Pages: 1 [2]   Go Down

Author Topic: Sigma DP1 - Good enough for stock  (Read 13507 times)

professorgb

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 55
Sigma DP1 - Good enough for stock
« Reply #20 on: February 18, 2009, 10:35:26 am »

It looks like you may have posted the wrong URL.  This file shows some pretty obvious artificacts--it looks like a magic wand selection tool was quickly used and either the castle was desaturated or the sky was saturated.  My guess is the sky was saturated, as there's some banding/pixelation.

Do you have a less-processed version?  I'd be interested in seeing it.

Quote from: james_elliot
You are an eagle-eye viewer...
The image posted was the last I sent to Alamy and it was really over-sharpened to see what they would tell me. And the answer was: "soft or lacking definition" again.
In fact, I sent the same image with different parameters three times to see if anything would change...
If you want to look at the original just out of the raw converter that I sent first, it is there:
http://www.photo-lovers.org/portfolio/priv...xO_raw_copy.jpg

And if the sky is so blue, it's because I used a polarizing filter, with the sun exactly in the right position...
JMA
Logged

james_elliot

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 57
    • http://www.photo-lovers.org
Sigma DP1 - Good enough for stock
« Reply #21 on: February 18, 2009, 11:37:38 am »

No, I checked again and it is the output of DxO right out of the box, with saturation set to intermediate, and sharpening set to -0.5 (yes, less than 0, which is the default).
Nothing was selected, saturated or desaturated.


Quote from: professorgb
It looks like you may have posted the wrong URL.  This file shows some pretty obvious artificacts--it looks like a magic wand selection tool was quickly used and either the castle was desaturated or the sky was saturated.  My guess is the sky was saturated, as there's some banding/pixelation.

Do you have a less-processed version?  I'd be interested in seeing it.
Logged

250swb

  • Full Member
  • ***
  • Offline Offline
  • Posts: 220
Sigma DP1 - Good enough for stock
« Reply #22 on: February 18, 2009, 11:51:39 am »

Quote from: james_elliot
You are an eagle-eye viewer...
The image posted was the last I sent to Alamy and it was really over-sharpened to see what they would tell me. And the answer was: "soft or lacking definition" again.
In fact, I sent the same image with different parameters three times to see if anything would change...
If you want to look at the original just out of the raw converter that I sent first, it is there:
http://www.photo-lovers.org/portfolio/priv...xO_raw_copy.jpg

And if the sky is so blue, it's because I used a polarizing filter, with the sun exactly in the right position...
JMA

This image is soft and lacking definition, I think you could be confusing the high contrast with sharpness. If it hadn't have been 'soft and lacking definition' Alamy got you for it would instead be dirt in the sky, low down near the tower on the left. It may be a bird, but a buyer isn't to know that from the shape of the 'dirt', so it needs cloning out.

Producing a photo for a good stock agency (which Alamay is compared to some) is about making an image that is a ready for the buyer to work with, and while the buyer may want to saturate the image more than how you supply it, they can't easily desaturate this image without leaving behind the tell tale artifacts of high contrast and bold colour. The contrast is perhaps damaging any inherent sharpness and perhaps you could try a slightly flatter approach? Just an idea.

Steve


Steve
« Last Edit: February 18, 2009, 11:53:11 am by 250swb »
Logged

professorgb

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 55
Sigma DP1 - Good enough for stock
« Reply #23 on: February 18, 2009, 12:25:42 pm »

Just out of curiosity--I don't use DxO, but I'm considering it for use with my ultrawides.  On this image, did you use the standalone app, or did you use the Photoshop plugin?  Have you found an advantage either way?

Quote from: james_elliot
No, I checked again and it is the output of DxO right out of the box, with saturation set to intermediate, and sharpening set to -0.5 (yes, less than 0, which is the default).
Nothing was selected, saturated or desaturated.
Logged

james_elliot

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 57
    • http://www.photo-lovers.org
Sigma DP1 - Good enough for stock
« Reply #24 on: February 18, 2009, 02:06:19 pm »

It's the standalone application. I have also the photoshop plugin but I don't use it. The standalone app is much easier to use in a workflow. I usually create a 16 bits TIFF with DxO, without really processing the image except for sharpness.
I finish the job in photoshop regarding contrast, saturation, etc...
However there are drawbacks with DxO. For example, you should never use their perspective tool: the algorithm for resizing the image is terrible; photoshop does a much better job.
DxO is (according to me) very good for retrieving small details. I did one year ago a test of different raw conversions programs (regarding only small details retrieval) that can be found there:
http://www.photo-lovers.org/fconvert.html.fr
However I did again the test last month with the new versions of every program and some of them have been seriously enhanced. The new lightroom 2 for example is much better than lightroom 1. I will post that very soon on my website.

Quote from: professorgb
Just out of curiosity--I don't use DxO, but I'm considering it for use with my ultrawides.  On this image, did you use the standalone app, or did you use the Photoshop plugin?  Have you found an advantage either way?
Logged

John Clifford

  • Newbie
  • *
  • Offline Offline
  • Posts: 25
    • http://
Sigma DP1 - Good enough for stock
« Reply #25 on: February 23, 2009, 01:27:58 am »

Quote from: james_elliot
You are an eagle-eye viewer...
The image posted was the last I sent to Alamy and it was really over-sharpened to see what they would tell me. And the answer was: "soft or lacking definition" again.
In fact, I sent the same image with different parameters three times to see if anything would change...
If you want to look at the original just out of the raw converter that I sent first, it is there:
http://www.photo-lovers.org/portfolio/priv...xO_raw_copy.jpg

And if the sky is so blue, it's because I used a polarizing filter, with the sun exactly in the right position...
JMA

I have to tell you, that original "just out of the raw converter" image is also oversharpened; there is a grey halo around the ruins. In short, it's not a very sharp image, just as the first posted image wasn't very sharp. In fact, the only difference I can see is that the grey halo is much more distinct on the first "really oversharpened" image... but it's there in both images.
Logged
'Do you think a man can change his destiny?'
'I think a man does what he can until his destiny is revealed.'

james_elliot

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 57
    • http://www.photo-lovers.org
Sigma DP1 - Good enough for stock
« Reply #26 on: February 23, 2009, 11:48:53 am »

Quote from: John Clifford
I have to tell you, that original "just out of the raw converter" image is also oversharpened; there is a grey halo around the ruins. In short, it's not a very sharp image, just as the first posted image wasn't very sharp. In fact, the only difference I can see is that the grey halo is much more distinct on the first "really oversharpened" image... but it's there in both images.

I magnified and cropped a part of the image. By "halo", are you referring to the slightly whiter area that can be seen around the ruins?

JMA

[attachment=11704:crop.png]
Logged

Plekto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 551
Sigma DP1 - Good enough for stock
« Reply #27 on: February 23, 2009, 04:12:14 pm »

No, look at the tallest tower on the top right side.  See how the straight edge of the building's side has a couple of pixel wide gray line?   This is a sure sign of bad processing.  Now, this might be the camera itself being set up wrong or sigma's software, but it's there.

It's also noticeable on the leftmost tower - Some are real shadows, but it looks exactly like the effect from an area selection tool(ie - select sky "area" and apply changes).  You can plainly see that something is tracing the outline of the towers.
Logged

professorgb

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 55
Sigma DP1 - Good enough for stock
« Reply #28 on: February 23, 2009, 04:14:58 pm »

Actually, I think this is what he's talking about (100% crop, from a random part of the structure):

[attachment=11705:fragment.jpg]

This is what I was interpreting to be selection and enhancement within the image.  I'm pretty sure you've got the wrong file; the EXIF info shows that this was a TIFF file that was edited in CS2.  It's not straight out of DxO, nearest I can tell.

Quote from: james_elliot
I magnified and cropped a part of the image. By "halo", are you referring to the slightly whiter area that can be seen around the ruins?

JMA

[attachment=11704:crop.png]
Logged

james_elliot

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 57
    • http://www.photo-lovers.org
Sigma DP1 - Good enough for stock
« Reply #29 on: February 23, 2009, 08:03:21 pm »

Interesting story... To answer all posts in one:
-It was edited under photoshop to add the copyright notice and nothing else. It was not oversharpened.
-Conditions : Canon 1Ds MarkII on a tripod, Canon 24-70 f/2.8L, aperture f/8, speed 1/100, focal length 24mm, end of october, around 1PM solar time, very sunny day but not extremely warm (around 12 Celsius), no wind

Just to be sure, I decided to take the original RAW and to run it into another converter (Digital Photo Pro) with contrast and saturation set to normal and with sharpness set to 0 (and 0 is NEVER used by anyone; the least you use is 2, and most people on landscape shots use 5, or even more).

Guess what? The famous "halo" was still there (I was not exactly surprised by the way).

In order to make it more clearly visible, I loaded the DPP image into photoshop and modified the hue (and ONLY hue; no saturation, no contrast, no lightning, no sharpening).  With the proper angle of hue, the halo is extremely easy to see especially above the stones of the tower: a clear black/yellow line that separates the stone from the (now orange) sky.
It is less clear along the vertical wall of the tower, and even more difficult to see above the flowers on the left.
[attachment=11712:crop2.png]
So where does it come from? This is not oversharpening (difficult to believe with sharpening set to 0...), not saturation, not contrast and not demosaicing as it doesn't appear everywhere in the same way.

The answer is: it has nothing to do with sharpening; the air is warmer above the stones, and its refraction index is slightly different; it's just a reduced version of the hot road phenomenon.


Logged

Plekto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 551
Sigma DP1 - Good enough for stock
« Reply #30 on: February 24, 2009, 02:35:38 pm »

Not really, though.  If you look very carefully, it's everywhere in the photo.  It's just that it's easy to see in high contrast blue areas(appears to be affecting the blue channel).  It's also blurring the stones in certain areas as well.  If you haven't touched it, then the camera itself is doing something wrong.  Look at the leftmost tower.  It starts off very light brown/tan at the bottom and by the top it's clearly a dark gray.  The air would have to have a massive temperature difference to actually cause that much of a color change.

Ie - what you explained would result in a tinge of the same color over the area.  That the color of the line actually changes shows that something awful is happening between the sensor and the output.    If it is the camera, as it looks like it might be(we'd need other shots to verify), then yes, you do need to get a better camera.

What are the sharpening settings on the camera?  Are the levels correct?  I don't know what is causing this, if it's the camera - maybe it's bad settings.  But it doesn't look like it.
Logged

250swb

  • Full Member
  • ***
  • Offline Offline
  • Posts: 220
Sigma DP1 - Good enough for stock
« Reply #31 on: February 24, 2009, 05:21:51 pm »

Quote
Ie - what you explained would result in a tinge of the same color over the area. That the color of the line actually changes shows that something awful is happening between the sensor and the output. If it is the camera, as it looks like it might be(we'd need other shots to verify), then yes, you do need to get a better camera.

I suppose that depends on what you mean by a 'better' camera. Alamy have no problem accepting images from 6mp cameras, yet they are rejected from this Canon 1Ds MkII. So its technique, or a case of the camera showing inadequacies in the lens (out resolving it), or the camera itself is faulty. So for a 'better' camera you could say the Sigma DP1 from the start of the thread is at least no worse in that it gets thrown out for the same problems, but at least its cheaper :-)

Steve

denisv

  • Newbie
  • *
  • Offline Offline
  • Posts: 2
Sigma DP1 - Good enough for stock
« Reply #32 on: March 04, 2009, 04:07:56 am »

Quote from: Plekto
Unfortunately, they are right.  The Sigma/Foveon sensors are in need of a serious upgrade.  They really put out about 4-5MP and that's just not going to compete with the 12-25mp crowd which can literally throw away pixels to smooth out the image and still show more detail than the DP1.

Check out the Fuji S5 Pro - its sensor technology produces very clean results like the DP1.  Very clean, smooth, and good contrast(on-sensor HDR bracket and blending/noise reduction).  Not the highest resolution option out there, but it's very easy to shoot with and get back good results.  Especially outdoors.
I think you're mistaken.

Compare Foveon and S5 pro
http://www.sigma-dp.com/DP2/sample-photo/i...igmaDP2-011.jpg
http://a.img-dpreview.com/gallery/fujis5pr...ls/dscf0365.jpg
Logged

Plekto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 551
Sigma DP1 - Good enough for stock
« Reply #33 on: March 04, 2009, 03:54:40 pm »

Quote from: denisv
I think you're mistaken.

Compare Foveon and S5 pro
http://www.sigma-dp.com/DP2/sample-photo/i...igmaDP2-011.jpg
http://a.img-dpreview.com/gallery/fujis5pr...ls/dscf0365.jpg

The S5 looks much better to me, so I don't see your point.
Logged

denisv

  • Newbie
  • *
  • Offline Offline
  • Posts: 2
Sigma DP1 - Good enough for stock
« Reply #34 on: March 12, 2009, 08:17:16 am »

Quote from: Plekto
The S5 looks much better to me, so I don't see your point.

By "much better" you surely mean "much softer"?
Logged

Plekto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 551
Sigma DP1 - Good enough for stock
« Reply #35 on: March 12, 2009, 06:17:26 pm »

Quote from: denisv
By "much better" you surely mean "much softer"?

The man's beard and his pimples are well defined on the Sigma.   For so few pixel locations, it's doing a fantastic job.  I don't see many moires or patterns, either. (though the color is blown out on his shirt - it's not as accurate with colors IMO, or as wide a dynamic range)

The second one has better colors and dynamic range, but it's not really any higher resolution that I can see.  It does seem to be doing a LOT more NR than the Foveon sensor, which is understandable as it's really a modified Bayer pattern.  

The Sigma seems sharper but it doesn't look as realistic to me.  The Sigma is gorgeous, but the dynamic range and colors of the S5 look closer to good film and less "digital".

I suspect with the AA filter removed, though, the S5 would jump at you.
Logged
Pages: 1 [2]   Go Up