Alrighty then … much has been made of the HS10’s long reach and there has been much debate about whether that reach can be effectively matched using a modern dSLR and a megazoom like the 18-200VR by Nikon. I happen to have the D300 and the first generation version of this lens, so I thought I’d pop by Henrys and take the HS10 and my kit out for a spin.
So, the particulars of this test:
Stand in one location and shoot something very far away at full zoom on the HS10. Shoot at least 3 images and take the sharpest one (and in all cases, there is always a sharpest one, even with the dSLR … stabilization is a *statistical* thing.) Then crop the image with the wider field of view to match. Finally, normalize all images to a specific size. I chose a 10” final image to which to normalize, which means 3000 pixels on the long edge. For a 3:2 image from a dSLR, that’s a 6mp image, and for a 3:4 image from a small sensor, that’s a 7.2mp image.
Finally, process all the crops for maximum detail and sharpness. Sorry, but the default settings of these cameras are different enough that comparing just the out of camera images is a total waste of effort. It says nothing of what can be achieved with a little skill.
This methodology gives the HS10 a significant advantage. First, I am reducing it image slightly, which helps alleviate some of its artifacts. Second. I am magnifying the other’s images dramatically. From 300mm to 720mm entails a magnification of 2.4x, which makes every artifact that much more visible. And. of course, for very fine details there are only so many pixels to go around when cropping that substantially. So the HS10 has every opportunity to spank the dSLR in this test.
Part 1 is a based on a shot of a house that is a few hundred feet away from where I was standing. There is a car in a garage and I want to try to read the fine print along the bottom of its license plate. Let’s see how we did.
The HS10 shot this image at 720mm.
The D300 shot this image at 300mm effective.
And from that image, I cropped this image (the section in the yellow surround above.)
And just for fun, I shot the same image with the F70EXR at 270mm.
And cropped this image.
Click through to see any of these at 800px.
Now, at these sizes, the dSLR and HS10 images are about the same. There is no reason to assume that the HS10 is going to produce a better or a worse image than the cropped dSLR at web sizes. Which means that if this sort of image is your target, then buy what you can afford and don;t spend much time worrying about it.
On the other hand, how about when we crop these images at 8x10size? And remember that you are viewing on a medium that expands everything 3 times, so in fact this view is the equivalent of printing at 30” on the long side.
Click through to see the crops, and be warned that they are fairly big.
So … what can we observe from the crops? Bearing in mind that this is like looking at a 30” print from 20” away :-)
- I can (barely) read the fine print on the Ontario license plate. But bear in mind that I know what it says, and that may help.
- I cannot quite read the license plate on the dSLR shot, but the letters look a little smoother.
- The HS10 shows some chromatic aberration around the license plate, the “PATHFINDER” lettering and indeed along the creases of the tailgate.
- The bricks have better contrast and are slightly sharper. More aggressive processing of the D300 image would probably make that a lot closer, but then the smoothness of the rest of the details would be compromised.
- The F70 image is surpsingly good when you consider the levels of magnification of a *5* megapixel image. And you can read the license plate with ease … the big letters at least :-)
Note that an actual 8x10 print is 1/3 this magnification so none of the minute difference would be visible. The shots would all look pretty decent, although the F70 image would lack some detail for sure. But if your goal is to see tiny details in 30” prints, then the HS10 wins this round. But the caveat is that you must process away the chromatic aberration at this magnification.
Fo all practical purposes this is a tie. For anyone shooting for the web, this is a tie. But I have other images with a lot more detail in them to check out. See part 2 for the branches test.