Wednesday, April 14, 2010

Depth of Field … One More Time *updated*

Moose responded in the comments section, see my reply and new shots at the end of this article.

Ok, just one more time. Taking a quote from an early mini-review of the HS10 … I already addressed this once, but I just have to do it again.

If you plan to take most of your photos outdoors in bright light situations (kids soccer games/wildlife/landscapes), the HS10 can stand up to to many entry level Digital SLR’s in terms of image quality. Pro photographers may be able to see subtle differences between the images when blown up to 100% on their computer monitor, but in a 4×6 print most people wouldn’t be able to tell the two images apart. Again, this only applies to photos in outdoor (bright light) situations. Once you go indoors, DSLR’s have the advantage due to their larger sensors and ‘noise free’ images in low light situations.

The key comment in there is “in a 4x6 print most people wouldn’t be able to tell the two images apart” …

Now, for landscape shots in good light this is probably fairly true. So long as the ISO is low and the print very small, the two will look a lot alike. But tiny sensor cameras tend to smudge diustant foliage and other details, and that will show on a print. I don;t think he can say such a thing with a lot of caveats.

But when he mentions kids’ sports, things get *really* dicey. Say you are shooting from short to medium distance at focal lengths like 80 to 100mm, what difference would we see?

To show the potential differences, let’s take the Panasonic ZS3, a camera with a sensor about the same size as that in the HS10, and the D700, a full frame dSLR. This test simulates the kind of differences one would see when you are shooting at 30 feet and there are people and fences on the other side of the pitch.

So … the ZS3 cannot throw much out of focus at these relatively normal distances. Here is what this test looks like.

P1000321_zs3

The background is slightly blurred, but it is pretty obvious what it is. In the soccer scenario, the background would be pretty clear, if not perfectly focused.

Now the D700 rendering from the same position.

DSC_0068_d700

Sorry, but I don’t care how small the print is … you’d have to be blind to see no difference in the subject isolation. And that’s why people buy dSLRs …

Disclaimer: There are more variations of subject and background distance than you can shake a stick at. So don’t bother pointing out that there will be situations in which the difference is less than this … much less, even. That’s all pretty obvious. But there will *always* be a difference. And it won’t be all that small if the dSLR has a nice lens on it.

Moose’s reply: You can read the whole reply in the comments section. But the gist comes down to these points:

1. Test shots between 400mm and 720mm have similar DOF to 200mm shots on D90 (which is 300mm effective.)

2. Most soccer moms won’t use a D700 with 70-200mm f/2.8 lens.

3. 300mm on HS10 gives wonderfully blurred shots at similar distances to shooting soccer.

4. Most DPReview members have confirmed shallow DOF at longer focal lengths. Fuji did a great job.

5. dSLR has advantage of closer distances in my earlier test. Great portraits can be had at close distances by zooming to 300mm and beyond.

Now, my responses:

1. Irrelevant. First, I perform apples to apples tests, not apples to oranges. Second, shooting at massive focal lengths means stepping a long way back, behind the audience, to get the same framing. Not goona be as useful as it sounds, unless you want head shots all day, and there are much better situations in which to shoot those.

2. Fair enough. So I will repeat the test with the D300 and a consumer lens.

3. No, it does not. I will prove that in a moment.

4. DOF Master will confirm that based solely on the laws of physics. We don;t need “confirmation.” Now, if you had said “wonderful bokeh”, that would be a different statement, since that depends on control of aberrations in the lens.

5.  Not sure how the dSLR can have an advantage when shot from the same location.

So … let’s repeat the test. D300 (same sensor size as *all* Nikon consumer dSLRs) with 70-300VR (inexpensive consumer lens) set at 180mm (270mm effective) shot against the F70EXR at full zoom (which is 280mm or 270mm, I forget … I put 280mm on the test image.) 6400iso for the dSLR against 800 ISO with flash for the F70EXR (to get something close to sharpness indoors.)

Both shot at f/5.6 and the the consumer zoom shot at f/22 for illustrative purposes.

Same subject as the original test, shot at about 12 feet and the background (same hound’s tooth as last time) about 8 feet behind. This more or less simulates shots just past half way across the pitch. The relative amount of blur will be consistent, if somewhat more or less depending on subject distance.

d300_f70_bokeh_test

A few observations:

1. Laws of physics are not to be denied. The DOF on the small sensor cam (and this sensor is *larger* than that in the HS10 remember) is not even as shallow as the consumer cam at f/22.

2. The clarity of the consumer dSLR at 6400iso (and remember that the D90 is slightly cleaner and the D5000 slightly cleaner again than the D90) is far better than that of the F70EXR at 800iso with flash assist. Translated to the pitch, this means that details in shadow (midday creates those) will be awful. And the struggle for shutter speed for evening games (kids play a lot of those) will be never-ending. Boosting ISO will not be a good answer, even with the alleged extra stop or two of efficiency on the BiCMOS sensor.

My point is simple … for soccer moms shooting the latest kits (D5000 2-lens kit) that cost a small premium on the price of the HS10, the output will destroy that of the HS10 in all lighting conditions. Which means that even 4x6 prints will show obvious differences in overall image quality for most shooting circumstances.

Which further means that I cannot agree with any of your responses. Sorry about that.

Update 2: I just calculated the crop factor for this camera … a staggering 5.7 (720mm effective / 126mm actual.) Wow … that’s gonna give you an equivalence at f/5.6 of f/32 for depth of field, f/22 on APS-C. Not too shallow, although at 720mm you will still get blur if close enough to the subject. The wide angle at 2.8 will have depth of field equivalent to f/16, f/11 on APS-C. Again, very wide depth of field, especially at 24mm effective.