2009-09-28

California Screaming

Yes, the title of this post is a stupid pun on that old song from The Mamas & the Papas. :-) I just returned from a 14-day visit to southern california, visiting & meeting new friends that I made through Picasa Web - and I want to begin sharing my experiences with a post about sensor dust that is meant as an update to my existing "Sensor cleaning mythbusting" series (parts 1, 2, 3 and 4), especially to part 4 (regarding the Arctic Butterfly, of which my effectiveness rating wasn't all too cheerful, ahem).

I was sharing lenses with fellow Nikon F-Mount user Shuwen because it was just very versatile - she was using her 16-85VR, I was using my 70-300VR (or 12-24) or the other way around, and whenever one needed just the other part of the zoom range, we would swap lenses quickly without having to unmount the backpack, find the lens, fiddle around with the lens caps, etc. - what I want to say is: we swapped lenses often.

And now look at this photo (yes, its not exactly a masterpiece, but thats not the point). That day, we got up at 4am (yes - holy shit!!!) for a shower, quick breakfast and a one hour drive from Palm Springs to Joshua Tree National Park to catch the sunrise, and then this:


Dawn, Joshua Tree NP (FinePix S5Pro, 1/7s @ ISO 100; f/16, 18 mm DX)

The evening before, I checked my photos - and there was no dust. That nasty little hairy fluff must have waited patiently in the mirror chamber of my camera to strike - it first appears on the 5th photo from that session. That's the problem: you change the lens, and you might just get some fluff into the mirror chamber of the camera. No immediate effect, but the movement of the mirror keeps it in motion and eventually it will be attracted by the charged sensor enough and there you have it. See above.

(of course, stopping down to f/16 is a lot, and thats why the hair shows so clearly, but nevertheless, it would have been just as easily visible at f/8 - because the spot in the middle is there right from the 1st photo of that day)

Guess what - after I removed that fluff from the sensor with the Arctic Butterfly the exact same thing happened again to me again only two days later. I never had that happen to me so much before. So, I've learned my lesson from that, and its:

First: carry something like the Arctic Butterfly with you at all times for "emergency situations" like these (because mine was in the hotel room a one hour drive away, in other words: useless). Unfortunately, my camera has no built-in sensor dust removal feature - I can't really give any good advice what would be a good practice for cameras that have this.

Second: check your photos often for dust and fluff - you can do that even with the camera's tiny display, and here's how I do it:
  1. Stop down the lens as far as possible. That will be somewhere in the range of f/16 to f/32 usually and this will make the dust really sharp and visible. Camera shake is mostly irrelevant, because the dust moves with the sensor.
  2. Point the camera at an evenly lit surface or light source (the blue sky, a white wall, whatever).
  3. Set the camera to manual focus and completely de-focus it (ie. set the lens to its close focus minimum so that everything that you normal would see is completely blurred)
  4. overexpose by at least one stop (no matter which metering method you use) so that you get a bright enough image even when you're in direct sunlight
  5. release the shutter and use the image control at maximum magnification. Scroll across the entire image and inspect it carefully - when you're in bright sunlight, put the camera in your camera bag or hide it in a jacket so that you get a good look at the display.
And with a big fat fluff on the sensor like shown above, it actually might work to just set the camera to cleaning/inspection mode (mirror locked up, shutter open) but keeping the lens mounted, then pointing the camera down to the ground and shake it wildly. I haven't tried that because I noted that dust fluff too late (and I hope that I don't have to try it, but I think it will happen).

But there's more to come. When we got back to the hotel I removed the fluff with the Arctic Butterfly. And yes, I have to correct my first evaluation: it actually is somewhat useful. It just needs a lot of swipes and spins to get small normal dust particles (but with fluff like above, one swipe should be enough).

But somehow, I managed to get some greasy dirt on my sensor while cleaning it. It might have been lurking somewhere, or I somehow contaminated the brush of the Arctic Butterfly with it, I don't know (what I do know now is why they sell a version with replacement brushes that is even more expensive! LOL). A big fat smear covered the lower right corner of the sensor - and that's the upper right corner of the frame, what an excellent spot for dirt when you're about to make photos of sea & sky...

So, the total desaster was there: I needed a wet cleaning to remove that grease. And of course, there's no camera store in Palm Springs (golf players don't need it), the next store that had sensor cleaning swabs was a half-hour drive away - and they ran out of the cleaning fluid! So, what to do?

I still had some cotton swabs (Q-Tips) with me and after a pilgrimage through three local pharmacies (well, more precisely, two - because the first one that was listed in Yelp didn't exist!) one was actually carrying high percentage alcohol and donated a small portion of it for my cleaning purposes (so: thanks to the friendly pharmacist of the medical court at East Tachevah Drive, Palm Springs:-).

Problem: my cotton swabs apparently weren't clean. Or the alcohol I got from the pharmacy wasn't pure enough. Because after it dried, I had a nice hazy residue on the sensor - which I carefully removed with a microfibre tissue. After that, I gave the sensor 20 swipes with the Arctic Butterfly, then checked again as described above: clean at last!

One problem with using cotton swabs though is that you might actually touch the mirror chamber of the camera with it, and you bet that one fluff of that cotton will come off there, going to ambush you... therefore, I think I'll carry some wet & dry sensor cleaning swabs with me from now on...

2009-09-06

Resolution - S5pro vs D90

Thanks to a friend who lent me his D90 for a weekend I was able to compare the resolution of the D90 (real 12mpx) with the S5pro (fake 12mpx, the special honeycomb layout of the Fuji SR sensor applies).

It was quite clear that the D90 would be the winner, but I wondered how big the difference would be. The subject of the test was once more my trusty CD rack (lots of fine details), I set both cameras to their base ISO (S5pro: 100; D90: 200) and used my AF-Nikkor 50mm/1.8D stopped down to f/5.6 for the test (it performs best in the range of f/4 to f/5.6 or so).

Both cameras where on the tripod of course; the S5pro was released with manual mirror lock up and cable; the D90 with mirror delay and self timer (I had no cable for it).

White balance has been neutralized with the picker on one and the same CD (not in the example frames, only in the full pic), Fluke's "Six wheels on my wagon", it's not possible to simply set the absolute same color temperature because of the different AA filter that the S5pro uses.

Lightroom's standard sharpening for Landscapes was applied to both images (yes, its common knowledge that the S5pro needs more sharpening because of its stronger AA filter, but what good would the comparison be with different sharpening?).

I picked two crops from the resulting frame to show the differences (please click on the photos just as usual to see the "real" 1:1 crop).


S5pro with 50/1.8D @ f/5.6
(FinePix S5Pro, 1.6s @ ISO 100; f/5.6, 50 mm DX)


D90 with 50/1.8D @ f/5.6
(NIKON D90, 1"s @ ISO 200; f/5.6, 50 mm DX)

Its quite visible IMHO that plain color surfaces (like the orange on the Etnica album's spine) are much more smooth and homogenous with the D90; the S5pro shows a certain grain even at base ISO (thats something I noticed for quite a while; I always considered it part of the special appeal of S5pro images).

The second set of crops clearly shows the limits of the S5pro. Check especially the spine of the Dire Straits albums and the Dub Trees album:


S5pro with 50/1.8D @ f/5.6
(FinePix S5Pro, 1.6s @ ISO 100; f/5.6, 50 mm DX)


D90 with 50/1.8D @ f/5.6
(NIKON D90, 1"s @ ISO 200; f/5.6, 50 mm DX)

Its quite remarkable IMHO that the "cheapo plastic" prime AF-Nikkor 50mm/1.8D performs so well even on the D90 with its 12mpx sensor (and if you listen to the herds in some forums you usually hear its "nowhere near as good" as the 3 times more expensive 50mm/1.4).

You can download the original files (full resolution, sRGB JPEGs, 90% quality export from Lightroom) from my webspace: Fuji Finepix S5pro and Nikon D90. Take care: each file is about 4MB (right click and "save as" if you don't want to view it in the browser).

And yes, the CD collection was more expensive than all the camera gear that I own today. :-)

2009-09-01

Liquid Light


Liquid Light (FinePix S5Pro, 1/200s @ ISO 400; f/22, 28 mm DX)

The Schwarzachen creek near Laubau, Ruhpolding in backlight. I was on my way back to the car from a little hike into the middle one of the three Kraxenbach valleys (german side of Sonntagshorn peak) and the sun was already long at half past 6 in the evening.

The title "Liquid Light" just popped up in my head instantly as I saw the sunlight glistening on the water. At first I thought it would be futile to try and make a photo in such a strong backlight situation.

And in addition to that, coming out of the shade of the trees and making some hand-held photos of Toni jumping around and being funny, I still had the camera at ISO400 - argh! Its a challenge sometimes to stay calm and focus on the sheer technical aspects during the excitement of a beautiful scenery and photo occasion.