Showing posts with label Photography. Show all posts
Showing posts with label Photography. Show all posts

Sunday, 19 September 2010

More manual focus

I got a response from Nikon service reasonably quickly, however the response itself was fairly useless. They couldn't give me a price for doing any work without seeing the camera and said that if the focus screen wasn't a Nikon one (as far as I know Nikon don't make split prism screens for their DSLRs) they may replace it during the service.

So, next stop was to try and find a local camera repair shop and see what they could do. I phoned RCCR as they're pretty close by and had a very helpful conversation with Ray. Whilst he couldn't actually help me he was very happy to explain in detail what might be going wrong.

Conveniently I've also currently got a trial version of Lightroom 3 installed. I'd read that this allows you to do tethered shooting, and sure enough after changing the USB setting on my camera, tethered shooting worked without a hitch. This allowed me to check the focus of my photos at 1:1 on my laptop as I took them, which is a lot more useful than looking at them on the back of the camera.

Ray was worried that the mirror adjustment might noticably mess with the autofocus (contrary to what I'd read previously). It seems that he was right, thanks to the tethered shooting I saw a noticable change in sharpness as I moved the mirror adjustment this time.

According to Ray I should have seen a difference in focus performance after taking the existing shims out. The first time I took them out and put them in a few times and didn't really notice anything changing. This time I thought that instead of using the autofocus and seeing how far misaligned the split prism was I would try focusing with the split prism and see if I could tell the difference in the pictures. This turned out to be a much better idea and showed that focusing with some thick paper shims I made was clearly worse than no shims.

Unfortunately because in both cases the focus was off in the same direction I have to conclude that either my camera just can't do accurate manual focus, or that I've got a duff focus screen - if it's thicker than the existing screen plus the shims then it's probably going to be impossible to use.

So for now I've re-installed the original screen and gone back to my mostly reliable autofocus. Drat.

Saturday, 11 September 2010

Manual Focus

I've been interested in getting a manual focus screen for my D300 for a while, but the price of the Katzeye screen has put me off. However, I recently found out that you can get focus screens from on eBay for a fraction of the price, so I thought I'd take the plunge.

I chose a 180 degree split, which just means that the split line is horizontal. That means that it's easiest and most accurate to focus on vertical lines and impossible to focus on horizontal lines. The other common types are:
  • 45 degree single split. This is the same as the 180 degree but with the split line at 45 degrees. One of these screens will let you focus on both horizontal and vertical lines but with less accuracy. For maximum accuracy you'd need to find a 45 degree line.
  • 45 degree double split. As the name suggests this screen has two splits. Apparently the purpose of this is to help get the sensor plane aligned with what you need to focus on and is mainly useful for photography flat objects like coins and books.
Removing the focus screen on a D300 is slightly fiddly as you really need to remove two tiny screws (it's probably just about possible to change the screen without removing the screws but I wouldn't recommend it). Once the screws, and small plate they hold, have been removed you just need to unhook the wire holding the screen in and replace the screen. If you plan to do this you should follow some proper instructions.

Unfortunately, so far things have not gone to plan. Fitting the screen is in theory fairly simple. In practice the particular screen I have was about 1mm narrower than the original screen and so the wire clip holding it in slipped off one side and so didn't hold it as securely as I'd like. To counter this I removed the clip completely, bent it in slightly with some pliers, and then put it back. Not too hard but it shouldn't have been necessary and required some confidence.

Once securely fitted I tested the screen by setting the camera to autofocus and seeing if the screen agreed. It didn't. I half expected this as there were two shims between the original focus screen and camera body and I'd read guides that said to leave them in and other guides that said to take them out. So I tried removing one, then both shims. Unfortunately that didn't really help.

At this point I had to do some more research. One thing I learned is that with a 1.5mm allen key you can adjust the mirror resting position to a small degree by turning a small screw next to the bottom right hand corner of the mirror (as you look into the camera). This helped and got the focus quite close, but I couldn't adjust it far enough. Also please do note that I later found a few warnings about how this adjustment might mess up your auto focus, however, I found just as many notes saying that people had done this without any problems. For the record I have not seen any problems with my auto focus but haven't tested it thoroughly.

Finally after more research I've found a few people saying that modern DSLRs are simply not set up for manual focus at all accurately (although mostly they're not far off). Given the general lack of complaints about these screens I think in my case I may have got unlucky and got a camera that is so far away from accurate manual focus that the small adjustments available to me aren't enough.

So, I've admitted defeat and emailed Nikon Service to see what they can do adjust the focusing and how much it will cost. More details to follow as things progress...

Sunday, 25 April 2010

Printing pictures

So today I wrote my first plug-in for the GIMP (it's really just a script but as it's written in Python it goes in the GIMP plug-ins directory).

The reason for the script is that I've found a print lab in the UK that will accept files over the internet and just print them without messing around trying to "correct" them. They give you an icm profile for their printer so that you can convert your images. They're called ProAmImaging. There are other places on the internet that seem to do the same but the sample prints I got from ProAm look pretty good so I'll be using them for now.

Anyway, the flip side of them not touching the images is that I need to scale and crop the images to the exact number of pixels needed for the print size I want. This gets tedious very quickly so I decided to write a script to automate things. What the script does is take an image size in inches, and a desired DPI, and uses them to scale and crop the image to the correct size in one step. It's nothing you couldn't do with the existing tools, but it halves the number of actions you need to do, and does the maths to work out exactly how many pixels to crop automatically.

I'd like to improve it so that it activates the crop tool with the correct aspect ratio so that you can select the crop manually, however it's not obvious to me whether that's even possible so I'll leave that for now.

A couple of things worth noting when writing Python scripts for the GIMP:
  • If python doesn't like your script for some reason the GIMP won't tell you why, your script simply won't appear in the menus. There's probably a way around this but it never got so annoying that I had to figure it out. If you start with a known good script (there are loads at registry.gimp.org) and test regularly it's OK.
  • The GIMP/Python documentation seems to be rather sparse (http://www.gimp.org/docs/python/index.html) and the python object interface is quite small. However you can call all the libgimp methods from Python using pbd.methodName().
  • For debugging pdb.gimp_message("Your message here") is rather helpful.
Please note that this is the first time I've written a GIMP plug-in and almost the first time I've used Python so the take the above with the appropriate dosage of salt.

And finally the script itself - ScaleAndCrop.py, I hope someone else finds it useful.

Tuesday, 20 April 2010

High pass filter effect


This technique has been written about a number of times, not least by Strobist, but I thought I'd see how it works on this rhino rather than a person and what happens if you mess around with it a little.



I use the GIMP for post processing my photos, and by default it doesn't have a high pass filter. So the first thing to do is to head to the GIMP plugin repository to get one.
On Windows put the .scm file in "C:\Documents and Settings\Your User Name\.gimp-2.6\scripts"
On OSX put the .scm file in "~/Library/Application Support/Gimp/scripts"

If all went well you should now have a "High Pass Filter" entry under Filters > Generic.

Now we're all set, so:
Open up the photo you're going to work on
Select Filters > Generic > High Pass Filter
I'm working on a picture that's about 800 pixels wide so in the dialogue I chose:
  • Filter radius: 16
  • Contrast Adjust: 0
  • Mode: Greyscale
  • Keep original layer: Checked

Other values will also work (see below for what happens) but a radius of about 1/50th of the image width is a good place to start.

Change the layer mode on the new grey layer to Hard Light
Tweak the opacity of the layer to taste

The image below has had a different setting applied to each quarter for comparison.

and the original again for comparison:
My favourite here is the radius 16 contrast 0 section. Adding extra contrast to the high pass layer adds contrast to the final image, bleaching out some of the colour. Increasing the radius seems to emphasise larger features turning the effect from a subtleish sharpening effect to something actually affecting the image more obviously.

Saturday, 6 March 2010

New Lens

I recently bought a second hand Nikon 70-210mm f4-f5.6 AF lens from eBay for £65 as I'd read some good things about it, and it comes from the same era as my 35-70mm f2.8 which I like very much (but which was also rather more expensive).

I already have an 18-200 VR so this new 70-210 is going have to be better than that lens to avoid going straight back on eBay.

I set the camera up on a tripod pointing down the garden and took a few shots with each lens at 200mm and 210mm with varying apertures. These are 100% crops from near the centre of the frame (unsharpened). The last shots are shrunk versions of the full image to give an idea of the extra reach the 70-210 has.














These results are a bit disappointing, the 70-210 is clearly putting more pixels over the same area, but I'm not convinced it's actually capturing much more detail (especially at f5.6).

I've also just taken the 70-210 down to the park to take some photos of my daughter, and I've learned how important shutter speed is with this lens at 210mm. I guess I've been spoiled by the VR on the 18-200 but it does seem that you need a shutter speed of 1/400s or faster when hand holding which means you need quite a sunny day to also get down to f8-f11.

That said, I did get some quite nice shots in the park and noticed that it might be front focusing slightly which I should be able to correct for in the camera.

In terms of build quality the 70-210 wins, feeling much more solid. But it loses out in terms of focus speed to the 18-200's built in AFS. The AF-D version of the 70-210 is supposed to focus twice as fast which would certainly help (but they are also rarer and more expensive on eBay).

So for now I'm leaning towards the 18-200 as it's only marginally less sharp, but focuses faster, has VR and can go all the way to 18mm. However, if I didn't already have the 18-200 I'd certainly be happy with 70-210, especially at the bargain price it can be had for.