A CONVERSATION WITH GREG DOWNING: TRAVEL PANORAMAS, 3D TECHNOLOGY AND HDRI
by Michelle Bienias
Downing also sorts out the confusion over HDRI in 'High Dynamic Range Imaging, Panoscan & Spheron'. and discusses 3D Imaging in 'CASE STUDY, TRIBUNAL PLAZA, NICE'
Greg Downing of www.gregdowning.com has traveled the world shooting location panoramic photography for Intel in Russia, Jordan, France, India, Vietnam and Australia. He has produced HDR images for Pixar and was an HDR imaging consultant for Spheron. Downing is also a current board member for the IQTVRA and is chair of the membership committee. He's a busy man - these are only a summary of his accomplishments over the past two years.
Downing's specialty is travel panoramas and the convergence of 3D technology and panoramic photography. He is constantly utilizing his considerable technical talents to produce new applications for VR panoramas (see the Panoramic Image Modeling section of his website and the VRMAG article on the Tribunal Plaza) and High Dynamic Range Images (HDRI).
In our extensive interview with this interesting photographer, Downing discusses his current and past projects, technology and his worldwide travels.
Note: The first panorama will apear and you can pan around but you will not be able to click on the next node until the entire movie is downloaded. This is only a problem because there is no visual feedback to let you know it is downloaded. If you wait for the movie to download (about 2.5 minutes over DSL) you will get an arrow icon for your curser when mousing over the cathedral and then the movie will work correctly. We are working on getting some visual feedback into this as soon as possible!
Can you tell us more about the "panorama based 3D virtual set" you produced for the multimedia park in Turin?
I realized a dream in this project I have had since the first time I saw Paul Debevec's Campanile movie (my first exposure to image based modeling). I had always wanted to use the omni-directional nature of panoramas to capture images of a location and use it to recover a complete 3D model of the location that had the realism of real photography.
In this project I photographed three panoramas in a public square in Nice. I then used photogrammetry techniques to create a 3D model from the photos and texture the 3D model from the photos. The result is a very fast method for obtaining an accurate model of an environment that is as real looking as photographs and has the interactivity of video games. I was recently told by a military contractor who works in government simulation training that my technique is two-thirds faster than what the U.S. military currently uses for its environment simulation and that the result is more realistic looking. Considering the amount of money and resources that they spend developing their techniques I found this pretty encouraging.
VRMMP will be compiling the model into a standard Virtual Reality SDK called Vega, from Multigen. This will allow them to use 1028x768 active stereo goggles, head mounted motion tracking and VR gloves for navigation. I am very excited to see my finished model in this environment. They'll be using the model primarily for Virtual Reality research.You have been traveling a lot for Intel image-based technology demos. Can you share some highlights of these trips?
In some ways this was a dream assignment for me and in others a nightmare.
The best thing about these assignments was that I got to photograph panoramas in some really incredible locations: the Taj Mahal in India; the unknown soldier and tomb of Ho Chi Min in Hanoi; the Opera House in Syndey; Red Square in Moscow; Petra in Jordan (my favorite!); and in Paris, Notre Dame Cathedral. These are the same places I would have been taking panos had I traveled there on my own, so it was nice to have a gig that paid me to go.
The assignment was not a "work made for hire" so I own the copyright to the images I took. This has been great, because I have been able to license my images to the American Museum of Natural History in New York. In September 2003, they will be opening a five-year traveling exhibit on Petra and featuring my panoramas and artifacts from Petra. At the entrance to the exhibit there will be a three-walled room that visitors can step into. The walls will be made up of three HDTV projections, allowing the visitors to experience a very high-resolution (panoramas are 16000 x 8000 pixels) immersive experience.
The most difficult part of the job (other than being rushed by soldiers with machine guns pointed at me in India, being sent to the hospital via an ambulance in Vietnam and getting arrested in Russia) was not having any lead-time on the assignments or advanced arrangement for the photography. A typical scenario would go something like this; I would get a call on Tuesday afternoon saying that funding for a project I had didn't even know was being considered, just got approved. If I wanted to accept the project I would have to be on a plane to Russia Thursday, on to Jordan Sunday, France on Tuesday and be back with completed panoramas for all three countries by Saturday.
I was stitching low-resolution versions of all the panoramas on the plane between countries and uploading them via FTP from hotels. By the time I left France, I'd get an email telling me which image Intel wanted and I'd start stitching the full resolution version of those panoramas on the flight back to the Untied States. Once I landed, I spent a day or two in postproduction on the images doing the Photoshop work and exposure compositing, if needed. Temis Nunez who played a design and directorial role on the project, created and integrated the animated bicycles in 3D Studio Max. Roger Jones of www.throbbing.com then tied all the media together with Lingo in Macromedia Director. I was working in a completely exhausted state for the entire trip. My preference definitely would have been to have more time in each country, and time to arrange permission to shoot before arriving on location.
My favorite location was Petra in Jordan. Unfortunately, I only had one long day in Petra and getting there wasn't easy. I was booked in a hotel in Amman, a three-hour drive across the desert from Petra. I took a horse, camel and donkey ride to transverse the terrain with my camera equipment and computer on my back and climbed more than 700 steps. So even though I left my hotel at 3 a.m., I only made it to the first location by 9 a.m.
I was lucky in that for the most part I had the entire park to myself. The local bedouin told me that Petra usually receives 6,000 visitors a day but since September 11th Americans and Europeans stopped coming for vacation, devastating the local economy. There were less than 200 visitors on the day I went and the park is so large that it takes three days to see the whole thing. Most of the people I saw were Bedouin and not tourists, which helped me get the shots I needed. One of the requirements of the assignment was that there be no people in the panoramas, since animated 3D characters were to be placed in them, a "frozen" person would break the illusion.
Petra was one of the most magical ruins I have ever seen in my life. The facades that were carved directly out of the sandstone cliffs were enormous. To give you a sense of the scale look at Al Dier (aka the Monastery) . It is 40.2 meters wide, the doorway itself is 8 meters high. The bottom of the door was at eye level and I am 6 feet tall (~1.8m). It was amazing to think that this was all hand-carved; inside the tombs you could see individual chisel marks going up the walls. Some of the facades were in perfect condition and some were rounded to soft contours by the elements and in almost all of them you could see the beautiful patterns of the natural strata cutting through the stone. The amazing part is that there are hundreds of these tombs in the valley and I saw less than a third of them. I can't wait until the day when I can go back and photograph the rest of Petra.
Even with all the difficulties of these trips, I loved the experience. It wouldn't qualify as an adventure if there weren't incredible obstacles to overcome. What were your most pleasurable experiences from your work as application designer and workflow analyst for Realviz, where you enhanced Sceneweaver, Image Modeler and Stitcher?This was my first experience working with highly skilled programmers. As a content developer I am well versed in working around problems and making applications do things that they were not necessarily built to do. It took me a while to get use to the idea that the program itself could be changed so quickly to solve a problem. I would make changes to the workflow or some other aspect of the application design and have a working copy, often within a couple of days. Often they could change the applications faster than I came up with a work around for a problem. That was pretty exciting for me.Your work was featured at Apple Expo Paris Keynote, why do you think your work appealed to them?This needs a little further explanation. While I was working at Realviz, I worked with Realviz's USA president, Emanuel Javel, and was involved in establishing a relationship with Apple. Apple was very interested in all of Realviz's applications and wanted to see them ported to OS X. They were especially excited to see Image Modeler come over to OS X and agreed to feature it as the finale in the Apple Keynote in Paris. They thought it would be a great fit to feature a cutting edge European company at the Paris Expo; Apple had just done all the work in getting AliasWavefront to port Maya to OS X and were pushing the Mac as a 3D platform.
We had jumped through all the hoops, preparing the material for people at different levels at Apple, and had been to numerous meetings with them. I was quite excited to have the opportunity of getting to meet Steve Jobs and show him what I could do with this application and then demonstrating it to the world, on Stage at the Expo, with Jobs and Emanuel. Our first meeting with Steve Jobs was scheduled for September 12, 2001.
When the tragedy of September 11th happened, like everyone else I was shocked but I had to bury my emotional reaction for another day and continued to work on a demo. It was a strange day in the San Francisco Realviz office, working under such an important deadline with the news on an office television repeating the airplane crash into the buildings in what felt like an endless loop.
We had three demos for the show; one was a modeled and animated woman's head, a unique-looking complex French mailbox that could be modeled in three minutes, and a highly realistic model of a San Francisco building that suddenly and very realistically explodes, done by Temis Nunez at Realviz SF. On September 10th, it was by far our most impressive demo of the product because it was so realistic and so dramatic, it looked like something done in the movies and was made by one guy on a very short timeline. On September 11th, we found out just how realistic it was and suddenly it was the worst possible demonstration, so we had to throw it out of the presentation. Our September 12th meeting got pushed back; Steve Jobs had a personal friend that was a passenger on one of the planes and he was in mourning. That was okay by me and the rest of the team as we all wanted a little space to consider what was happening. A few weeks later, Apple pulled the plug on the Apple Expo Europe because no one from the American companies was willing to fly.
I feel fortunate that the events of September 11th only touched my life professionally and did not take the life of anyone I knew personally; even so, the professional loss was quite a disappointment.You created an impressive virtual tour of Notre Dame Cathedral for GorillaGuide.com, combining QTVR and image based modeling. The Notre Dame project was a pretty big project at the time. I had done one proof of concept before this using the same technique with the San Francisco Museum of Modern Art. I landed the GorillaGuide gig while traveling, taking panoramas on my vacation in Europe. I had saved up three years worth of vacation and decided that if no one was going to pay me to do what I really wanted to do with my life then I would just go and do it for myself. To my delight, as soon as I started doing just that, I found someone there willing to pay me.
I was staying at a youth hostile in Rome when a video crew made up of American 20 year olds came through, shooting video of all the youth hostiles in Rome for a travel website targeting American college students. I showed them some panoramas of Rome and the SF MOMA project on my iBook and they got very excited, called their boss on the spot, handed me the phone and I suddenly realized I was being interviewed. I was hired without ever having met the people who hired me. A few weeks later they sent me on to Paris to plan a similar project of Notre Dame. Unfortunately, I had my homemade multi-row bracket stolen while I was staying in Florence so I could only shoot cylinders.
I was in Paris for 10 days, during which I think I saw the sun for about 20 minutes. This was fine, as I wanted to make sure that the textures were within the dynamic range of my camera and I didn't want conflicting shadows if I had to shoot at different times of day. It was difficult to get a complete panorama with no tourists between the two bell towers, so I had to wait until the last tour of the day and make sure I was the last person allowed through. The shots taken from the ground had to be taken at the first light of morning before all the tourists showed up.Back in 1999, you realized an educational exhibit for Sonoma State University where you used very particular techniques for this job, could you tell us about them?The Sonoma State Exhibit was a photo exhibit that was the result of a long collaboration with Charles Evans (www.digitalpanos.com). We did a photo installation at Sonoma State's Fairfield Osbourne Preserve and sold the prints as a fundraiser. For two years, Charles and I would go out shooting panos at the preserve just about every week. By the end we had hundreds of panoramas and some of them were pretty good. The preserve is still using one of the images on their website.
I built the project when I got back to the States. My new boss was thrilled with the results and I started on a second, even more ambitious project of the Louvre. I was quite excited about this because the Louvre was interested in licensing the project as well. Unfortunately, the GorillaGuide went out of business before the second project was finished.
I wanted to make the exhibit as immersive as possible and have always had an interest in obscure photographic techniques. Along with the panoramas that were displayed flat along the walls, I had one printed on canvas and spray mounted this onto both sides of ABS plastic and then riveted this into a cylinder that we hung from the ceiling via monofilament. The inside print was a mirror image of the outside print so that when you approached the ring you would see a matching view on the inside of the cylinder.
I also took a series of photos with a three lens stereo camera that allowed me to make lenticular 3D prints for the show. In addition to that, I created one laser transmission hologram for the show of the different plants and lichen that could be found at the preserve. Steven Anderson at Sonoma State University was kind enough to teach me how to make holograms. I found this a fascinating, if under-explored, medium.Please tell us about your most important publications over the years.
So far I have found myself having to write quite often. I wrote the first photography for a photogrammetry manual with Charles Evans that was published with an early version of Image Modeler. This was a good process for me as I had to take the gestalt of my photography that I gained by a trial and error process and translate it into a literal set of rules. My photography for photogrammetry improved after writing it.
My first article for a 3D magazine was a good lesson; it showed me how easy it was to make a useful contribution. The article was on how to make linear movies from panoramas. In the film industry they call this a postproduction pan. It makes compositing special effects easier, helps control camera artifacts (shutter speed and speed of camera movement is more flexible with a post production pan), and it offers the director more control over how a camera movement is made.
One other memorable moment for me was when Paul Debevec (http://www.debevec.org/) used my stitching HDRI tutorial in his Image Based Lighting course at Siggraph 2002. His work has been a big inspiration for me my whole career, so it felt pretty good to both give back and get some recognition.When did you start doing VRs? Do you still have your first VR?I think it was around 1997 and yes, I still have it. It is pretty sad, lots of banding and when I printed it out it only looked good when it was printed 2 cm tall. I remember I made a wristband of it and carried it around with me for weeks, showing it to people and saying "isn't this neat!" It took another six or eight months of practice and a better camera before people started agreeing with me.
What can you tell us about the recent changes in the IQTVRA board, and your future involvement?
I was happy to be elected and also a bit surprised that I had that much name recognition. I have served on non-profit boards before, so am familiar with the struggles of making so many decisions with a group of people, but also of the sense of accomplishment when you achieve the goals you set. We are lucky to still have some of the old board members; they offer us a sense of continuity and having some new blood has energized things. I am glad that the board still has an international flavor with some representation from Europe and Australia.
While I was living in France I went to an IQTVRA meeting in Köln Germany and really enjoyed meeting so the thirty or so European members that attended. I think it?s important that we continue to develop our membership and leadership worldwide; it makes for a much richer community.How do you see IQTVRA's role in our industry and community evolving?
I see what we are calling interactive imaging expanding, both as an art form and an industry. There are half a dozen image-based technologies that will find their way into our work as photographers in the coming years and I want to make sure our organization can grow with it and fan the flames of its growth. There are many things that we can do as an organization that would be too difficult for us to do as individual photographers.
As our technology, community and industry are maturing, I would like to see the IQTVRA do a better job of serving the specific interests of our community. I think that there is a lot more that we can do for professionals. I want our organization to behave more like a trade organization and do more marketing of VR, and I think that we can do this while still maintaining inroads for the hobbyists and amateurs, we all started out that way. I think there is more that could be done for educators as well.Some people, including our director, Marco Trezzini, believe 2003 will be THE year of VR, when it will finally break out of the boundaries of our community and industry. What are your thoughts about this?
I think he could be right. Despite the tech crash, there are several observable positive trends: we have grown as a community, our technology has matured, and even our industry is expanding. If you look at the traffic on the QTVR lists, it has probably tripled in the last couple of years. Now there are many different solutions for capturing panoramas and about a dozen methods to deliver them. In the last year or two, we have also seen the development of several regular publications, such as VRMAG, that are serving our community specifically. I think that tells us that our industry is maturing.
Although I think it is probably still pretty difficult for most people to make a living purely on VR, we are finally outgrowing much of the damage done by IPIX. We are also seeing more new markets for our imagery open up (i.e. 3D, selling high quality prints, and security).Any comments, critical or otherwise, about VRMAG? Anything we should be doing but aren't? And how you see its role within our community/industry.
I think VRMAG is really great! It is exciting to learn about my peers and what they are doing, how they got started and where they are going. This is really a first for our community. There are so many people in our community that I know some from the postings on the listservs but it is so much more revealing to see the spotlight or guest artist on them in a VRMAG article.
More of Greg Downing's work:
Downing’s Garfield Trailer
Petra Exhibit at the American Museum of Natural History in New York, featuring a Surround movie generated by Downing’s panoramas