Wednesday, May 12, 2010

HS10 versus S100fs versus D5000 – Cropfest (Mushfest?) 2010

*Note* I’ve seen people already lambasting imaging-resource for using the original firmware that ships with the camera for their tests. After all, the blur was removed by that firmware. Now here’s my problem with all the apologists pounding on that point: the issue was specifically isolated to the use of dynamic range processing, not to the use of higher ISO. *It should have no effect on this test.* And note that I shot all of my tests on the V1 firmware. Including my ISO ladder. What’s good for the goose …

I tend to like the images at http://imaging-resource.com because they have become quite consistent over the last few years, and because they have the mannequin series, which tends to expose heavy noise reduction better than any other series of images I have seen. The hair on the dummy looks terrific on cams that don;t wipe detail with abandon and horrid on cams that do.

So it was with great anticipation that I went to the site when I read that they had shot some images with the HS10. Of course, it turns out that they did not shoot the mannequin series, so I will have to make do with the bright light still life series. Not bad, though. You can still get a good idea if the cam has the chops.

I used these three images:

HS10 at 1600 ISO
S100fs at 1600ISO
D5000 at 1600ISO

I created a series of crops without touching the images. These are presumably straight from the camera.

mushfest_2010[1]

Remember to click on the image to open it in a separate tab or window. And you may need to click a second time to open it to 100%.

So what do I see?

  1. The HS10 is blurred and mushy throughout. I am very surprised at its lack of detail.
  2. There is discoloration in crop 1 at the edge of the inner fabric swatch. What’s up with that? This cam may have more serious dynamic range issues than the size of its sensor would predict. Perhaps isolated to the blue channel.
  3. Edge integrity is a problem with the HS10. It’s actually quite good with the S100fs. And the D5000 is close to perfect of course.
  4. Where did all the low contrast detail go in crop 3 (fiddler)? The HS10 has wiped it clean. The S100fs makes a reasonable go of it.

No need to go on … there really is no contest. The S100fs crushes the HS10 for detail retention at this focal length, and the D5000 does the same to the S100fs.

The laws of physics prevail yet one more time …

4 comments:

Coolsiggy said...

At 1600iso this is not surprising, at 100iso the S100fs would improve a fair bit...then it's more down to the glass.

Kim Letkeman said...

Coolsiggy: Sorry, but 100ISO in a studio setting is not interesting at all. 1600ISO tests the ability of the camera to function indoors and in very early and very late light (the light that professionals call "sweet") ... I have no interest in checking how well the S100fs shoots in blazing sunlight. I think we all know the answer to that.

Anonymous said...

That firmware did a lot more then to fix blurring at elevated DR. I shoot at normal 100DR and have noticed a marked improvement in JPG sharpness in both indoor and outdoor shots, wideangle as well as zoomed in. Also, FUJI has already been shipping the HS10 with the new firmware. So for you to do this review without the new firmware is pointless.

Kim Letkeman said...

SPFH: Hardly pointless. I shot all my own tests with a V1 f/w and my images were sharp. The big announcement from Dave Lloyd was that he found the difference to be based upon the use of DR200. Everyone bought into that. The fact that there was no proper test to isolate the exact issue is simply typical of that side of the Fuji forum. But I saw pretty good detail retention all the way up to 3200ISO (have you seen my ISO ladder?) So what am I to believe? It's all anecdotal until someone posts a proper v1 versus v1.0.1 comparison. I will, however, agree that it would be nice for imaging-resource to reshoot with the new firmware to remove the doubt that some will have.