig, where the camera can move across the field on wires  would be a good idea to keep the action dominant in frame without the use of a telephoto lens that reduces the 3-D effect.

The projection system involved two Christie projectors, each with a polarising filter.

The use of two projectors creates a brighter image on the screen but also is another  part of the chain that needs to be perfectly matched and aligned if one is to tweak the most out of a 1920 x 1080 pixel  system. I believe that once projectors become a little brighter it will be the realD system (it  uses one projector running at higher frame rates) that will be the obvious choice for this application.

Analysts in the United States expect significant revenues from 2D and 3-D theatrical pay per view. 

Watching a live HD event in 2D with fast cutting multiple replays is one thing but live 3-D sets itself up to be as good as the being there and can’t rely on fast cutting and multiple cameras to create eye candy as the 3-D effect is too jarring.

But having seen demonstrations of 2D Ultra HD 4320x7680 pixels (or 16x the number of pixels of HD) there is no question that this is a far more involving picture than regular HD resolution and ideally suited for 3-D applications of live events.

NHK reckon that distribution isn’t feasible until 2025 so lets aim a little lower and set our sights on 4K (or 4x the number of pixels as HD) for live 3D events.

Its not that fanciful a concept, with Sony’s 4K projectors, live 4K cameras from Olympus, NHK showing its UltraHD system and RED digital cinema camera (suitable for  4K recording only as there is no live 4K output) we are already on the way to 3-D 4K.

But in the meantime it is feasible that existing 3-D HD with a mixture the ‘2D fudges’ should maintain the momentum of the latest 3D voyage long enough for 4K to become a reality.

‘Fudges’ and high tech working hand in hand, sounds like a classic TV business model me to me. 

Posted on May 1, 2008 and filed under 3d, broadcast, case study, mike brennan.