Jeffrey Martin creates gigapixel panoramic shots by stitching together thousands of photographs, mainly urban landscapes, Peter Dench finds out more
If your pictures aren’t good enough, you’re not close enough,’ is the famously over-quoted mantra by Hungarian-American war photographer and photojournalist Robert Capa. It has since been adopted by many photographers as motivation to keep going forward, over borders, front lines and into living spaces. To get as intimate as possible with what they are photographing while remaining an objective observer.
Jeffrey Martin’s photographs get in close, then closer, then even closer – enough to placate Capa’s doctrine. They are also shot from very, very far away and are very, very big. An entrepreneur and designer of camera systems, Martin is also a Gigapixel guru creating some of the largest panoramic photos in the world. Gigapixel photography is the craft of shooting hundreds or thousands of photos and joining them together into a single, seamless, ultra-high resolution image (a Gigapixel image is a digital image bitmap composed of one billion pixels 1000 times the information captured by a one-million-pixel digital camera). These colossal images are typically shot with a complex set-up involving long lenses and a programmable robot, typically requiring weeks or months of post production.
‘I’ve always been interested in photography since I was a kid. My dad was a keen amateur photographer and my mother’s dad was a keen amateur photographer. I have his slide collection in a couple of large boxes and his Leica rangefinder from a long time ago. My first photography course was in 1996, learning black & white photography in High Desert, New Mexico. Amazing, the original type of photography, shooting film, developing it yourself, printing it yourself.
When I got my first digital camera around 2000, a Canon digital ELPH, it had its own panorama mode and primitive stitching software where you could shoot overlapping photos. This struck me very deeply as something fundamentally different than analogue photography, to fuse photographs together, that’s pretty wild! That was a big thing for me. Changing from film to digital.’
Martin started playing around with early photo-stitching programs developing his 360° photography technique. He Googled, reached out to website forums, Yahoo groups and slowly learned from those more experienced in the field – which lenses to buy and how to use a panoramic tripod to shoot the perfect spherical photo. ‘Pretty quickly after that I was doing some fashion photography for work and had a 70-200mm lens which is great for portraits and also for shooting landscapes.
If you shoot 20 to 30, 70mm pictures in a row overlapping, you’ve got an extraordinary high-resolution landscape or skyline. The first Gigapixel type of image I tried was the Prague skyline; the beer garden near where I live has a great view. I zoomed in and joined them all together and from, say, a six-megapixel camera got a 100, 200-megapixel image which in 2002 was outrageous. How far can this go, how far can you take it? Since then I’ve been trying to find out.’
A decade later Martin did find out. Shot over three days from the 29th floor of London’s iconic 177-metre-tall BT Tower, working alongside colleagues Holger Schulze and Tom Mills, they set a then-world record for the largest panoramic digital photo, a staggering 320 gigapixels. ‘We were asked to create the largest photo in the world. I’ve thought for a few years the BT Tower would be the perfect location for that, it’s like a giant tripod, designed to stay still,’ says Martin. The plan was military.
Four Canon EOS 7D cameras each fitted with a 400mm lens and 2x converters attached for even more zoom were mounted on a Clauss Rodeon VR Head ST motorised mount, 3D stereoscopic laser robots moved them around incredibly accurately to within five-thousandths of a degree. The constantly moving cameras had to focus, shoot and fire around 13,000 times successfully. Each photo would be around 18 megapixels.
‘The full set of images took only 90 minutes which is extremely fast but it took six attempts and several days. One day the wind was 50 miles an hour and my partners didn’t want to shoot at all, we had an argument, someone almost quit, it was not smooth. Generally, it’s good to shoot more than one set because you’re building a large mosaic of images and you want all of the images to be high quality. You don’t want any to miss the focus or have motion blur or miss completely.
You’ve got a programmable robot and it’s moving the camera and firing. London has 48,640 images, it’s not too outrageous to imagine it might miss a couple. You know how many pictures it’s supposed to take. You check the file number when it’s done and pray it’s the right number. Also you can listen – it’s clicking regularly and if you’re paying attention it’s easy to hear if it’s missed, but then if there’s four cameras running at the same time and only three photographers as it was in London, that’s not going to happen!’ explains Martin.
The project used nearly 3TB of data shooting JPEGs. Stitching the images together took close to 200 hours of render time on a 32-core computer with 256Gb of RAM. Each photograph taken on the day represents 1° of the finished panorama; there are 85 rows altogether.
Before you open and interact with the London 320 Gigapixel Panorama Photo, clear the diary. Time is gobbled away as you forensically swoop into the minutiae of the metropolis: football players warm up in Regent’s Park, Big Ben chimes 1.30, visitors pack onto the London Eye, a boy scoots along New Cavendish Street, I think I spot my wife!
You can recognise someone’s face, read street names and number plates at around half a mile away. Between six and 12 miles the detail holds before the air quality intervenes. Shapes are recognisable at 20 miles. It’s impossible to tell the photo was made from thousands of smaller photos.
Martin has gone on to beat his own record with the 2018, 405 Gigapixel of his hometown Prague, shot from the Astronomical Clock Tower on Old Town Square. The world record is currently held by the 2015, Kuala Lumpur 846,071,539,488 Gigapixel – the final image combining over 31,000 individual images in a panorama. Martin has also photographed New York from the Empire State Building, Rome from the Torre delle Milizie and Paris from the top of the Eiffel Tower. He would like to photograph Mumbai and Rio de Janeiro.
The release of Auckland, Chicago and Barcelona Gigapixels is imminent. His favourite Gigapixel photographs are the ones he shot in Japan from the Tokyo Tower and Roppongi Hills Mori Tower. ‘Tokyo is such a dense, endless, amazing city with all these weird hidden places. In the middle of skyscrapers you’ve got these little houses and parks, interesting glimpses into a far-away land that people may not have the chance to go to.’ Martin challenges the viewer to find the baseball players, temple worshippers, woman hanging her laundry and the salaryman who went to sleep on a bench, fell on the ground and kept on sleeping.
With Martin’s up-close and personal probing of people in public and private spaces, has he experienced any protests or backlash? ‘These images are made infrequently and I don’t exactly reveal when they are shot as an additional layer of privacy. If someone raises the issue, I kindly point out there are more severe, threatening issues of privacy around CCTV cameras and so on that are literally watching you all the time and the ability of law enforcement to link your phone metadata with your bank metadata with your transit card metadata – this is far, far more worrying to me in general than a single Gigapixel photo which is meant as a work of art.
There have been street photographers for decades, I would see this as an extension of that. In the case of a city, there is a law called the freedom of panorama. In general a provision for photographing a cityscape that they cannot possibly get permission from every owner of a building. There are general laws in most countries where people who are in public in view of a public place don’t have an expectation of complete privacy.
‘That being said, in a couple of Gigapixels there have been things that I would not want myself on the internet like that. We practise our best judgement and blur out what might be embarrassing. There has been at least one naked person and in one Gigapixel shot by a friend, they found a gun on a roof just sitting there; they informed the police who went and took the gun away,’ reveals Martin.
Robots, multiple cameras, lenses, permissions and teams of photographers… it all sounds a bit overwhelming to start experimenting in Gigapixel photography. ‘You can just go to your favourite vantage point or tower in the middle of a town square and take your 50mm lens, try shooting four pictures in a row that overlap about 25% and boom, you’ve got a picture that is about three times the resolution of your camera. You don’t have to take it to the absolute extreme like I do, but it’s a useful trick.
‘For example, shoot a portrait of a couple of people and have some extra photos on each side, a longer photo of higher resolution. There’s some free software that works totally fine called Hugin – it’s open source, the user interface is not the easiest but the quality is great. It can be used in a fairly automatic manner. If you shoot these four photos, put those into Hugin and click start, then it should give you the stitched image error-free.
‘If it’s something you enjoy, try taking more photos with a longer lens. There’s no functional limit in terms of number of photos for most people, you can make 100 or 1,000 photos and thread them together. Of course it gets more necessary to understand some of the technical issues behind how to do it if you want the results to be better,’ advises Martin. If your photographs aren’t good enough, perhaps you’re not far enough away.
Jeffrey Martin is an entrepreneur and designer of camera systems, being a gigapixel photographer in his spare time. He has shot record-breaking large-scale photographs, with more information available at jeffrey-martin.com
Follow AP on Facebook, Twitter, Instagram, YouTube and TikTok.