Monday, August 19, 2013

Fun with filters

We fit protective filters to all our lenses which can accommodate them. Some people prefer to use them, especially on expensive hired equipment, so hopefully they'll be happy. Some people prefer not to, and that's fine. The most common reason for people not wanting to use protective filters is their concerns about what it does for image quality. So we thought we'd do some experiments.

First off, here's our man Richie with his test setup. From left to right:

Here's a close-up of the test rig. As you can see these are all decent quality Hoya Pro 1 Digital filters, not your cheap rubbish. You're looking at about £1000-worth of filters on the end of that lens.

Here's another view of the test set-up. You can see that the front filter isn't exactly in optimal condition. It's got a huge crack and it's covered with dust and fingerprints and other crap. And ALL 34 OF THEM are rejects too: cracked, chipped, scratched, indelibly marked, that sort of thing. You wouldn't want to put one of these on your lens - and we wouldn't want you to - let alone 34 of them.

This is a test shot taken at f=80mm, without any filters. We'll use this to check what effect the filters have.

And this is a test shot taken at f=400mm, without any filters.

Here's what we got when we used all 34 damaged filters at full zoom of 400mm. Hmm. The contrast and sharpness aren't quite what they were. Still, there's no sign of the cracks and scratches in all the filters.

And this was an attempt to clean up the image. We've massively boosted the contrast, clarity, saturation etc. It's much better than the straight-out-of-the-camera shot, but it's still not brilliant. Still, we think it's surprisingly good bearing in mind what was on the front of the lens!

Here's what we got using all 34 filters with the lens set to its minimum zoom of 80mm. The vignetting is a little extreme, isn't it? That's because the lens is effectively looking down a long tunnel created by the filters. At 400mm, the field of view is narrow enough that the tunnel walls don't affect the picture; but at 80mm they do.

And finally this was an attempt to salvage the best image we could from the second image. Again it's not brilliant, but again all that crap in front of the lens hasn't done as much damage as you might think.

So what have we learned?

Well, probably not much that's of practical value in the real world.

We've learned that scratched, cracks and other crap on your filter don't necessarily show up on your image. Sure, you can degrade the image in terms of contrast and sharpness by attaching a really crappy piece of glass to the front of your lens; but the defects in the glass don't show up directly.

We've also learned how to make photos look just like they're on Instagram, using just a DSLR, a £2000 lens and £1000 worth of filters.

What about if we'd only used one filter?

Some people might be wondering what difference it would make to the image if we only used one good filter, instead of 34 bad ones.  The answer is that in most situations a good filter will have no noticeable effect on your image quality.

In the real world, we can only think of about three situations where using a *good* filter is going to noticeably degrade your image:

  1. If you're shooting into the sun, a filter can make the lens flare worse.
  2. With some telephoto lenses - especially image stabilized ones, we suspect - a filter can make the bokeh (out of focus background) look more harsh.
  3. Filters can cause 'ghosting' if you have lots of bright points of light in the image, as in astrophotography or night-time city photography.