Category Archives: Enormousness

Adventures in Upsampling

Interesting recent finds:

  • Google Zoom Enhance. “Using generative AI, Zoom Enhance intelligently fills in the gaps between pixels and predicts fine details, opening up more possibilities when it comes to framing and flexibility to focus on the most important part of your photo.”
  • Nick St. Pierre writes, “I just upscaled an image in MJ by 4x, then used Topaz Photo AI to upscale that by another 6x. The final image is 682MP and 32000×21333 pixels large.”
  • Here’s a thread of 10 Midjourney upsampling examples, including a direct comparison against Topaz.

A giant space telescope made from plastic

Normally I’d call a 600mm lens big—but this bad boy dwarfs even the Hubble: 

DARPA says the MOIRE program is currently in its “second and final phase,” and has already been prototyped. When completed, such optical systems could result in more powerful telescopes that fit in smaller rockets and shuttles, with eventual use in tracking weather systems or for reconnaissance. Here’s a clip explaining more about how it works:

 

[YouTube]

DARPA tech: Cool with a side of creepy

Generating 600 GB of data per second (eat your heart out, RED cam), the 1.8-gigapixel ARGUS camera will hitch a ride on drones, spotting targets as small as six inches from an altitude of 20,000 feet. TechCrunch reports,

The camera uses 368 five-megapixel camera sensors aimed through a telescopic array to pick out birds in flight and humans on the move on the Earth’s surface. ARGUS stands for Autonomous Real-Time Ground Ubiquitous Surveillance Imaging System.


Elsewhere, acronym-lovin’ DARPA wants you to design an amphibious Fast, Adaptable, Next-Generation Ground Vehicle (FANG). Somehow I’m kinda weirded out by their appropriating a breezy whiteboard aesthetic more often seen in TED talks:

Save big Photoshop files up to 20x faster

Time is money, and storage is cheap. If you use large Photoshop docs & don’t mind them becoming much bigger, you may be able to save them much faster.

You’ll want to update CS5 to 12.0.4 (via Help->Updates), then grab this plug-in (Mac)/registry key (Win).  Photoshop PM Jeff Tranberry writes, “Saving the file can be much faster (20x in some cases), if you have a fast hard disk and enough disk space to hold the larger file size,” and performance expert Lloyd Chambers goes into detail on his site.

A Blade Runner tribute made from a giant image

François Vautier built one enormous PSB* file to create this short film:

Created by extracting 167,819 frames from ‘Blade Runner’s final cut version, then assembling all these images to obtain one gigantic image of colossal dimensions : a square of approximately 60,000 pixels on one side alone, 3.5 gigapixels. A virtual camera was then placed above this big picture which creates an illusion, because contrary to appearances, there is only one image. It is in fact the relative movement of the virtual camera flying over this massive image that creates the animated film, a kind of “zootrope effect”, like a film in front of a projector.
The whole concept echoes one of the signature scenes from the film where “Deckard” (Harrison Ford) analyzes a photograph via voice recognition software.


* Like PSD, but scaling beyond 2GB and meaning “Photoshop Big.” Yes, really.

A great digital imaging project honors the fallen

Photographer Peter Krogh (author of the excellent The DAM Book, the Rapid Fixer extension for Bridge, and more) recently completed an ambitious & enormous digital imaging project: photographing all 58,256 names listed on the Vietnam Veterans Memorial, enabling the creation of an interactive online version of the wall.  By stitching together some 1,494 digital images into a 400,000 pixel by 12,500 pixel monster, Peter & colleague Darren Higgins were able to help create a Flash-based presentation that enables you to search for names, read servicemen’s details, and add notes and photos to the wall.

The presentation site features some behind-the-scenes production info, but figuring there was more to the story, I asked Peter for details.  He kindly provided them in this article’s extended entry.  Read on for more.

Continue reading

Gigapixel panos through Flash

GigaPan.org is "sort of a Flickr for zoomable panoramas," notes Photoshop engineer (and Photomerge creator) John Peterson. The site makes it possible to upload & browse gigapixel-sized images, then navigate through them via a Flash interface.  Here’s a shot of Adobe HQ, taken from nearby Caesar Chavez park* in downtown San José.  (Bustling, isn’t it? ;-))  The site is labeled "beta," and the viewer currently leaves much to be desired (quit squirming around, dammit!), but it’s a very cool project nonetheless. [Via]

For more in this vein, see previous: Colossal images through Photoshop & Flash; 13 gigapixels or bust; 3.8 Gigapixels of Half Dome.

* I’m sure I walk by it all the time, but until seeing this image I never noticed the deeply gross sign in the park.  Click the second of the two snapshots below the Adobe pano to read it.  I’ll never think of the fountain in quite the same way.

13 gigapixels or bust; sketchy photogs; more

  • In Spectacle, photographers David Rockwell & Bruce Mau "celebrate the phenomenon and history of communal, awe-inspiring public performance worldwide–from the stadium to the streets, from religious festivals to political marches."  Dig the really well-chosen type treatments as well. [Via]
  • For a different kind of spectacle, see Harlem in 13 Gigapixels. Photographer Gerard Maynard & software developer Alexandre Jenny have teamed up to create a massive image of the famous New York neighborhood.  With results spanning 279,689 x 46,901 pixels, the project’s raw numbers
    are pretty eye-popping:

    • 2,045 individual photos from a Nikon D2X
    • 21.49 GB of compressed raw data
    • 1 day for image placement and color correction
    • 46 hours of rendering on an 8-core Xeon system with 8GB of RAM
    • Results: A single 48.8 GB image stored in the Photoshop Large Document format (.PSB), converted via Zoomify & displayed through the Flash Player.
      [Via Maria Brenny]  (If this is up your alley, see previous.)
  • Ah, the 1950’s, when you had to be the lookout for "corn-fed belles" hanging out of trees along the road, ready to disrobe in your U-Haul trailer.  At least that’s the world conjured up by the (more than a little creepy) Glamour Photography magazine–one "designed to give the camera man a better understanding of the technical and philosophical aspects of photographing pretty girls."  Philosophy–yes, that’s it. [Via]
  • Elsewhere in history, here are 50 years of a woman’s life, as told by photos bought at a garage sale.  Note to self: Keep trying not to get old.
  • Clayton James Cubitt shares portraits of Hurricane Katrina survivors.  (I’m a big fan of Flash galleries in general, but in this case I think the jerky transitions distract from the subject matter.) [Via]

More gigantic typography

  • 6,272 Post-It notes form a giant, editable "TO DO" on windows in Brooklyn’s DUMBO neighborhood.  Passersby are invited to jot their own to-do lists on the notes.  I love it.  (Consider this "Solve Gordian knot of ever-increasing power & complexity in Photoshop; also buy new shoelaces," written in absentia.)   More photos of the work are on Flickr.
  • In her Type the Sky project, Lisa Reinermann captures buildings that form letters against the sky, creating a photographic font.  [Via]
  • For more big letters, see previous type entries filed under Enormousness.

This font goes to 10,116pt.

  • The designers at Pentagram talk about how they created a giant NY Times logo (10,116-point Fraktur) for the publisher’s new headquarters.  Interestingly, each letter is comprised of numerous small, three-dimensional “beaks” that enhance the sign’s visibility from the street.  [Via]
  • How about lettering via “military-like technology for criminal mischief”?  We Make Money Not Art hosts an interview with the Institute for Applied Autonomy.  Their Streetwriter is a giant printer disguised as a cargo van, while GraffitiWriter offers radio-controlled pranking:

    “Studies have shown that in nearly 100% of the cases, a given agent of the public will willing participate in high profile acts of vandalism, given the opportunity to do so via mediated tele-robotic technology.”

  • From the Ministry of Silly Type Tricks: Flip text using Unicode. [Via]
  • Graffiti artist “Eine” has painted a set of very cool East End Shopfront Letters. They can be assembled into words via this little app. [Via]

[Update: In response to Ramón Castañeda’s comment below, Thomas Phinney replies, "Ramón is right. Fraktur typefaces usually have a forked top to the ascenders (h, k, etc.), more curves in the lowercase (less rigidly hexagonal shapes than Textura), and all (not just some) of the caps will have curvy or squiggly shapes replacing vertical lines.  This page even shows the NYT logo among the Textura samples, an unexpected bonus).  Not that I think this is a big deal, by the way. If the worst typographic errors we have to worry about were people confusing different styles of blackletter, we’d be in pretty good shape. :)"]

 

My screen can beat up your screen

The CineMassive MasterPlex 21T (<–double intercaps = double the power) offers flat-screen madness: a 21" LCD center with five 17" arrayed around the sides.  That works out to a cool 8,473,600 pixels for $3,299.  Is it just me, or does it seem like the 20" Apple Cinema Display was just introduced* for $3,999?  [Via John Agger]

I have a soft spot for this stuff, maybe, as I haul around my second 17" Mac portable. (I’m much happier now that I have a bag that doesn’t feel like it’s cutting my shoulder with piano wire.)  Of course, it’s always theoretically possible to carry around an even bigger screen… [Via]

Marginally related: I enjoyed CNET’s story "You call that high-def?," listing some of the more tenuous (okay, entirely unwarranted) ways the "HD" moniker is being slapped onto various endeavors.  And with that, I’m off to pet my GatoHD’s high-def coat.

* 1999–for real? I am getting old.

The 66" negative

AutoWeek has the interesting story of how photographer Rick Graves uses a modified, motorized camera back which feeds a continuous roll of film past the shutter while it’s open, creating a very wide negative (like this one; scroll it to the right):

"Each image Graves makes is from one exposure on an entire roll of film, not a composite of several different images.

"’A number of people have tried to build this type of camera,’ Graves said, likening it to the finish-line cameras used at horse races. ‘But the difference with my camera is that I have 66 inches of movement [of the film] in one second. The film is moving relative to the moving subject. I developed this camera as a better way to capture motion.’

"The secret to the system is not the camera itself—a standard 500 Series Hasselblad—but in the film back, which contains a small motor and various electronics adapted from the robotics industry. This setup gives Graves control of how fast the film moves when he opens the shutter. If he gets it right, the film is moving at the same speed as the cars, allowing for a photo with dozens of speeding cars, all razor sharp."

NASCAR sells prints that are 4 inches tall by 8 feet long.  Check out many more examples (not all automotive) in the DistaVision portfolio. One slight bummer is that because of the ubiquity of Photoshop-edited composites in the world, a lot of viewers may think these works are simply digital collages. [Via Joe Ault]

On a related note, I happened across an article on slit-scan photography that features a rather trippy photo produced using related methods. [Via]

The 66" negative

AutoWeek has the interesting story of how photographer Rick Graves uses a modified, motorized camera back which feeds a continuous roll of film past the shutter while it’s open, creating a very wide negative (like this one; scroll it to the right):

"Each image Graves makes is from one exposure on an entire roll of film, not a composite of several different images.

"’A number of people have tried to build this type of camera,’ Graves said, likening it to the finish-line cameras used at horse races. ‘But the difference with my camera is that I have 66 inches of movement [of the film] in one second. The film is moving relative to the moving subject. I developed this camera as a better way to capture motion.’

"The secret to the system is not the camera itself—a standard 500 Series Hasselblad—but in the film back, which contains a small motor and various electronics adapted from the robotics industry. This setup gives Graves control of how fast the film moves when he opens the shutter. If he gets it right, the film is moving at the same speed as the cars, allowing for a photo with dozens of speeding cars, all razor sharp."

NASCAR sells prints that are 4 inches tall by 8 feet long.  Check out many more examples (not all automotive) in the DistaVision portfolio. One slight bummer is that because of the ubiquity of Photoshop-edited composites in the world, a lot of viewers may think these works are simply digital collages. [Via Joe Ault]

On a related note, I happened across an article on slit-scan photography that features a rather trippy photo produced using related methods. [Via]

Laser graffitti + chrome spheres

File under Enormousness:

  • You’ve gotta love any ingredient list that includes the phrase, "1 60mW Green Laser (super illegal in a lot of places and very dangerous)."  And you’ve really gotta love what the Graffiti Research Lab does with theirs, lighting up a Rotterdam building with all kinds of hand-drawn art.  Big style points for the dripping paint effect!
  • I’ve always really liked mosaics and particle systems, and I used to browbeat a friend in Illustrator engineering to convert their mosaic filter to create symbols (good for turning artwork into particles that could be animated, kind of like these fish).  That hasn’t happened, but in the meantime I can enjoy Danny Rozin’s shiny balls mirror (see video).  Comprised of "921 hexagonal black-anodized aluminum tube extrusions, 921 chrome-plated plastic balls, and 819 motors," the system reflects the viewer twice: once in each ball, and once in the entire piece. [Via].
    See also his earlier wooden mirror (video).  And lastly, his Time Scan Mirror reminds me of the Scanner Photography Project (the site for which is now down, unfortunately).

[Update: Speaking of mosaics, how about a cereal Seinfeld? [Via]]

65MPx Mother of All Touchscreens

The folks at Virginia Tech’s GigaPixel Project have been busy, creating a 50-monitor display prototype.  Comprised of 21" flat-panel touchscreen monitors (the perfect complement to this stuff?), and driven by a cluster of 25 small PCs, the setup promises a resolution of at least 12,800×5120 (65,536,000 pixels).  To afford the sucker, you could do what David Pogue suggested for that 108" Sharp TV: build a new house with the display as one of the walls (waterproofing recommended).  They’ve also done what any good college students should, rigging up a 24-monitor display wall to play Quake.  [Via Jon Williams]

3.8 Gigapixels of Half Dome; Great Flash panoramas

The folks at FlashPanoramas.com sell a utility for displaying spherical panoramas via the Flash Player.  They’ve now updated their technology to take advantage of the new full-screen mode enabled in the latest rev of Player 9.  Check out some very cool examples, or get the tool for €39.95 from their site. [Via]

Elsewhere, Greg Downing & co. at xRez.com are working on Extreme Resolution panoramic image creation.  Check out this 3.8 gigapixel* spherical panorama of Half Dome, displayed via the Google Maps API. 
Although the subject is nearly a mile from the camera position, you can zoom in and see a climber on the face of Half Dome, as well as someone standing on the visor & and hikers along the Merced river in the valley below.
Wicked!  "By the way," Greg writes, "Photoshop large document format [PSB] was a lifesaver on this project!"

The xRez site shows off more examples and goes into plenty of technical geekery for those so inclined.  Greg’s own site offers other interesting bits on HDR panoramas, and this QuickTime slideshow nicely demonstrates how various elements of a scene can be displayed at different exposures.   (Aside: Is that thing a naval mine or an interrogation droid or…?)  A test render of 3D objects lit with an HDR lighting map shows the power of sampling this data from a scene, then feeding it into a 3D rendering package.

*According to Wikipedia, a single gigapixel contains 250 times the data captured by a 4MP sensor. (Of course, at any given moment Wikipedia might claim that I personally have invented over 350 uses for the peanut–but I think it can be trusted in this case.)

Pleasures of the Flesh… with Toast

Clearly the smell of slow-cooking turkey meat wafting down the hall is getting to me, and soon enough I’ll give this laptop a much-deserved break.  But before that, here’s a wee cornucopia of hopefully interesting bits:

And with that, I wish you good eating, good health, & a day free from turkey frying disasters. [Via] Happy Thanksgiving!

–El Tryptophan

Biggest. Typography. Ever.

  • geoGreeting leverages Google Maps, letting you assemble animated greetings by using satellite photos of letter-shaped buildings.  Ridiculous!  Check it out! [Via]
  • Along vaguely similar lines, FireHorse Studio has fun with lettering in a 15 second teaser trailer for Toyota. [Via]

[Update: That giant “MOMO” signature (mentioned recently) would qualify as well. And so would the 8-foot LED letters of Mary Ellen Carroll’s Indestructable Language installation in Jersey City.]

And speaking of gigundo pixels…

  • Duncan Wilson has created pixelnotes, a wallpaper constructed from Post-It-style chunks of paper, four deep, each of which is a different color.  "Pixelated formations and shapes develop according to our patterns of use."  Groovy.  See also his funky Cup Communicators (heh), as well as his PSP cocoons.  (And I’m envious that he gets to work with someone named "Sirkka Hammer."  That’s the kind of name that belongs on the Photoshop splash screen.) [Via Julie Baher]
  • Krazydad (Jim Bumgardner) turns vintage sci-fi covers into pixels, arranging them into a "coverpop."  Covers are arranged horizontally by time, and vertically by average hue.  Jim provides an interesting backstory (combining Flash, Processing, ImageMagick, and more) and visuals here. [Via Marc Pawliger]

Photographic sculptures, giant graffiti, & more

Photographic sculptures, giant graffiti, & more

Colossal images through Photoshop & Flash

  • Jean-François Rauzier has developed techniques for creating "Hyperphotos"–panoramas that can be printed some 30′ x 10′. "When looking at a Hyperphoto," says his press release, "at first you think you’re
    looking at an enlargement of a panoramic photograph.
    Not quite. Look more closely and you absorb a strange
    atmosphere that distances the viewer from the real world
    and sucks you into a universe of dizzying amplitude.
    Each Hyperphoto is a gigantic hyper-realist puzzle,
    created by assembling hundreds of close-up shots
    taken with a telephoto lens." 

    Jean-François reports that although he tried other software, Photoshop was the only tool capable of handling his 30-40GB images.  He displays them on his site using Flash, though for sheer scale I’d love to see one in person.   More info (in French) on his process is here and here.

  • Rob Galbraith has the story of HAL9000, an Italian team that has created a whopping 8.6 gigapixel stitched photograph of an Italian fresco.  They won’t go into the details of how they stitched 1,145 Nikon D2X frames into a 96,679 x 89,000 behemoth, but it looks like they use the excellent Zoomify technology to make the results visible (a la Google Maps) via Flash.  Check out the results on their site.

Hmm–using Photoshop and Flash together to make sharing high-res imagery a snap; seems like something the Grand Unified Adobe might want to consider… [pulling chin thoughtfully]

Colossal images through Photoshop & Flash

  • Jean-François Rauzier has developed techniques for creating "Hyperphotos"–panoramas that can be printed some 30′ x 10′. "When looking at a Hyperphoto," says his press release, "at first you think you’re
    looking at an enlargement of a panoramic photograph.
    Not quite. Look more closely and you absorb a strange
    atmosphere that distances the viewer from the real world
    and sucks you into a universe of dizzying amplitude.
    Each Hyperphoto is a gigantic hyper-realist puzzle,
    created by assembling hundreds of close-up shots
    taken with a telephoto lens." 

    Jean-François reports that although he tried other software, Photoshop was the only tool capable of handling his 30-40GB images.  He displays them on his site using Flash, though for sheer scale I’d love to see one in person.   More info (in French) on his process is here and here.

  • Rob Galbraith has the story of HAL9000, an Italian team that has created a whopping 8.6 gigapixel stitched photograph of an Italian fresco.  They won’t go into the details of how they stitched 1,145 Nikon D2X frames into a 96,679 x 89,000 behemoth, but it looks like they use the excellent Zoomify technology to make the results visible (a la Google Maps) via Flash.  Check out the results on their site.

Hmm–using Photoshop and Flash together to make sharing high-res imagery a snap; seems like something the Grand Unified Adobe might want to consider… [pulling chin thoughtfully]

Pixels a Go-Go

Think you’ve got a lot of data after starting to shoot raw? The new RED ONE camera system eats bandwidth for breakfast & then asks for thirds. The recently announced system is said to generate an insane 60fps of 4520×2540 (11.4 megapixel), 12-bit raw data (see specs). Put that in your hard drive & smoke it (literally).
Elsewhere in Tales of the Ginormous, a Microsoft researcher is stitching together hundreds of photos to make a 4-gigapixel monster (with 2.5x bigger on the horizon). And John Carmack of Doom/Quake fame is using a 5GB “MegaTexture” to power his next title. [Via]
These guys ought to hire The Humungous to be their pitch man…

15,000-layer Photoshop file

‘Tis the season of gigantic PSDs: Digital painter Bert Monroy sees Kevin Hulsey’s work and raises him a couple of gigs. Bert, a former matte painter at ILM and elsewhere, has been pushing Photoshop since v1.0, and at Photoshop World he unveiled his latest creation: a monster painting that’s 1.7GB (when flat!), comprised of some 15,000 layers, 500 alpha channels, and 250,000 paths. Man… what a testament to Bert’s artistry & commitment to his craft. [Via]
Photoshop handles tasks from creating sub-1KB Web graphics to wrangling files of basically unlimited size, and that makes it tricky for us to ship the app with settings that address all scenarios optimally. The Support team publishes some tips on optimizing performance (Mac/Win), and we’re looking at ways to make it easier to tune the app.
[Update: Tobias Hollerich points out that the site has been “dugg,” making it slow to load. The Digg.com entry lists some mirror sites & links to videos of Bert in action.]

720 hours in Illustrator; Painting with light

  • Think you’ve got patience and attention to detail? You’d need it to match the work of illustrator Kevin Hulsey. His site features a variety of tutorials that demonstrate techniques for building up artwork in Photoshop and Illustrator. Creating “Radiance of the Seas” took some 720 hours, and it’s fascinating to watch a fully rendered ship develop from line work in Illustrator into color in Photoshop, rendered in lavish high resolution. Kevin’s ghosting technique produces some lovely translucent results, too. [Via]
  • Playing with blending modes in Photoshop can enable some very cool painting with light (a la Picasso, not this dude). Computer Arts features a rather neat tutorial on the subject.
  • On a related note, this month’s Surface Magazine features a piece on Swedish collective Front, who use motion capture technology to sketch with light pens, then render the results as functional furniture. Check out images as they move from conception to finished pieces. Kind of reminds me of Moholy-Nagy’s Leda and the Swan. [Via]