I had a chance to do another large banner project for the YMCA. This time it’s for a 13×17ft rear-lit sign to hang on the corner of their newly remodeled building in downtown Dallas. The last banner job was going to be viewed from close enough that I decided to shoot with the 24.5 megapixel Nikon D3X. This new banner would be viewed from the street so I shot it with my 12 megapixel D300.

We shot several setups, knowing that one of the shots would be used on the building and the others would be part of a billboard and bus campaign. This photo of a group of Y members was the one chosen for the big sign. There were several changes and repairs to be made.

Quite often on discussion groups you’ll hear people get snarky and say something like “I would have just shot it right in the first place instead of spending so much time in Photoshop”. The people who take that attitude may not have shot in a live, slightly chaotic location with a dozen people waiting to have their picture taken. With paid models you can sometimes take a little longer getting things “right”, but you don’t often have that luxury. Also, what’s “right” may not have even been decided yet!

We shot this group of people near a railing, in front of a glass wall with the basketball courts in the background.

Once the shot was selected it was time to clean it up a bit: (from L to R)
- Stretch the top part of the frame to match the aspect ratio of the sign.
- Repair the woman’s eye which was hidden under her bangs. I found another eye from a different shot.
- We didn’t have a good shot of the second man looking at the camera so decided to replace him with a woman from another setup.
- Replace the logo on the man’s shirt with a Y logo and change the shirt color.
- Remove logo from the red jacket.
- Remove logo from shorts.

The idea was to have some good background action happening on the basketball court. I had the actors step out of the frame and I shot a couple dozen frames of court action at 1/8 sec to get some blur. Keeping the camera locked off made it a little easier to composite the background elements:

- The woman with the blue shorts was used in place of the man from the original image. I had to first clear out a space for her by cloning over the edges of the man to give her some extra background. Then the adjacent actors were masked and she was placed behind them. Shadows were painted in to match the lighting from the left. The client also asked to add some color to her shirt.
- The basketball court was placed into the main shot, along with background action from three shots.
- The last inset shows a photographer friend of mine with the Lastolite Easy Balance card for white balance. (Wiley is his name and you can see his work here.)
- My client sent over the background graphic as a Illustrator file and I dropped it in as a Smart Object and needed to do some masking around the actor’s head
- Then there was skin retouching, some clothing repairs, local and overall color balance, hair trimming, sharpening. . . the usual ;-)
- The file was sized up to final print size at 50dpi, which is the native resolution of the XL Jet printer.

Here is the final composited shot after about 3 1/2 hours of work:

My son plays tuba in a rock band that’s been practicing at the studio for the last few weeks. I thought it might be fun to shoot some pictures at one of their rehearsals. I didn’t want to spend a lot of time setting up lights and wondered what I could do to get something interesting. They were playing in front of the cyc, but the white floor is in need of a coat of paint and I didn’t want to spend time cleaning that up in Photoshop.

A raw shot from the balcony with just the overhead fluorescents on:

I turned on a 2K (2,000 watt) tungsten light, widened the beam and just blew out the cyc wall behind them. That gave me some interesting contrast and solved the dirty floor problem since now the light is skipping off the surface of the floor instead of showing all the dirt. I did a little toning, tinting and negative clarity in Lightroom and go something we all liked.

Like most people, I love fireworks. And like all of you, I love taking pictures. So at least once a year I find myself with camera in hand, trying to capture some interesting photos of explosions in the sky. Last Saturday night my wife and I met some other couples near Lakewood County Club. I’ve been attending this little fireworks show for over a decade. There’s a street that bisects the golf course and we stake out a space early and then sit on our blanket while the mortars go off right over our heads.

I like shooting handheld, with the camera in my lap, so I can still watch the display and enjoy my surroundings. I started out by putting the lens on manual focus and setting it on infinity so I didn’t have the problem of the camera seeking an object in a dark night sky. I played with exposure until I settled on about 1sec at f8 at 200-400 ISO. I jiggled and spun the camera during the explosions, which gave me some interesting patterns.Where it got really neat was purposefully defocusing, holding the zoom barrel in my left hand and twisting the camera body on axis, zooming in during the explosions. The result is the almost aquatic looking images toward the end (my favorites). Even defocused, at the widest end of the zoom the light trails are relatively sharp. As I zoomed in they softened. The only processing I’ve done to these images is changes in exposure, fill light and black level in Light Room to bring out some of the trails hidden in the darkness – a very good example of why shooting RAW can really help.

In my bio I mention that I got involved in the motion picture special effects area of motion control for a while. What I built was a large Cartesian (XYZ) robot that could fly a camera around a 9×14 foot area. When I moved into the Spot last year I decided not to bring the  system with me. I hadn’t shot with it for a while and it was time to move on from that phase of geek insanity. So I’m selling it.

If you know anyone who is interested, or you just want to take a look, visit the gantryrobotforsale.com website I’ve put up. There are lots of photos, specs, drawings and some videos. This rig would be perfect for someone doing industrial plasma cutting or welding, or an artist doing intricate work in metal, wood, or anything else you can cut through.

UPDATE: The robot was sold to a special effects house in Chicago.

I was invited to speak to the Dallas Camera Club last night. I was really amazed to see almost 60 people attending the meeting. The DCC has been around for 75 years and has a very active calendar of meetings, competitions and field trips.They are a great group.

I spent about an hour talking about panoramic images, discussing why you’d want to shoot them, how to make them and then looking at the amazing software available to automatically stitch several images together into a seamless photo. I’ve been using AutoPano Pro for a few years and just love it. There are dozens of programs out there, many of them free or inexpensive. Panorama Tools has been around for a long time and has some great front ends like PT GUI. If you have Photoshop CS4, you already own a great stitching program, but I still far prefer AutoPano for its flexibility and ability to detect and extract panos from a folder of images.

I think panoramas are one of the coolest and most satisfying techniques to come out of digital photography. The images are wide format, more closely matching our natural field of view. You can create big impressive prints with even a low resolution pocket camera since the resolution of each photo is added together into the stitched final image. They look beautiful hanging on a wall and are easy to print at home. Watching the software assemble the images is really amazing.

At the meeting I showed some panoramic sequences from recent travels. I also did a live demo, shooting hand held with a little Pentax pocket camera and dragging the images into the AutoPano software.

These are the raw shots:

This is the result of the panorama stitch, processed using the “Spherical Projection” setting. This has the best relationship of sizes of the people in the room, but the most pronounced curvature of tables, which were actually in straight, parallel rows. This is what the shot would look like if taken with a scanning film camera like the Widelux or Horizon:

The same panorama, using the “Planar Projection” setting. The tables look better, but the people at the edges are stretched, as they would be using a traditional super wide angle lens:

This is the same Planar projection as above, but taken into Photoshop for some warping. I did a Select All, used Transform:Warp, and pulled in the sides. There is less of a stretched feel, but you start to loose the wide format of the panorama:

I highly recommend shooting some panoramas. With AutoPano, and others, you can shoot multi-row panos, and even shoot any group of images covering a subject, without them having to be aligned, level – or even all horizontal or all vertical, as long as there is overlap in the images the software will find the pano in there.

Well, here’s a product I didn’t know existed but was happy to find. About a year ago I built a new machine for the studio, mostly to run Lightroom. I like Lightroom and want to love it, but have had speed problems with it since day one. So I built a quad core machine with 8gb ram, fast ATI card (had conflicts with the Nvidia) and a serious RAID 6 controller from Areca for storage and cache. For my viewing pleasure I got two 27″ displays from Doublesight and glued it all together with Vista 64. The problems began when I had to install a couple of DVI video extension cables to the monitors. Every time the screen blanker kicked in, or when the machine was restarted or came out of sleep mode, the monitors wouldn’t come up at the right resolution or even left to right sequence. About half the time I had to go in and reset the layout and resolution. It was one of those really frustrating things that I let go for two long.

While I was trying to get some HDMI problems solved (another day, another rant!) I asked John Johns, the video wizard at In-Sync, Inc. in Dallas, what he would recommend. He referred me to Gefen, a company that makes a crazy array of video conversion boxes. On their website I found a little box called the DVI Detective and an explanation of what the problem was with my monitors (and my HDMI troubles). Both of those connections transmit an EDID code. According to Wikipedia: “Extended display identification data (EDID) is a data structure provided by a computer display to describe its capabilities to a graphics card. It is what enables a modern personal computer to know what kind of monitor is connected.” As with many things electronic, that doesn’t always work out in practice. What was happening was the EDID wasn’t being picked up and weird things were happening.

The DVI Detective goes inline between the computer and the monitor. You power it up with the included wallwart, push a tiny button on the little box, a light flashes telling you the code is being received and stored, you flip a switch to lock the setting – and you’re done. The power can now be unplugged. I was hesitant to spend 2 x $60 (from Buy.com) on these little gizmos but my problem is completely solved. I love these things.

They are tiny boxes and come with nice, stout turn around DVI patch cables. A solid product that does exactly what it claims to with very little setup. This is what the mess looks like behind my monitors until I dress this up a little:

UPDATE: 7-4-09
I’ve been using the Gefen DVI Detective boxes for a few weeks now and I can happily report that they have completely solved the problems I was having! Money well spent.

I have a collection of toy and cheap cameras that I started about 30 years ago. My rule for the first couple of years was that I would’t spend more than a dollar. Sometimes I would find 5 or 6 at a thrift store. Then my family started looking for them. I still get a bag of garage sale cameras every Christmas from my brother.  I now have probably 500-600 cameras stashed in cardboard boxes on some industrial shelves at the studio.

For a while there has been a tiny leak in one of the concrete studio walls and I was dutifully collecting the water in a bucket. When I went to the studio yesterday during a downpour I discovered water dripping from a new place, right under the shelf that holds my collection. Turns out a trickle of water had been going into one of the boxes for the last few months. Unwrapping the cameras was gross and sad. I had to throw away about a half dozen, including a nice small wood view camera that had fallen apart. The biggest shock was this Argus C3, which was in its original box. The box was a black, sodden, smelly mess. I’m guessing that the combination of water, darkness and the leather case made a perfect breeding ground for mold and mildew.

A few more photos:

I’ve had an Epson 7600 for a few years and always thought of it as a big machine with an appetite for ink and a producer of impressive big prints. Up to 24″ wide by any length. Boy, have I had to recalibrate that thinking! I just got through shooting a job for some interior banners that will hang at the newly remodeled downtown Dallas YMCA. They are being printed on an HP Scitex XL Jet which prints up to 5 meters wide (196″). The images will be full bleed and the size is 9ft x 18ft. I shot using a rented Nikon D3x. I’ve written about that experience here. The resolution of the D3x is roughly 4000×6000 pixels. The long dimension of the banners of 18ft, or 216 inches. Dividing the resolution (6000 pixels) by the printed dimension (216 inches) will give us the maximum pixels per inch that I would get on the print: 6000/216=28. If I crop the image I would have fewer pixels. Is 28 ppi (pixels per inch) enough? Absolutely, because of viewing distance.

When I do my own exhibition prints, I have found that anything over about 150ppi is seldom seen in the print itself. If the image is a technical or highly detailed shot involving product with sharp edges, a super high-res multi image panorama or overlayed vector art or text, then going to 200 or 250 ppi can make a difference. Those images are hung on a wall and viewed from a few feet away – unless you’re like me and always lean in, tip your head, and look to see what kind of detail there is in the print. Its sometimes an unkind thing to do, but I’m always interested in printing technique and it doesn’t get in the way of enjoying the image as art.

These large banners will hang overhead and the minimum viewing distance from a 6′ person’s eyeballs to the bottom of the banner will be 7ft, assuming they are looking straight up. To the center of the banner would be about 11ft. More realistically, the viewing distance as people move around the room will be well over 15ft. If 150ppi is great quality at a viewing distance of 3 feet, then what happens at about 15ft? Subjectively, I can tell you that they look great. I’ve seen the prints at Color Place, who is doing the printing, and was pleased with how crisp they look from a few feet away. The numbers support this. There is a simple, linear relationship here. A viewing distance of 3ft is 1/5th of a viewing distance of 15ft. Take our target high quality image of 150ppi and divide by 5 and you get, yup, 30ppi. There is a good article on this topic here.

The ink droplets from the HP Scitex XL Jet printer that Color Place uses are pretty big. Well, big in the ink jet world. They are specified at 80 picoliters. By comparison, my Epson 7600 has a minimum drop size of 4 picoliters. That’s billionths of a liter, which is small! What that means is that the image from the XL Jet, when viewed closely, has quite a bit of dithering, a kind of noise pattern that results from the low printing resolution. That is actually a very good thing in terms of percieved sharpness. At a distance, that noise actually sharpens the image. Try it yourself. Print a photo, then add noise to it in your image editor. Put them side by side and step back. Often the noisier image will appear sharper.

I visited Color Place when they were printing and finishing the banners. The printer is really big, and so is the head assembly. I complain about buying 220ml cartridges for my printer. These guys buy ink by the gallon. When the print technician saw me photographing the bottles on the floor he said “Oh, that’s just junk ink. We can’t use that. Here’s the new stuff” and then showed me the cabinet with the new gallon jugs.

Perhaps the biggest surprise is how inexpensive this kind of banner printing has gotten. It costs me about $2/square foot in ink and paper to print on my Epson printer. You can get banners made for around $3 to $4 per sq. ft. You aren’t going to hang them in a gallery, but it is a blast seeing images this size.

I needed something higher resolution than my D300 for the big banner project. I’ve been a Nikon user forever so was leaning toward the D3x, since it would give me twice the resolution of my D300 plus a full size sensor and lower noise. I also considered the Canon 5D MkII and some medium format cameras. The Canon was a little lower resolution than the Nikon and online reviews and samples showed the D3x to have a (sometimes small) lead in image quality. The Hasselblad CF22 shoots at 4080 x 5440 pixel. The photos are to be printed at 18ftx9ft, or a 2:1 aspect ratio. So the number I care about here is the 5440 pixel horizontal resolution, which is less than the 6000 pixels of the D3x. I looked at specs for the Hasselblad CF39 at 5412×7212, which is more pixels and a bigger sensor. But considering that that’s only a 20% increase in resolution and it would be a totally new and strange system to work with, the D3x was the sensible choice for me. Plus, the camera is laid out just like my D2x so felt comfortable and familiar.

I rented it from an equipment house in Dallas and took it to the studio for some tests and to zero in the focus on the 3 prime lenses I was planning to use, a 50 f1.4, 85 f1.4, and 105 f2.0. I was also borrowing an 80-200 from a friend. I was going to be shooting close to wide open for limited depth of field and found that both the 50mm and 105 had to be compensated for in the setup menus.

This was going to be six shots spread out over 3 days. The first day of the shoot went well, having the D3x tethered gave me quick feedback on sharpness and cropping, although the Nikon Camera Control Pro 2 software just about drove me crazy. (that’s going to be another post!) The morning of the second day was when things changed. I was shooting some test exposures and when I looked down at the camera’s display, I was looking at the previous shot. I shot a few more and noticed that the counter was not incrementing. Another photographer was assisting me that day and he shoots with a D3, so knew this camera well. We had been shooting on dual CF cards in Backup mode, so the images were supposed to be written to both cards, but nothing was being written.

We pulled the cards and swapped them. Nothing. We tried the cards one at a time. No data was being written but the camera thought it was fine and showed no errors. Then I tethered to Camera Control and was able to control the camera but after exposing it wrote no files to the laptop. Then we tried Live View, both tethered and stand alone and neither worked. We did the 2 button reset, changed batteries, changed lenses, left it off for a few minutes with no battery, stood on one leg, basically tried everything. When we inserted and formatted yet a third card, it started shooting again but after 3 good frames the last one came up striped green on the screen, the kind of green that makes you stop breathing for minute if you’ve ever gotten that kind of data corruption before. That’s when we decided to stop shooting and reschedule for the next day. The camera had failed and I had a client looking at me from across the room. Normally, I would have a backup body, but this was a special situation with a rental and we just had to quit for the day.

We were really stunned that the D3x failed, as was the rental house. I was able to get another rental from another place in town and finish the job, using the same memory cards and same setup with no more problems. Unfortunately the rental house was unable to make it fail again and suggested that it was not the camera but the memory cards. That conversation didn’t end well, but I did get the job done, and the quality of the images was really impressive.

Three of the six shots were done with just ambient fluorescent light and a couple of reflectors. I used a Lastolite grey card for color and it worked well. A sturdy tripod was really important here since even a tiny bit of shake could really mess with a 1/4 second exposure with an 85mm lens.

Here’s the setup for two of the shots with the final images:

I wanted to be able to use my laptop tethered while shooting. I looked into a couple of commercial laptop trays, but they were kind of overkill. I used an extra Bogen Magic Arm and some plywood to make a usable platform. I cut a piece of 1/2″ plywood slightly smaller than the footprint of the Vaio laptop and inserted a 1/4″-20 threaded insert into the center, dressed it with some black gaffer’s tape, and made a strap out of velcro. It gave me a stable, adjustable platform for the laptop.