If you’ve been watching the Olympics, you may have seen some of the spectacular footage of Bode Miller racing neck and neck against Iyica Kostelic, or Lindsey Vonn and Andres Fischerbacher chasing each other down the super-G course (warning: Silverlight required). What you may not have realized is that these are solo competitions, meaning each competitor raced alone against the clock. What kind of strange quantum time dialation was creating the visuals on screen? A new video effect from Sportsvision powered by NVidia Quadro’s called ‘SimuCam’. As explained by NVidia’s Brian Burke (in this week’s nTeresting newsletter):
The SimulCam technology involves “background recognition,” a process that identifies the pixels that belong in the background and calculates how those pixels move throughout a series of successive images. Those calculations are made possible by leveraging the power of NVIDIA graphics processing units (GPUs). Differences in the camera angles between every two images of two videos are determined, and then every image of the second video is geometrically modified so as to match the viewpoint of the corresponding image in the first video. SimulCam then blends the two images together.
But that’s not all. A similar effect called ‘StroMotion’ allowed frozen images of an athlete to be overlain on his live video, creating a kind of “film strip” appearance of his actions. It was used in the Moguls competition, along with several other events. Again, from Brian Burke:
StroMotion similarly utilizes the power of NVIDIA Quadro GPUs to compute the camera movement between every two successive video images. Once determined, it stitches the images together, and using a high level of redundancy, it’s actually able to remove the moving object from the image. Then, from the computed camera movement, StroMotion can determine how each video image relates geometrically to each other and to the panorama. The identification of pixels belonging to moving objects is based on the change-detection of each video image within the corresponding area in the panorama.
If you can’t get Silverlight where you are, then check out this video from the CAR Sierra Nevada Freestyle ski for an example of StroMotion, and this video of Alexander Kerttu and Moa GS flying downhill together via Simulcam.
NVIDIA Quadro Pro Graphics Create Winter Wonderland for NBC
If you watched NBC’s coverage of the Olympics these past two weeks, you probably saw Bode Miller flying down the mountain racing neck and neck against Ivica Kostelic in the Alpine Skiing Men’s Downhill, and/or Lindsey Vonn and Andrea Fischerbacher chasing each other down the super-G course. If so, no, you weren’t imagining it— you DID see two skiers racing down the mountain to the finish line at the same time, mere inches from one another. Yes, these are solo competitions, with each competitor racing against the clock, but an NVIDIA Quadro GPU-driven technology from Sportvision called SimulCam gives viewers the ability to instantly compare one skier’s performance against another’s, helping better explain why one skier just beat out another by mere tenths or even thousandths of a second.
Or how about another intriguing NVIDIA Quadro GPU-driven video effect from Sportvision, called “StroMotion.” This one repeatedly freezes athletes in motion during a given segment of their routine to demonstrate, within a single frame, the entire evolution of their movements. A still photo is one thing. But a StroMotion-enhanced video sequence effectively lets the viewer see into the mind of an athlete as they execute a routine. StroMotion technology enhanced coverage of the Moguls competition, along with several other events.
The SimulCam technology involves “background recognition,” a process that identifies the pixels that belong in the background and calculates how those pixels move throughout a series of successive images. Those calculations are made possible by leveraging the power of NVIDIA graphics processing units (GPUs). Differences in the camera angles between every two images of two videos are determined, and then every image of the second video is geometrically modified so as to match the viewpoint of the corresponding image in the first video. SimulCam then blends the two images together.
StroMotion similarly utilizes the power of NVIDIA Quadro GPUs to compute the camera movement between every two successive video images. Once determined, it stitches the images together, and using a high level of redundancy, it’s actually able to remove the moving object from the image. Then, from the computed camera movement, StroMotion can determine how each video image relates geometrically to each other and to the panorama. The identification of pixels belonging to moving objects is based on the change-detection of each video image within the corresponding area in the panorama.
So no matter how you look at it, the power of the GPU continues to enhance the action you see on TV.
Keep your eyes out for more cool VFX driven by NVIDIA Quadro technology in the future.
