Human Factors Research Panel
September 22, 2011, 3-D User Experience Technical Summit, Hollywood--A panel discussed the needs and types of data that will most help the content creation community. The transition to 3-D is showing a large gap in the understanding of 3-D effects production and their affects on viewers.
Phil Lelyveld from the Entertainment Technology Center at USC moderated the panel. Members included Bill Chapman form Turner Studios, Sean Connolly from Indiana University, Buzz Hays from Sony 3-D Technology Center, and Howard Postley from 3ality Digital.
First, what to do about linear content?
Chapman responded that their focus is on extending their brands in news, sports, and cartoons. They need more information on viewer acceptance, but also need to get some idea of genre versus branding. Control studies are useful, but they also need to understand about story type, platforms, and distribution models. It would be helpful to understand how much or how many rest or break periods are needed versus the 24/7 content drivers. Producers need to learn how to interject 2-d into 3-D for rest periods and how to mix 2-D and 3-D to keep audience engaged.
Connolly added the need for practical implementation specifics as the industry transitions from the creation of 2-D to 3-D. Hays noted that Sony is trying to help with education through their Tech Center where over 2,000 people have gone though their programs on production and story telling. Sony is trying to help perfect the user experience.
Postley says the industry has accumulated lots of experience and data. Alignment, matching, and tracking are key requirements for quality work. The cost adder for 3-D has dropped by an order of magnitude from '08. The U2 concert cost over $20 M to produce and now would cost only about $2 M. So far, they learned that depth perception is a very subjective issue and that content affects the perception. The number of myths about 3-d are increasing, indicating a growing body of content and more viewers.
Small errors have a disproportionate impact on 3-D perception. A 0.01° rotation can completely destroy all perception of depth. 3-D errors are relational and relative and not absolute.
Changing creative for young versus older audiences?
Hays opined that content always has subjective and physiological causes for problems. Bad 3-D is bad for all people. Since most of the earlier releases were animation, content producers needed to be concerned about younger viewers because of the context and demographics. These films were designed with kids about 10 years old in mind. Changing the depth during the shot made the effect too deep to look at. Overall, the recent releases have enabled viewers to suspend disbelief.
In general, go conservative on the depth budget and try for a good balance. The screen size determines the viewing sweet spot, which can vary from 90' down to under 24". Do no harm.
Postley noted that vergence is an issue. Accommodation is the way that depth is perceived, so the depth and the budget are correlated. These are not the only factors, color grating, and contrast also contribute. A vergence point does not exist on the screen plane, so creators have to focus on the things that are in focus. Intensity, color, and changing the context of the shot can increase immersion.
Hays stated that aggressive depth changes can be uncomfortable. There are still many issues related to vergence and accommodation at distances of less than 2 meters. Creators still need to do lots of depth balancing. Unfortunately, filmmakers are still using 2-D techniques for their productions, because the eye changes require relearning their vision.
Connolly added that there are many effects that you shouldn't use in 3-D, but are still used based on the prior film art. There are too any myths, so creators need to try something and see if it works.
Genres and platforms?
Chapman suggested that the business case for genre makes more money, but you always have to match brands. Sports are good for 3-D. Sitcoms and dramas are questionable, 3-D will make stories more vibrant. Creators need to find what works, and what the audience wants. The industry needs to define what works at what level, especially for displays below 60 inches. Content at the personal device level is companion content, and is complementary for the different platforms.
Hays noted that filmmakers are trying 3-D to change the art of storytelling. It's not about genre but a point of view. The techniques and styles change and interesting things are coming from out in left field. 3-D provides a new vantage point.
Connolly stated that were working with hundred years of 2-D language. Now we're moving into what 3-D likes.
Editing and style?
Hays compared the change from 2-D to 3-D as the equivalent of changing from painting to software.
Glasses can be a disassociative change in expectations, especially the transitions from 2-D to 3-D within a film. (The latest Scorsese film actually requires viewers to put glasses on and take them off during the movie.) Hays considered that reviewers and 3-D TV excepted glasses as the only thing to focus on, which may be good. Glasses close out the rest of the world. Postley opined putting on and taking off classes during a movie is very intrusive. This would be an even greater problem with active glasses. Chapman added that the challenge of being cut off from the world is best resolved by leaving on the glasses.
Resources available to content creators outside major media centers?
Connally posited that performance, the technical aspects, come first and scripts follow. For example, dance and martial arts are good subjects. There are major differences between big-budget Hollywood productions and 3-D TV. Independent creators can access small audiences, but there is no discovery, marketing, or curation. The smaller budgets are not a major barrier.
Hays offered education is a critical item. Most films are high costs, so creators need to learn the processes first. The director of photography needs to know everything about the cameras and the capabilities of the creative team. This is especially important for the independents, who can save considerable portion of their budgets by reducing mistakes. Training in telling stories in 3-D must address fundamentals and basics.
Chapman continued the thread by noting that creators need to understand what works, especially when mixing 2-D and 3-D. People need to understand the changing methodologies help 2-D producers know what 3-D can contribute.
Viewing distances and rules?
Chapman stated that small screens still have disparity issues, because a home theater is not the same as a living room or bedroom TV. These differences become even worse on personal devices. Hays said that guidelines are not the same as rules, because he always need flexibility. Most 3-D productions have a parallelism budget of about 3 percent. The location of the focus point is in front of the screen for cinema and is mostly in front of the screen for TV.
Gaming tailors the screen to the game. The PS 3 game controls everything for one person. How do these things influence some?
Postley considered increased personalization. The difference in screens from 4 inches to 40 inches is much much greater than 40 inches to 40 feet. In all of these platforms, there are dark color issues. It would be nice if the industry can develop glasses to personalize the view just like games. The greatest challenge is a spectrum of responses across individual viewers.
Chapman responded that monocular vision still matters, because not all images just use binocular cues. Connolly argued that the industry is just starting to understand vision. People need depth cues, so directors need to understand framing and perspective to supply these. 3-D requires longer shots rather than quick cuts, but directors still need to take creative risks, because the technology keeps improving. Stereographers need to know what the equipment can capture.
Side-by-side images moving to 2-D plus depth?
Postley demonstrated the technical issues. Rigs have to change, replacing an objective camera with a witness camera. The relationship of the cameras, matching, rig alignment, and tracking would be very difficult. In addition, you need to ensure color and image correlation. There are many combinations of things to do, so cameras can change but using different cameras is a bad idea. Exhibition changes and technology changes over time. An alternative is to tailor data to deliver information to the platform. This made the data describe what to do. The problem with Whitefield optics is that they will generate multiple terabytes of data.
Hays agreed that a different solution might be important, 2-D plus depth is not working. It's very difficult to reconstruct the images from different cameras, and you may need greater than 4K resolution. An alternative are the next generation optics, which are fully digital and capture a light field, which allows total control of everything in the image. This technology can do 2-D plus depth in postproduction. The biggest problem is how to record the images at sufficient resolution. Today's displays do not have enough processing power to compensate these issues.
Connolly interjected the optics and resolution for many alternatives exist but all of them needs a lot of work before they're ready to be used in production and consumption. Chapman offered the possibility of copying biology. Other methods of capture or available but all have resolution issues. For a cable TV, you can get away with half resolution so 2-D plus depth is possible. The content must start from left and right images. The industry still needs lots of research on compressing the data and transforming or transcoding it.
How to develop 3-D literacy in young viewers?
Chapman considered it will happen for all ages. Hays stated that 2-D is a learned behavior that has been developed over the last 100 years. High health and visual safety is the greatest concern. Most artists are relatively young, but hundreds of people have been viewing 3-D content for over two years with no issues. People learn how to visualize 3-D over time. In the gaming world, users tailor the depth, but could establish settings that cause harm. Postley quipped that an extremely 3-D literate person is called a stereographer.