Tuesday, August 22, 2006

Maximize Sharpness!

having discussed sharpness and resolution, here's my formula for maximizing both.

1) use the most pixels you can. Stitch where possible, make sure that you frame carefully so you don't have to crop both dimensions. This of course requires an accurate viewfinder - often with less expensive cameras the lcd is accurate but the viewfinder is not so you can shoot and double check that you got it right. Know your minimum f stop for maximum depth of field without adding so much diffraction loss that you don't gain - for me and the full frame 1Ds2 this is f 16, for a consumer grade digital point and shoot this is probably f8, for a 4X5 it's f32. If you don't need the depth of field, then use a wider apperture.

2) use a decent tripod. Use as little centre column raise as you can get away with, normally none.

3) balance the weight of the camera and lens over the tripod head. For longer lenses this means using a lens collar - technically my 70-200 f4 lens doesn't need one but I get sharper pictures when I use one.

4) consider using an L bracket. When you flop the camera over on it's side you basically create a tuning fork - something sticking out to vibrate.

5) Use mirror lock if you have it - arguably it's most advantageous from about 1/30 second to 1/2 second but I use it for every shot - given Canon's awkward way to access mirror lock, it's easier to leave it on permanently.

6) Use your lens shade and shade it from direct sun with a hat or even your hand (not touching).

7) make sure that your foot isn't an inch away from the tripod foot and actually moving the ground. Consider standing upwind to block the wind or even better using an umbrella to block the wind. Keep in mind the wind will play on large camera straps and increase vibration.

8) make sure that all settings are locked down on the tripod and that there is no movement - if there is, track it down and fix or replace the offending part. I had a lovely Berlebach tripod which worked well in almost every way exc. the join between the centre column and the camera base which had a little play. If I attached the camera directly to it, all was well, if I had a tall ball head, there was enough leverage to create sig. movement. This raises another point. Lots of heads have cork or rubber surfaces but that inherently creates more movement. Remove this where possible - better you scratch the bottom of the camera than you have blurred pictures.

9) in windy conditions consider actually puting your hand on the camera (or lens if a longer one). This doesn't work with 2 second exposures but works quite nicely with 1/30 second exposures and a 300 mm. lens. Here I'd shoot several frames to give myself a choice.

10) Use a cable release and use it in a way that puts the least strain on the camera if you wiggle the cable release. I'm fortunate in one way that the Canon cable release is electronic and therefore very flexible and long so transmitting vibration isn't all that likely. On the other hand, it's long enough to get wrapped round tripod fittings so I still need to be careful.

A poor man's cable release of course is the self timer. I prefer a cable release so that I don't have to worry that all vibration is gone by two seconds, and also because when photographing in gusty winds, I want to shoot at the optimum stillness. It's not uncommon for me to let the 30 seconds activation time for the mirror lock to expire because the wind didn't settle and I just do it again.

11) In processing the image, use a good sharpening algorithm - mine is to raw process one size larger than native, use smart sharpen 300 1.1 and then do an output sharpen with photokit sharpener matched with the dpi of the image.

1 comment:

Anonymous said...

Hi George,

Enjoying your newly found blog. Learned about it by revisiting your site after seeing your pictures again in B&W magazine. I see that you are a fan of PhotoKit Sharpener too. I love it, but why are you using something different for a capture sharpening? Also why are you RAW processing at one step up in resolution rather than using the native resolution of your capture.

Thanks for any comments.