I was going through my collection of filters, digging out coloured and UV filters in preparation for my jump to film photography, and came across a Rodenstock Light Red #25 filter. I had never seriously used it back in my photography classes so I wanted to see what it does to a landscape. I grabbed my Canon 5D and an incident light meter and quickly headed outside for a little “assignment”.
The guide that came with the filter recommends an exposure increase of 3 stops for this filter. I needed a “control” shot so I took one photograph without the filter using the recommended settings from my incident meter for F13. I then screwed on the red filter and shot one frame with the exact same settings. I applied the 3 stop increase to exposure, as recommended by the pamphlet included with the filter. All three shots were taken with a graduated ND filter in place to keep the sky under control.
Finally I headed into the house, imported the photos into Lightroom, and applied a greyscale development setting. I immediately saw a contrast difference with the filter but it was obviously due to the darker exposure. The 3 stop increase with the filter lost all the tonal detail in the sky. I “cheated” and used Lightroom to apply a 1 stop increase to the metered exposure with the red filter. This gave an image similar in luminance to the control photograph but I immediately saw the difference in the contrast between the two.
Theoretically, whichever colour filter you apply, you will darken it’s complementing colours. The red filter is meant to bring extra “pump” to the blue and green end of the colour wheel. This increases the overall contrast of the outdoor landscape, usually dominated by blues and greens. It was a good exercise to perform and I thought I should share my results. I will now have a better idea of what the filter will do when I start shooting large format (hopefully in the next week) and there is less of a chance of an unexpected surprise in the darkroom.
When shooting for DeepSkyStacker you’re supposed to get the best results when you calibrate your camera. Unfortunately I didn’t calibrate my camera at all when I shot the night for the first time so out of curiosity I shot about 8 dark frames at ISO 1000, F2 and 2 minutes each. By the time I got to the 8th exposure my camera’s CCD was quite warm and you could see it in the images.
If you look at the right-hand side of the image, you’ll see a large red splotch developing and innumerable bright dots scattered about the image. Shooting these dark frames gives the software a reference point to subtract the hot pixels from the stacked image. I suspect that DeepSkyStacker may consider these hot points “stars” and it could mess up the registration of the stacked exposures, thus giving a good reason to shoot these dark frames.
I also shot about 35 flat frames. Here you can see a “Flat” frame taken, where I stretched two layers of a white t-shirt and shot on ISO 1000, F2 on Exposure mode. The camera metered for 10 second exposures. The purpose of shooting flat frames is to provide the software with a reference point to eliminate the vignetting around the edges.
This is a stack of about 35 flat frames. If you look carefully you’ll notice that the center of the vignette is not in the center of the frame. I don’t know if the lens’s vignette is off center or my method of shooting the flat frame was incorrect. Either way it looks like something I’ll have to figure out down the road.
So if you look at this image, you can see the CCD heating up on the right-hand side as well as the strong vignetting on the edges. Next time I shoot the night sky, I’ll be taking about 20 reference frames for flats and 20 reference frames for the darks so that the software can virtually eliminate both problems. The FAQ for DeepSkyStacker recommends that you stack not only your light frames, but your calibration frames as well for improved image quality. If you only include a few calibration frames it won’t be able to average out the noise and you’ll introduce even more sensor noise.
It’s been quite difficult finding documentation on exactly what Light, Flat, Dark and Dark Flat frames are. Light frames are the shots of the stars, the ones with an actual subject. Flat frames are shots of a smooth, evenly lit surface in order to accentuate the lens’s vignetting. This can be done with a soft-box (an aparatus placed over the lens) or with white cloth stretched over the lens hood. I haven’t figured out the exact difference between dark and dark flat frames, however, and will experiment with these the next time I shoot the night sky.
Later this week, weather permitting, I’m going to try my hand at astrophotography again. I’ve seen some awesome pictures of the Milkyway Galexy and even some nebulas from people who’ve used image stacking. I want to bring in much more colour and to do that I’ll need a serious mount for my camera.
This is a Scotch Mount and you can get the plans to build one yourself at Philip Harrington’s website. What it will (hopefully) let me do is follow the movement of the sky for each exposure. Previously I had to keep my exposures down to 30 seconds max. Any longer than that and the movement of the stars across the sky causes streaks in the exposure. I need longer exposures, however, to have a chance at gathering photons from extremely distant celestial bodies. This means I have to track the sky with my camera.
I could use a proper equatorial mount but those things start at $800 and go up to $15,000. I’m just dabbling so I googled for a while and found that the Scotch Mount, which was first published the April 1975 issue of Sky & Telescope, should allow me to do this without spending an arm and a leg.
The scotch mount, introduced by George Haig of Glasgow (hence the scotch name) is meant to be pointed at polar north, or Polaris, the north star. Then you crank the handle a quarter turn for every 15 seconds on your watch. It raises the platform at (almost) the exact speed of the rotation of the earth.
So according to Calgary’s weather report, at the moment, it will be mostly sunny on Friday. I plan on staying up until about 3am that night with a pair of binoculars and a camera, seeing what I can capture.
Update (June 16, 2010): After building the mount, I sat down and read the “Using the Scotch Mount” section. Then when I googled other people’s scotch mounts I realized mine had a design flaw: I can’t articulate the mount itself. I thought I would just line the hinges up with Polaris by looking along the top of the hing. I didn’t realize you had to orientate the entire hinge so that it’s shaft lines up with polaris, as though looking down the shaft, it’s spindle would block polaris. I figured out a temporary fix by unlocking the tripod legs and having two of them laid entirely flat and one only 1 click away from flat. This gave it enough of an angle upwards to line it with Polaris. Now I just have to wait for good weather…
I stayed up late tonight and did a little backyard astrophotography. I don’t know much about astronomy but I saw some pictures of nebulas captured by people with the digital rebels and DeepSkyStacker so I figured I’d see what I could do. DeepSkyStacker is windows freeware available here.
I followed the tutorial written up at A.S.I.G.N. and got some not-bad results of my own. No nebula but I did manage to get a lot of stars. The sky was quite clear but I did the shoot at about midnight. This meant the sun hadn’t quite set 100% and there was still a blue cast to the sky overhead. I wasn’t up for waiting until 3AM in the morning to shoot photographs. Thankfully the moon was only at 10% and had set before 11:30 pm so I managed to get something, at least.
The images were shot with my 5D, the Zeiss 35mm at F4 and 1600 ISO. There were fourteen 30 second exposures stacked using the software. I used Canon DPP to export the RAW files to TIFFs. When it finished stacking them I thought the software was broken but I realized afterwards I had to play with the RGB levels sliders to get the recorded tone range situated properly in the curve and white-balanced. Then the empty white image turned into actual stars. I rendered with the saturation at about 18%. There were no calibration images used whatsoever. Next time the weather clears up overnight I’ll shoot some calibration images. It’s supposed to greatly improve your contrast and reduce the noise.
I’ve tried out Photomatix 3.2 to process HDR images of some of the photographs from yesterday’s shoot at the Calgary Brewery. It’s much easier to push the microcontrast further than Adobe’s HDR Merge can with Photomatix, if that’s the image you’re going for.
I’ve been excited about a special location I’m visiting tomorrow for a shoot and I spent the evening learning a bit more about HDR photography. Adobe Photoshop has an auto-import wizard to merge multiple, bracketed shots into an HDR image. First you shoot several images, maintaining the same ISO and f-stop, changing the Tv. If you change the f-stop you alter the depth of field mid-merge and if you choose an ISO other than the smallest available in your camera you’ll just introduce more noise.
Then you select the photographs in the import tool. It gives you a quick thumbnail overview for all the shots; it’s good to pay attention to this because in one instance I accidentally imported a photograph that wasn’t part of a bracketed series and didn’t realize it until later on.
I used the local-adaptation method of HDR merge. Here you set the pixel radius for microcontrast, as well as the threshold. If you set the threshold too high you’ll get haloing/banding of your tones.
You can tweak the tone curve in this mode, just like regular curve window. Here’s an example of a tone adjustment I made. The left hand spike is the dark ground regions. The right-hand spike is the sky region. You can use multiple tweak-points to give a nice contrasty spike for whichever regions of the tone you want. Here I made the ground region ultra-contrasty.
One thing I noticed is that the branches of the trees in front of the sky were shifted in tone due to the micro-contrast radius setting of the HDR merge. I think I would need to lower the micro-contrast radius or try to keep things like that from happening in the original composition of the shot.
I performed these experiments so that tomorrow I have a better picture of what the end-results look like when photographs are merged together. I plan on keeping this in mind as I shoot so I can attempt to compose my images in the best manner possible to take advantage of what Photoshop’s HDR merging can bring.
Because of the gloomy Calgary weather I was stuck indoors tonight. Yeah, go figure? Calgary has rain!? Anyways I had some fun with my flash unit and a bunch of coloured gel sheets. I set my 5D on a tripod, adjusted it to bulb exposure and turned it to day-light white balance. Then I turned it to F22 and ISO 100 to make it as dark as possible. I even put my polarizing filter on it to give it another 3 stops of darkness. What I wanted to do was to make it drag out as long as possible so I could fire off my flash with the test button 2-4 times during one exposure, and not expose the background too much.
I tried to get a little creative with my poses. It’s kind of like acting, like drama back in junior high. I figured out how I could twist my body to end up in unique spots, and then twist my body further to fire the flash upon myself without dropping the coloured gel I’m holding over top of it. The results are single shots with multiple images of myself in them.
The key is that the photographic camera is a time-based medium, really. When you press the button, the shutter is open for a period of time. On bulb exposure it’s as long as you want. You’re recording something, not necessarily simply taking a photograph. You can get creative with this and lots of people do with tricks like multiple flash fires on a single exposure. This is what slow-speed-sync is. You get a still shot of your well lit main subject because the flash is fired for a fraction of a second. The rest of the exposure is recorded during the length of time that the mirror is up and the shutter curtains are out of the way. Once those curtains closed, you’re done your recording and left with a sharp image of your subject and a blurry, moving image of everything else.
In the early days of photographic technology, photographs were recorded over minutes, let alone seconds. Reportedly, the world’s first photograph of a person is Daguerre’s Boulevard du Temple, Paris 3.
It was a 10 minute exposure of a busy intersection of Paris. The only person visible is the man with his leg up on the post in the lower left of the frame. Everybody else wasn’t standing still so they weren’t recorded in the picture.
The principle applies to our digital cameras of today. Even shutter-less cameras perform in the same manner: they only record at a specific time period. This can be controlled in a creative manner that only the imagination can limit. What all this means is that even with digital your creativity can go beyond the perfectly still, perfectly sharp photograph.
Mouse-over to compare images (they may take time to load)
Filters make a big difference, especially for bright outdoor photographs that include a lot of sky. Lots of people think that photoshop is all they need to control their images but there is a lot you can do, even in the age of digital, before the light hits your camera’s sensor.
The polarizing filter is one if the first accessories any landscape photographer should have, or even photographers who like to include the sky in their photographs. The polarizer lets you control the tones of your sky and gets you much richer pictures. They can be tweaked as well. Polarizers cut out reflected light that comes in to the camera at a certain angle. The optimum angle is 90 degrees from the sun.
You can play with your polarizer and see the difference in the sky when the arrow on the rim is pointed towards the sun, and 90 degrees from it. Keep in mind that the sky is a dome, not flat. Your polarizer’s effect will change over your photograph if you’re shooting at a wide angle. You can see it in the above photographs, comparing the top of the frame to the bottom.
I purchased a Hoya 58mm circular polarizing filter for about $63 from The Camera Store. It has to be circular or else your auto-focus won’t work. I chose 58mm because most of Canon’s half-decent, non-L lenses use this size of thread. I purchased an adapter to step up the filter size from my 50mm 1.8 II so I could use the same filter on it. You can spend the big bucks and get a mult-coated Singh-Ray polarizer but if you’re on a budget like myself, the Hoyas are fairly well reviewed and getting those colour/contrast improvements now is far more valuable than spending $250+ for possibility of some added sharpness.
I pack an ND (neutral density) graduated filter with me whenever I expect to be photographing out-doors. These babies let you do HDR right in your camera! They are sheets of plastic, or glass, that gradate from clear to dark. You use them to darken the sky, keeping it from over-exposing, thus allowing you to properly balance the exposure of the ground with the sky. This is a very difficult task that would require either a sacrifice of either, or two exposures to be composited in photoshop.
I shoot with a cheap Cokin 3 stop ND graduated filter. If I’m shooting at 50mm or so, I use a holder that attaches to the lens’s threads. If I’m shooting wider I hold it in front of the lens with my hand, otherwise the lens-mounted holder causes vignetting at the corners. It cost me about $43 canadian from The Camera Store and was worth every penny. Highly recommended.
