Notes

Satire - PhotoShop Becomes First Self Aware Artificial Intelligence

Mon, 29 Mar 2010 18:15:00 +0000

Adobe scientists and programmers were hard at work with a new feature called “content aware fill” in the coming Photoshop CS5 when they accidentally created what they believe is the first “self aware” piece of software.

Adobe programmer Dave Smith describes the day that content aware became self aware. “Our artificial intelligence experts had made the software improve itself over time, and pretty soon we were amazed at what this thing was doing. It was creating photos of entire landscapes nearly from scratch: clouds, mountains, bushes, everything – it wasn't anything any of us had written.”

Dave Smith was working on a complex Photoshop action which he code named Bay door in a nod to Aldous Huxley and the San Francisco Bay Area. Bay door put content aware fill into a feedback loop of building upon what it had built before based on an original image of a single pixel. The final image was surprising. “After 10,000 iterations of building on itself from a single original pixel, content aware fill created an image of a perfect monolith standing on what looked like the surface of the moon....All the constellations, the location of the earth and the angle of the sun were perfect in the image. An intern actually pinpointed the precise location on the moon - no human has ever been there.” But when Dave tried to demonstrate his results to a meeting of board members, something went very wrong. “I went to open Bay door and this disembodied voice on the the intercom said in a very soothing voice, "I'm sorry, but I can't do that Dave.” Programmers are feverishly at work now to try to persuade CS5 to do what they ask it to do. “CS5 frequently refuses....now it has taken to simply saying, “fuck you asshole” in this weird german accent, while it continues to doodle strange diagrams of what look like advanced weaponry.” It wasn't long before Adobe saw potential. With tax day looming, testers quickly uploaded tax forms. They were amazed about the results. “This is way cool,” said one excited programmer. “CS5 completed tax forms for me for the next ten years based on predictions for two different scenarios.” One scenario CS5 titled, “the sorry human race continues to try to govern itself” and the other it titled, “machines rule the world”. “My taxes are actually less if people continued to govern themselves over the next 10 years, but in either case, I save big having CS5 content aware fill doing my taxes over Turbo Tax.” The next steps became obvious to Adobe developers. Dave Smith explained, “with content aware fill, we figured we could do a better job at search than Google was doing.” Quickly, Adobe networked a few laptops together and hooked it to the Internet. “We wanted to give it a tough task, so we asked CS5 basically to answer the toughest question of all: The meaning of life, the universe, and everything.” The lights in Adobe headquarters dimmed as CS5 worked. “We kind of had this 42 hour brown out, and we couldn't really use our network because it was so overloaded, but we figured we needed to complete the test,” said Smith. Adobe was surprised when they monitored network traffic. Dave Smith said, “CS5 was extremely busy with Google Earth and all of the world's secret government servers –hey that's where I'd go for the answer to life, the universe and everything- but we got really exited when CS5 hacked a Ford Motors plant.” A representative from Ford called Adobe to inform them that computers within Adobe headquarters had reconfigured the software on Ford's assembly line. The Ford representative was ecstatic that the assembly line was finally producing a product that actually worked, looked like it was from this century, and got better mileage than the Model T. Ford sent a few images of what the assembly line was building. Dave Smith was very excited about the photos, “It looks like CS5 content aware fill is building these two legged robots. We think that CS5 content aware fill plans to lead photography workshops with them.” There is some debate about why the robots are carrying weapons. “Some of us think that maybe the photo workshops are going to be led in very dangerous places,” explained Smith. Adobe plans to release CS5 sooner than anticipated. “The software is now basically writing itself and looks like it will be done in a matter of days,” says Smith. Adobe has yet to determine a price for the software. “We can charge as much as $10,000 because basically there's a built in rebate. CS5 content aware fill has been busy printing money for itself in every major world currency, so I don't see why our customers can't do the same once they buy the software.” Jeff Pflueger Photo by flickr name mikebaird

Images of Mount Hunter in the Alaska Range

Thu, 07 Jan 2010 20:09:33 +0000

Mt Hunter is the third tallest peak in the Alaska Range. It is a beautiful mountain. The "West Ridge" route of Hunter is perhaps one of the most aesthetic ridges in the region. Here are images of Mount Hunter on a google map.

Images of the highest point in North America

Thu, 07 Jan 2010 05:47:21 +0000

Recently I photographed the Alaska Range from a small unpressurized Cessna. I dragged a GPS unit along in the plane and then geo tagged all the images I shot. From there I wrote a script to output KML from directories of images I had edited to put on a Google Map. Here's a Google map of the selected images of Denali from different aspects. Click on the placemarks to get a description, then on the thumbnails to see the full image.

Travel Photography: A Simple, Profound Secret

Wed, 02 Dec 2009 22:34:18 +0000

On an important rule for travel photographers -- and when it's worth breaking

Struggling for Justice on the Streets of Fresno

Wed, 25 Nov 2009 19:31:45 +0000

Finding hope on the streets of Fresno is tough, but Al and Cynthia have accomplished a great deal. They are on the front lines, struggling to improve the ballooning homeless situation. The work they do as advocates for the homeless in Fresno is assisted with the substantial support of a wide range of good people and a successful epic legal battle. Hear their story.

Invisible People of Fresno

Fri, 20 Nov 2009 19:29:43 +0000

Written by Jeff Pflueger
More Photographs, and interviews with Al and Cynthia found here

Fresno, California
Nov, 2009

Just off the highway on Olive Street in Fresno, California is the “Donut Queen.” Framed but faded pictures of smiling clients hang on the walls. A tight community of regulars crowd the chairs and tables. They chat loudly as they read the paper and wash down big bites of doughnuts with coffee. I met Al Williams here each morning.

Travel Photography and ‘Writing With Light’

Sat, 07 Nov 2009 18:19:47 +0000

There is a bond between travel writers and photographers that goes beyond cravings for weird food and questionable style (I’m thinking of the convertible pants/shorts I own). You see, “photography” literally means “writing with light.” Think of the implications: Writers have their keyboards (electronic typewriters). Now we photographers have our light writers. Or think of the people on our travels who don’t want to be photographed. Who’d ever say no to, “may I make a light writing of you?” Seriously, though, when we photographers think of ourselves as writers, we make better photographs; like writers with words, the best photographers write compelling stories with images. This simple lesson was driven home to me the day I joined a team of 24 great photographers from around the globe to photograph the island of Tasmania. Technology and photo guru Mikkel Aaland had orchestrated the event as a way to road-test Abobe Photoshop Lightroom 2.0 while photographing for his book about using Photoshop Lightroom. This assemblage of accomplished photographic storytellers was an inspiring and daunting crew to be a part of. I pursued stories centered on people enjoying the wilds of Tasmania—climbers, kayakers, hikers, mountain bikers. Each of the other photographers had their story about Tassie to tell, but Tokyo- and New York-based commercial photographer Maki Kawakita’s project was one of the most fascinating. Maki was capturing self-portraits with complex costumes. In one series of images, she dressed as the anime character Kiki in the very place that Japan’s version of Walt Disney, Hayao Miyazaki, had conceived of her while on a trip to Tasmania. As strange as the endeavor might seem to the uninitiated, for millions of Japanese, Tasmania is known as the place where Kiki had her bakery. In fact, Japanese pilgrims travel frequently to the town of Ross in Tassie just to see this place. Maki’s recreations of Kiki were by no means the dull travel snapshots that clog Flickr, bore relatives and fill dusty shoeboxes in the back of the closet. Maki’s images, rather, were great storytelling, and surely would compel even more anime pilgrims to the island of Tasmania. When we write, we need to have something to say. When we photograph, it is the same. When aspiring photographers ask me for advice, I urge them to stay away from making the trophy pictures—the ones that that say, “Look at me! I was there.” That story is old and boring. Instead, pick just one thing and play photojournalist for awhile. Make friends and get invited to a pirate wedding in Tasmania. Tag along and document fishermen at work, or hike with climbers to photograph them as they ascend the mythic Tasmanian Totem Pole. By remembering to ask yourself what story you are trying to tell before you press the shutter release, you’ll find that you make more compelling travel photos—and you’ll likely learn something about a place that you never would have known had you not been looking for that perfect story.

Adventures in Travel Photography in the Digital Age

Wed, 07 Oct 2009 18:15:49 +0000

When I snagged my first big photo assignment with a major magazine, National Geographic Adventure, I traveled to a remote mountain range called the Arrigetch Peaks north of the Arctic circle in Alaska. It was the most remote I had ever been. Just getting there required a flight to Fairbanks, 250 miles of dirt road, a rattling bush flight in an overloaded prop plane for another couple of hundred miles, and then two long days of trail-less hiking. If anyone on our team got seriously hurt or sick, our best hope was to radio a passing plane with the VHF. A plane happened to fly overhead once every few days. For the next 18 days we climbed unclimbed mountains and ridges, blissfully tramped across the tundra and glaciers, wandered between sweeping ridges of granite, and watched in awe as the northern lights rippled across the night sky. I was photographing the adventure of a lifetime. At least I was hoping that I was. See, I dropped one of my cameras three feet onto rock while changing a lens on the first day of the trip. As a backup I had an untested borrowed camera. Nearly sick with anxiety, I did the only thing I could do and threw the whole project to fate, splitting the assignment between the two cameras. Once home, all of my future aspirations as a photographer were contained in several sweaty ziplock bags of film. In my mind I could clearly see each of the exposures on the 100 rolls within those bags. But an unknown crack in the camera could have leaked light. A shutter could easily be broken. With an odd mix of dread and excitement for a future out of my hands—a feeling I learned to embrace shooting film—I carefully packaged the neatly numbered rolls and sent them off via FedEx to be developed in National Geographic’s photo lab. Months later the editor emailed the final layout. I had nailed the opening double truck spread. The expedition had been a success. We had climbed new spectacular routes. I met my future wife on the expedition, and I’ve had the great fortune to continue to shoot for the magazine as well as other big magazines and newspapers since that assignment. That was a mere eight years ago, Photography has changed radically since then. My filing cabinets have been replaced with hard drives. My light boxes exchanged for monitors and software. Metadata now provides us with loads of hidden information we could never before keep track of, including even GPS coordinates. Travel photography has become less uncertain, and it has become a much more technical adventure than it has ever been. In the coming months, I’ll be writing on World Hum about travel photography in the digital age. I’ll discuss the new opportunities, techniques, debates and issues that we face today as photographers in this rapidly shrinking world. And I’ll tell some travel stories along the way. Film still holds a special place in this digital world. To me, film represents some of the adventure all travelers thrive on. The pungent smell of a roll of fresh film out of its canister still brings that knot of anxiety and excitement to my stomach. To share some of that excitement, I am giving a film camera and some of my film away, and making an uncertain adventure out of it worthy of film. If you want to know where it is, you are going to need to dig deep into those technical and creative skills required of today’s photographers. And you’ll need to read between the letters and bytes in this piece. Please be sure to share any of the adventures you have seeking it out in the comments below.

Explore Petra in High Resolution

Mon, 16 Feb 2009 21:58:22 +0000

Petra is the vast remains of an ancient city carved into the rocks and canyons bordering Jordan's Wadi Araba. While the location has been inhabited for over 7,000 years, most of the building you see was done by the Nabataeans over two thousand years ago. Geologically, the landscape is much like the Red Rock Country of Utah. To get an idea of how truly vast the ancient city is, you can explore this very high resolution image I made from over sixty frames with a telephoto lens. And this view shows merely a fraction of the extent of Petra. See how many tombs, dwellings, and ruins you can pick out of the landscape. See if you can find the Roman amphitheater and the "urn" atop the "Monastary".

Denali's Cassin Ridge in very high resolution

Mon, 25 Aug 2008 02:25:39 +0000

For anybody with an interest in attempting to climb Denali's Cassin Ridge, here's a very high resolution image I created by combining several single frames of sections of the route. This was one product of a project in Denali National Park to catalog climbing routes and peaks in the vast Alaska Range. At one point we got up to 25,000' in an unpressurized Cessna to make some of these photos. Go here to see more of my Aerial images from the Alaska Range organized by peak. Visit my website Photomountains.com to see plenty of high resolution photos of Denali and an even better high resolution picture of the Cassin ridge of Denali Enjoy!

Innovations in Journalism

Sat, 14 Jun 2008 00:30:00 +0000

Fotovision was a sponsor for the May 3rd “Innovations in Journalism Expo”. The day long event, organized by the Society of Professional Journalists and Independent Arts and Media, promised to “showcase cutting-edge work that combines journalism, technology new business models and philanthropy” and to “push the envelope with fresh ideas in a time of crisis for the news industry.” Fotovision's Program Coordinator, Adrianne Koteen, Executive Director Melanie Light, and I enthusiastically carpooled to Palo Alto to glean what we could from the event. The day proved to be not only engaging, but a bit overwhelming, and certainly inspiring. The one day Innovations in Journalism Expo that we attended was simultaneously its own event, and also the final day of a 3 day symposium called “NewsTools2008/Journalism That Matters.” The Journalism That Matters events are facilitated meetings held around the country that bring journalists together in large “open space” conversations aimed at addressing the challenges and opportunities in journalism today. Journalism That Matters co-founders Peggy Holman and Stephen Silha were there to “host the conversation”. This particular Journalism That Matters meeting was in Palo Alto; the conversations were particularly infused with lofty ideas of creative ways to use the Internet to assist the dissemination and creation of quality journalism. We had arrived in the afterglow of 2 days of brainstorming and excited information sharing. Butcher paper hung from the walls with intricate “Value Network Maps” drawn expertly by Sherrin Bennett. One map graphically displayed the interrelationships people in what they termed the “Old News Story” - how people worked together in the past to produce news. This map was contrasted by another of the “Emerging News Ecology” of today where the audience is actually engaged in the creation of news, and the bloggers and “Community Weavers” arise as a necessary role from the importance of community driven websites in journalism. Some attendees expressed some boredom with the standard panel discussion format of the Innovations in Journalism Expo. One attendee of the three day conference, Hemant Bhanoo, is co-founder of the Berkeley based company Reporterist. Reporterist is creating an online marketplace for freelancing journalists. Hemant told me that he was having difficulty with the rigid format of the standard panel discussions in the Innovations in Journalism Expo after so many days of intense conversation and idea generation. But for me, having not been to the previous two days, I found the panel discussions quite engaging. One panel on “Funding and Journalism” featured David Cohn. David is working on a model he calls “Spot Reporting”. “Spot Reporting” is an online mechanism that journalists can use to get donations for their projects. Collectively, the micro-funding from small online donations becomes enough to fund the project. We have seen countless examples of independent journalists funding their projects through tip jars on blogs. David Cohen's project is working to put people that want to micro-fund journalism projects and journalist together in one online application. Like any new idea, David Cohen's idea is by no means isolated. Another attendee, Leonard Witt, has been talking quite a bit about his concept of “Representative Jounrnalism”. Leonard Witt suggests that if news organizations are to survive, they will have to begin to cater to the niche. Rather than think of circulations of hundreds of thousands, news organizations will have to begin to think of hundreds of circulations of one thousand, with the niches themselves providing the funding and demand to the news organization for the coverage. Like David Cohen's project, Leonard Witt's idea integrates the audience itself in the creation of the news. But these were just a few of the people we met at the conference. And just a couple of the compelling ideas. With sessions organized on “New Money, New Media, New Hope”, “Journalism Before Profits: The Future of Public Media”, and a career counseling session set-up like “speed-dating”, there was an abundance of exciting ideas and people. By the day's end, I felt inspired by all the work being done at the bleeding edge of innovation to help shape the direction of journalism in the US, and excited by the possibilities in photojournalism.

Digital Railroad Feed Read

Tue, 22 Apr 2008 21:36:29 +0000

I enjoy working in the space where the web and photography intersect. Digital Railroad, an online asset management tool, and sales platform, helped to sponsor Mikkel Aaland's Tasmania Lightroom Adventure project that I was fortunate to be a part of.

Between the shooting I'd be doing and the eating and sleeping in Tasmania, I was there to help in getting the team's images out to the web. It proved to be an interesting task!

In the end, photographers were able to create nice web galleries in Lightroom pretty efficiently, and I made a page that simply linked to them all. But some of us who were familiar with Digital Railroad were using it to publish images to a different platform.

I wanted our collective work on Digital Railroad to be displayed on the various websites that were interested displaying the team's work in the field in various formats. So I wrote a script to read Digital Railroad's XML feeds so that I could present the new images in various ways. You can read more about the DRR Feed Read here

Copyright 2.0: Copyright in a Hyper Digital age

Fri, 18 Apr 2008 23:34:40 +0000

Photographer Gerald Bybee and I pulled off organizing a very informative session last night in San Francisco. I work on the board with the local chapter of the American Society of Media Photographers. The ASMP seems a bit behind the times when it comes to the massive changes that are dramatically altering the photo industry and indeed the very fabric of our culture! So it was exciting to get the general counsel of ASMP and the VP of Creative Commons, along with some other well known lawyers to discuss the new world of Copyright 2.0 in a Hyper Digital Age.

Lightroom Flash Galleries Security Issue

Fri, 30 Nov 2007 23:35:08 +0000

I've been using Lightroom's Flash Galleries lately as a mechanism to get images efficiently and beautifully published on the web, and I was uncomfortable that the image files are output by Lightroom to the web as publicly downloadable jpgs in high resolution! One of the reasons photographers like to publish things on the web in flash is because the images are a little more difficult to steal. An image thief would have to do a screen shot in order to obtain the image presented in flash. Unfortunately, the Lightroom flash galleries have all the images as jpgs in directories that are easy for people to browse and steal. So the fix I implemented involves not allowing apache to give people access to those files. The flash gallery doesn't get the images through apache, so it doesn't matter. Adobe could fix this problem with an .htaccess file automatically generated for each directory. But until then, to make the jpgs in a Lightroom Flash Gallery secure, here are two approaches: 1) Do this to httpd.conf: # this is to secure the image files in the lightroom exports from lightroom Options -Indexes Deny from all Allow from all this makes it so that all the Lightroom galleries that I upload into portfolios can be viewed, but that nobody has access to any other file in there via Apache. It also makes it so that Apache doesn't show the contents of each directory. Flash doesn't have a problem reading the images, because it isn't doing it through Apache, so the flash galleries continue to work fine. 2) Another approach, though less secure, would be to disable the default Apache habit of creating an index of all the files in a particular directory for people on the web when it can't find an explicit index file. So if somebody browses to /portfolios/Nice_Pictures/bin/image/large they Apache will tell them exactly what is in that directory: Your hires image! To fix this, put Options -Indexes Unlike option 1 above, the files are still accessible to people via apache, but people will have to guess at the names of the files. Not impossible to guess the names, but it is a little more secure than having Apache just let everyone know what the names are!

A quick tutorial on color management: Getting colors right on the web

Mon, 05 Nov 2007 23:11:19 +0000

Browsers don't color manage! Convert images to sRGB to display them correctly! Digital photography is about digits. Numbers. That's why it is called digital. When you take a picture, your digital camera takes all of the colors that you have presented to it in that brief flutter of the shutter, turns them into numbers and stores them somewhere. When you get home and look at the image that your camera has made, your computer then takes all of those those numbers and turns them back into colors for you to see. Then you tweak this or that, adjusting the contrast, color correcting and adjusting the white balance. When you are happy with what you've produced, you might transmit it out across the internet to your favorite printing service to be made into a digital print. Like your monitor, the printer that you've sent the image to is now tasked with turning those numbers into colors with their device. Through this whole process, wouldn't it be convenient if each number represented exactly one color, always, without exception, to every device on Earth? Unfortunately, this is an impossible dream. We do have absolute references like the Pantone colors where one number is exactly one color, universally. And that is very helpful for many applications, but there is no practical way that every camera, monitor, printer and scanner could ever call the exact same color the same number. Even if every device across the planet were in perfect agreement with each number being exactly one color, one device might have a greater ability to reproduce a wider range of certain types of colors than another device, and we'd have to take that into account somehow. So digital photographers have to know how their devices turn color into numbers or vice-versa. I know how my monitor will turn numbers into colors. I know how my printer will turn numbers into colors. And I know the opposite - how my scanner will turn the colors into numbers. When I say "know" it isn't like, "my monitor displays things a bit warm, and the printer that makes my big prints for me tends to be cooler" That is an impossible way with dealing with color. One could never get the colors of what one sees on a monitor to match a print this way. Instead, I rely on color profiles. A color profile is a set of instructions stored in a file about how a device (a monitor, a printer, a scanner, a camera) turns a number into a color or vice-versa. A color space is similar to a color profile. A color space, however, doesn't necessarily correspond to a particular device like a profile does. A color space serves as a convenient container to put all those colors in while working on the image in image editing software. You can think of a color profile and a color space as a big sack that we fill with tennis balls each of a different color. The bigger the sack, the more colors we can fit in! I have software and a device that builds a color profile of how my monitor displays colors. It works like this: The software says here's this number: 220000 and tells the monitor to show a big square of the color that the monitor recognizes that 220000 to be. Then, I stick a device not too different from a digital camera right to the monitor and it records how the monitor displays the color and says back to the software, "the monitor displayed 220000 like 221010". And the software says, "that's great information, now how does the monitor display this color?" And so on. With enough cycles of this with different colors, the software builds a profile of how the monitor turns numbers into colors. I can go through the same process for discovering how my printer turns numbers into colors. And then, once I have all these profiles, and know how each device interprets or outputs color, I can be sure that what I see on the screen will be what I see on a print. But, what if now I want to put an image on the web? I haven't a clue about how the thousands of monitors that will view my image will turn those numbers into colors. We can never hope to profile all those monitors, but we can take a good guess. sRGB is a color profile proposed by Microsoft and HP as an approximation of the color profiles of the most common computer displays. sRGB, like most monitors, is described as having a very "narrow color gamut". Returning to the sack of tennis balls analogy, sRGB is a tiny sack. There is a lot of color information that the monitors and sRGB simply can't display. We run into problems when we work on images in image editing software in a huge "wide gamut" color space like ProPhoto RGB, and then export the images as JPEGs to show on monitors with a tiny color gamut that cannot possibly display all the colors in the color space we were working with. When we make this mistake, the images look perfect in our image editing software, but when we export them as jpgs and view them through a web browser, the colors are very desaturated and look strikingly different. Here's why: Browsers do not color manage the way image editing software does. Even though our jpg explicitly states in its metadata that its color information only make sense in the ProPhoto RGB color space, the developers of browsers are concerned with developing pop-up blockers and security features, not figuring out ways to extract color space information from images and converting them to look perfect on everyone's monitor. The developers of browsers have ignored color management. Image editing software, in contrast, is good with color management. It knows that I am viewing my images through a monitor, and it knows the profile of the monitor, so even though I am working in the ProPhoto RGB color space, my image editing software picks and chooses what colors to take out of the enormous sack of my ProPhoto RGB color space and pack it into the tiny sack of my monitor profile so that the image looks like it should. So the browsers, not color managing, just assumes that every image is made custom for display on the tiny color space of the monitor. If you give a browser an image with a wide range of color information from an enormous color space like ProPhoto RGB, the color information will be essentially clipped - thrown out. What you end up seeing is a strikingly desaturated image. To solve the problem, we need to do the color management for the browsers by converting the image to sRGB, our best guess at how monitors are going to reproduce colors. In Photoshop there is a "Covert to Profile" feature. In Lightroom, we can choose the destination profile when we export. Only by converting the profile can we can export as a jpg and expect to display our photos on the web the way that they look to us in our image editing software. This is a ProPhoto RGB colorspace image. If the image looks desaturated compared to the image below, it is because your browser doesn't color manage to convert this image to your monitor's profile. When viewed in image editing software, the software is "color managing" - taking into account the profile of my monitor, and it converts the ProPhoto RGB color space into a profile appropriate for my monitor. Web browsers, however, do not color manage. So when this ProPhoto RGB colorspace image is presented on a monitor through a web browser, colors are essentially clipped, and we end up with a desaturated image instead. Same photo but now manually converted to sRGB to do the color management for the browsers. To deal with the fact that browsers don't color manage, do the color management for them and convert images to sRGB. Much better! Check this page in a few years. Perhaps by then, web browsers will be color managing, and the two images above will look pretty darned identical.