Now more than ever, it would behoove photo-makers and -viewers - i.e. everyone - to read Sontag's "In Plato's Cave" as a reminder that there is nary a thing as a real photograph.
> Whitney feels that researchers are overcoming the computational obstacles and that computers, despite their considerable expense, have already been proved cost-effective. With a computer, one will need a much smaller movie staff. Lighting crews, carpenters and cameramen are expendable. One can film in any weather and dispense with travel to exotic sites.
thanks for posting this! in reality over the last 40 years, movies got bigger and employed far more people
> With a computer, one will need a much smaller movie staff.
Part of the problem is that their estimate for the number of computers was off by many orders of magnitude. In a decade or two, with all of the generative tools, they might be much more wrong on the compute side of things, but more right on the direct-staffing side.
If you can do the same thing as today's biggest movies with fewer people, it means that you'll need to do more to separate your big movies from the rest of the pack.
TV and movies today use extensive visual effects for stuff that would've been shot on-set or on-location and not retouched at all 20 years ago. All the soccer field and stadium stuff in Ted Lasso, for instance. The 20 year ago version probably would've had stock photos/video and hard splits between shots of "the stadium fans" and shots of "the coaches by the sideline," and nobody would've complained... but try to get away with that today as a top-end TV show and you'd get panned for how fake and cheesy your penny-pinching looks compared to your competition.
I certainly don't think I can predict what that "more" looks like in twenty years, but I doubt it's purely "throw more generative tools at it." It could be something like "throw an army of hundreds of people to review, collate, and cherry-pick the coolest out of hundreds of thousands of generative outputs" - if everyone has access to good tools, then maybe the truly breathtaking stuff will be the stuff from people who can afford the time to curate to a specific vision.
> The National Aeronautics and Space Administration regularly employs similar techniques to clarify images transmitted from outer space. The computer can sharpen fuzzy pictures or add color to satellite pictures of land masses so that it is easier to decipher the details.
lololol
> It is no accident that unfamiliar environments such as outer space or the inside of a computer, with its high-tech components, have been favored for the computer-generated segments of movies.
> In search of aids to more realistic simulation, researchers comb the various branches of mathematics, and even art. For instance, Geoffrey G. Gardner of Grumman Aerospace in Bethpage, N.Y., has been analyzing Impressionist painting techniques in order to understand how it was possible to use such large brush strokes and still represent the essence of a scene without reproducing unnecessary detail. Gardner would like to apply such a technique to the computer.
> Though no researcher has yet synthesized an image of a human being that anyone would mistake for a photograph of a person - real or fictional - scientists at the University of Pennsylvania Computer Graphics Research Facility have created a humanlike figure that moves on a video screen through zero gravity like an astronaut. This figure will be used for tests within a simulated space shuttle, such as checking on the size of passageways and the accessibility of instrument panels.
hehe
> Recently, Whitney says, his production company used the data from a NASA photograph to simulate the surface of Jupiter for the soon-to-be- released movie ''2010,'' a sequel to the film ''2001.''
> "If we really want to, we can convince folks that it's real" .. What interests Smith is "going to be to find out how far you can go away from reality and still have people follow you."
Adobe Photoshop has just introduced their "generative fill" tool in the new beta version. It requires an internet connection and it goes out to the web and uses AI to fill photo regions with whatever text request the user inputs. While in a way it does nothing that could not have been done before with PS, it now can do in seconds - and by someone like me with very limited PS expertise - what previously may have taken hours and required considerable PS skill. We've come a long way since the predictions of 1984 (and not just as regards Orwell).
The affect of computers on photography has grown dramatically in the last decade, its been very impressive seeing all the 'tricks' that we have incorporated into mobile phones to deal with their shortcomings.
But, more excitedly, we are seeing these advancements coming into the professional field - such as with Sony's new mirrorless A7RV which has a dedicated AI processor for autofocus.
All these additions help aid in composition, understanding what a scene needs and helping produce the most accurate image. As someone who still regularly shoots with film, it is amazing to go back and see just how many additions we take for granted when taking photos today compared to in the past.
23 comments
[ 4.8 ms ] story [ 60.8 ms ] threadhttps://openlab.citytech.cuny.edu/chengphotoarth1100f2019/fi...
thanks for posting this! in reality over the last 40 years, movies got bigger and employed far more people
Part of the problem is that their estimate for the number of computers was off by many orders of magnitude. In a decade or two, with all of the generative tools, they might be much more wrong on the compute side of things, but more right on the direct-staffing side.
TV and movies today use extensive visual effects for stuff that would've been shot on-set or on-location and not retouched at all 20 years ago. All the soccer field and stadium stuff in Ted Lasso, for instance. The 20 year ago version probably would've had stock photos/video and hard splits between shots of "the stadium fans" and shots of "the coaches by the sideline," and nobody would've complained... but try to get away with that today as a top-end TV show and you'd get panned for how fake and cheesy your penny-pinching looks compared to your competition.
I certainly don't think I can predict what that "more" looks like in twenty years, but I doubt it's purely "throw more generative tools at it." It could be something like "throw an army of hundreds of people to review, collate, and cherry-pick the coolest out of hundreds of thousands of generative outputs" - if everyone has access to good tools, then maybe the truly breathtaking stuff will be the stuff from people who can afford the time to curate to a specific vision.
lololol
> It is no accident that unfamiliar environments such as outer space or the inside of a computer, with its high-tech components, have been favored for the computer-generated segments of movies.
> In search of aids to more realistic simulation, researchers comb the various branches of mathematics, and even art. For instance, Geoffrey G. Gardner of Grumman Aerospace in Bethpage, N.Y., has been analyzing Impressionist painting techniques in order to understand how it was possible to use such large brush strokes and still represent the essence of a scene without reproducing unnecessary detail. Gardner would like to apply such a technique to the computer.
> Though no researcher has yet synthesized an image of a human being that anyone would mistake for a photograph of a person - real or fictional - scientists at the University of Pennsylvania Computer Graphics Research Facility have created a humanlike figure that moves on a video screen through zero gravity like an astronaut. This figure will be used for tests within a simulated space shuttle, such as checking on the size of passageways and the accessibility of instrument panels.
hehe
> Recently, Whitney says, his production company used the data from a NASA photograph to simulate the surface of Jupiter for the soon-to-be- released movie ''2010,'' a sequel to the film ''2001.''
> "If we really want to, we can convince folks that it's real" .. What interests Smith is "going to be to find out how far you can go away from reality and still have people follow you."
quite far indeed!
I've been waiting for the new Jimmy Stewart/John Wayne movie for years now. AI, step up your game!
But, more excitedly, we are seeing these advancements coming into the professional field - such as with Sony's new mirrorless A7RV which has a dedicated AI processor for autofocus.
All these additions help aid in composition, understanding what a scene needs and helping produce the most accurate image. As someone who still regularly shoots with film, it is amazing to go back and see just how many additions we take for granted when taking photos today compared to in the past.