Sunday, May 16, 2010

HS10 vs D300 vs F70EXR – A Test of Reach – Part 2


In part 1 of this series, I outlined the test methodology I use to test reach on the HS10 against the possibility of cropping a typical dSLR sensor and megazoom to match. I even add the F70EXR, a long zoom camera itself, to test whether something like that might suffice for some people.


Round 1 was a vehicle in a garage and the HS10 eked out a narrow victory, albeit with caveats (CA being the main one.) In round 2, we turn to a pretty classic subject: distant foliage. A jumbled set of branches behind which are many leaves.

Since the HS10 does resolve a bit more detail than the 18-200VR after cropping, it should have the advantage. But, at the same time, the HS10 suffers from some edge integrity issues. I didn’t mention them in part 1 because they actually enhanced the sharp look of the brick details. But branches need to retain a rounded look … they are definitely a 3-dimensional element, so anything that causes edge issue is going to make it difficult for the HS10 to maintain the 3D illusion.

The F70EXR, on the other hand, resolves distant detail rather poorly by comparison. At this distance and magnification, it is simply overmatched. But that does not mean that it will suck at web sizes. It may very well do just fine. Let’s see.

Note that I added Topaz Adjust 4 to extract maximum detail from all three. So the potential of your images does rely on some skill with post processing.

So, the first image is of course the baseline against which all others will be compared. The HS10’s shot:

R2_HS10[1]

That’s one jumbled image :-)

Now, the D300’s image showing the crop, followed by the crop itself.

R2_D300_CROP[1]

R2_D300[1] 

You will note the fact that I also needed to rotate the image a bit to get them to match. This puts the D300 at a further disadvantage, since its pixels have already been rather heavily processed by the rotation algorithms. But we press on.

The F70EXR rounds out the test. And its image is surprisingly good at 800px, don;t you think? (Remember to always click thorugh to see the bigger images.)

R2_F70EXR_CROP[1]

 R2_F70EXR[1]

And finally, what you’ve been waiting for … the crops. Note that this original file is almost 1MB in its original size, so be warned. It is much bigger than 800px on a side, so you may need to click twice, once to load the full image and once to expand to original size in your browser window.

R2_CROPS[1]

 

And some observations:

  • The smaller branches do retain their fine appearance. This gives the impression of better detail in some areas.
  • The main trunk up the left side is absolutely more detailed in the dSLR shot, despite its many disadvantages in this test. Here is where the quality of the pixels come to the fore.
  • Similarly, the main branch in behind slashing from the middle of the frame upwards and to the right has a lot of detail in the dSLR image and is essentially smooth in the HS10 image. This is noise reduction rearing its ugly head.
  • As you penetrate deeper into the image, you see places where the leaves are well defined in the HS10 image and then the leaves behind or between are mushed together. The D300 tends to retain more definition between the leaves. They retain more of an individual character, despite the edges been a bit thicker (from the fairly severe upsizing.)
  • The F70EXR retains a surprising amount of fine detail … but the background is completely gone. It is overmatched again. Too few pixels and too much enlargement. Its noise reduction would likely have gone to town on the smaller low contrast details too, and that is plainly evident in the crops.

So, the landscapers and birders of the world remain better off with the dSLR combination, which should really come as no surprise. Even cropped, there is more available low contrast detail and the 3D effect remains largely intact throughout the image. Of course, you cannot get at all of that detail straight from the camera. You must process to extract it. So of course, if you prefer to shoot and post or shoot and print, the HS10 is still your camera. It does quite well in this test. The detail it squelches is not all that visible in 8x10 prints, so you can do well with the bridge cam.

But if you are on the fence about what to buy, and you want all the other advantages of a dSLR, then don’t fret that the 720mm reach of the HS10 always crushes any dSLR you can buy at a reasonable price, because it doesn't. The modern version of this lens can be had with a refurb D5000 body for about 2.5 times the cost of the HS10. So if you are not heavily budget constrained, and if you think you want to have the higher performance of the D5000, then you can surely consider a combination like this.

It’s all about what *you* want. Obviously, both can do the job at typical print sizes and web sizes.

2 comments:

ropster said...

I think the image given here by the HR10 is far superior to the other 2 I dont realy understand your comments

Kim Letkeman said...

ropster: EVeryone who sees images like these focuses on different aspects and thus concludes something different. My perspective is that I strongly prefer the dSLR image because there is simply more detail in it ... that being details in the sections of the large branches that are wiped flat in the HS10 image by noise reduction. On the other hand, some people focus on the finer details in the HS10 image and decide that's what matters most, ignoring the smoothing that happens on all the thick branches and the trunk. Each to his own I suppose ... but I prefer to see less of the image smoothed into oblivion ... that is *always* the position I adopt.