Back in July, Bernie and I discussed the possibilities of putting together a large-scale effort to photograph a number or world-class momuments. This would be done with immersive images, of course, and Bernie had already started talks in this regard. He has a meeting in a few days to take this futher, so he asked for some advice, which I've just sent him. It's long, so you may not want to read it all, but this is where my mind is these days.

----------

In launching a campaign to create highly detailed immersive interactive panoramas of world monuments, I would want both flexibility and the highest quality possible. Any system used to create the images should not limit itself to a particular technology, and should be able to handle future developments in the field gracefully.

To my knowledge, a Panoscan Mark III is the best digital capturing system presently available to the public and should receive top consideration for a recording system. It offers outstanding resolution, the ability to capture high dynamic range information (HDRI) in a single pass, and works quickly. The drawbacks to digital scanning cameras such as Panoscan are their cost, some problems with people or other moving objects in the scene, direct sunlight, and less portability.

With regard to cost, the price of a Panoscan system can easily be offset by the reduced time needed in post production. Problems with movement can be avoided if shooting a still scene, or scanning multiple times to produce a composited image free of movement. Direct sunlight problems can be avoided by careful placement of the camera, shading, or multiple scans. Digital scanning cameras must currently be connected to a computer and power source in order to operate, and must rely on a tripod. This makes then less portable, which may or may not be a problem for a given site.

Competing against digital scanning cameras are digital still cameras (with stitching software packages), which have recently become better in quality. Cost-wise, such systems weigh in at a small fraction of the price tag of a Panoscan, however their photographs require significantly more handling, expertise, and time in post production in order to create a comparable immersive image.

Still cameras have no problem with in-scene movement, although objects in their scenes may have to be edited out if moving from one frame to another or not removed by automatic blending processes when on the edge of a frame. Direct sunlight poses no problem for still cameras, and they are quite portable, and can even be used without tripods (although a lack of tripod requires a very high level of expertise to blend stills for panorama creation). Still cameras on a tripod can also capture HDRI, but only with frame bracketing, which is much more labor-intensive. They are also best used with high-quality mounting hardware.

Both of the above approaches assume a digital approach is best – each produces generic types of files that can be applied to any future technology capable of producing immersive images from high-quality digital files. Traditional film can also be used to capture the original images and then scanned to produce digital files. Scanning cameras exist for film, as do still cameras of course. At this point it is probably debatable whether film has been superceded by digital chips, and may very well depend on budget, formats, chip sizes, etc. Lens quality will also play a significant role in image sharpness.

Whether to employ film or digital capturing technology appears to be a significant question. Each has numerous pros and cons, which could be listed in a large matrix, but any final answer would depend highly on the resources behind the project and its ultimate goals, and perhaps even on the preferences of individual photographers. Resources don’t simply include funding, but also locations to store negatives or digital files after shooting, routines for archiving and retrieving photographs, and methodologies for processing in the field.

Given adequate funding for a large project, my preference would be the use of a Panoscan Mark III digital scanning camera with an identical system available as a backup, and a high-quality still camera fail-safe system for tertiary back-up. It seems to me the hardest parts of any systematic moment visitation program will be arranging permissions and dealing with too many people in the scene. Thus, after all arrangements have been made, the worst thing to happen would be equipment failure. Redundant equipment would prevent an embarrassing waste of effort. Now, should funding not allow a backup scanning camera, a still one would have to do. Without funding for even one scanning camera, a pair of high-quality still cameras would have to suffice.

File formats in either case would be camera RAW files, so that no reliance is placed on the camera's software for image coloring. Instead, any image program can be used to create the colored images from the RAW files, and if necessary changes can easily be made in the future. RAW files can be batch-processed into JPG or TIFF images at very high resolution for final output.

Exciting new developments with Flash, specifically Spi-V, may prove to oust QuickTime as the preferred method of immersive imaging, at least for the moment, although neither is globally accessible. Spi-V offers many new capabilities such as allowing dynamic changes to an image on-screen as the user pans or tilts between areas of low and high light (in which case the image lighting changes to lighten or darken accordingly), or the user switches attention between foreground and background components (in which case the image blurs itself oppositely). Such changes are produced through a masking file, but require twice as much storage space compared to non-dynamic images.

A project intending to visit world monuments should concentrate on photographing them well, using a generic file format that can be post-produced to the current best technology available, where "best" can mean most available, not necessarily most technically advanced. QuickTime and Flash both have many things to recommend them. Java can also be used, but is not currently the favored medium by those currently producing high quality immersive images.

Sites should be covered both inside and out. Whether they should be shot only from a person's perspective depends on the project's goals. If a site is being documented for posterity, then perhaps cranes or other machines should be used to shoot from vantages not normally possible. If the desire is to show things as they are from the ground, there would be no need for unusual equipment. Obviously the more complicated the photography, the higher the cost of the project. Whether or not to use any artificial lighting also depends on the project. If certain site elements must be brought out to their best, then strategic lighting may be required. Any lighting powered by cables or normally behind the camera in a traditional setting would be visible in a panorama, and subject to digital editing or more cumbersome arrangements during the shoot.

It seems to me that it's not always necessary to shoot every site equivalently, if that’s even possible. All sites should be shot representatively at least, thoroughly if possible. But again, it depends on the project's goals. Documenting a site should be thorough. Producing something for comparing and contrasting sites should concentrate on those different aspects and coverage need not be exhaustive. Actual coverage is also subject to many forces. If time permits only one day on a site, lighting or weather conditions may not prove agreeable, or even possible. Scheduling can make a big difference -- perhaps none is larger -- in how a site looks. This ranges from natural lighting to background vegetation to cloud cover to crowd control.

Speaking of people, decisions will have to be made whether they appear in the photos or not. There might be a preference for empty spaces, but that can make sites look dead. On the other hand, people in a scene can block the view. In-scene people can serve as living scales, but could pose legal problems if they are recognizable. They can also date photos by hair and clothing styles, which may or may not be desirable. Dating problems can also crop up from vehicles in the background, or the appearance of modern contraptions such as wiring, exit signs, fire extinguishers and the like. Whether or not these should be edited out depends on project goals and resources.

Camera positioning for most sites should be considered so that the advantages of immersive imaging bring out the site's best. Shooting from the middle of an interior produces a good view from all around, but placing the camera nearer one wall makes the experience more interesting in many cases. Shooting from both vantages is even better so that the user can choose which view to experience. Being able to jump from one vantage to another within a series of immersive photos often requires line of sight from one camera placement to the next, but this feature is not necessarily desired, and need not be implemented even if the technology allows it. Navigation from one vantage to another might be best done via an accompanying plan of the site with selectable viewing options.

For any project, I prefer to photograph it from as many vantages as time permits. Whether shooting with a digital scanning camera or a still camera, time-on-site is the most limiting resource there is. Take as many photos as possible, then decide which to use in the end product later. In the case of still images, there’s often no need to spend the time stitching after reviewing the sequences of images to decide which would make the best immersive offerings.

Prior to shooting a site, I would think it best for the photographer to scout it for best camera locations. In addition, if an expert familiar with the site (or type of site) can be consulted about suggested perspectives unique to it, then that would be helpful. Thus, I would expect each site would be visited not just for a few hours of photography but might be considered for a couple of days to find the optimal conditions for shooting. But again, if the goal of the project is simply to make a few quick immersive images, this might not be necessary. If time allows, the photographer can work on post-production.

That being said, it's a good practice to record the locations of each shot for later reference. Digital cameras automatically record the standard photographic references, but do not yet have GPS built into them. A portable GPS unit for the photographer would also be handy. Actually, at least one assistant would be even better for this purpose. An assistant would also be very helpful in transporting, setting up, and operating a scanning camera, or in field-testing still camera image stitching.

If you have more details about the goals and resources of the project I could offer additional specifics. I hope your meeting goes well, and I remain most interested in participating in a major role.