Reading mode: Light Dark

Bayer yecch

Go to last post
M

Matthew Cromer

Veteran Member
Messages
5,079
Reaction score
0
Location
US
I'm gonna rant a bit here.

I shoot a lot of landscape with my Sony 717 and generally like it and the results. But it's fall foliage time where I live right now, and I've just finished processing the images. For red leaves especially, Bayer just looks like cr*p at 100%. And I find myself pining for the pixel-color accurate world of Foveon, where a pixel is a pixel instead of a guess.

My next camera is probably going to be a Sony 828, as it has the form factor and handling I need to get into the artistic "flow". But I've got my eye on the Foveon sensor and if a higher-megapixel model comes out next year I'm almost certainly going to pick up one because I just LOVE those true-color images. To get that kind of fidelity from Bayer you have to downsample to 50% in each direction, for 1/4 the number of pixels.
--
my favorite work: http://www.pbase.com/sdaconsulting/featured_art&page=1
 
Matthew Cromer said:
I'm gonna rant a bit here.

I shoot a lot of landscape with my Sony 717 and generally like it
and the results. But it's fall foliage time where I live right
now, and I've just finished processing the images. For red leaves
especially, Bayer just looks like cr*p at 100%. And I find myself
pining for the pixel-color accurate world of Foveon, where a pixel
is a pixel instead of a guess.

My next camera is probably going to be a Sony 828, as it has the
form factor and handling I need to get into the artistic "flow".
Click to expand...
I was a lot like you.... When I went from SLR to PnS style cameras it was like being cut loose from a ball and chain. I like taking a camera everywhere I go and its a lot harder with an SLR.

Do you really want to keep shooting with bayer cameras wile looking at others foveon images? Its a lot more fun making your own.

The switch back to SLR has its drawbacks... the SD9 is a monsterous camera compared to my tiny point -n- shoots. But I'm not going back!
--
Obscura
Join the GREY CARD ARMY!
 
I told you you should have gotten an sd9 for your landscape work but you said back then that the colors weren't accurate, nice to see you've seen the light.
Matthew Cromer said:
I'm gonna rant a bit here.

I shoot a lot of landscape with my Sony 717 and generally like it
and the results. But it's fall foliage time where I live right
now, and I've just finished processing the images. For red leaves
especially, Bayer just looks like cr*p at 100%. And I find myself
pining for the pixel-color accurate world of Foveon, where a pixel
is a pixel instead of a guess.

My next camera is probably going to be a Sony 828, as it has the
form factor and handling I need to get into the artistic "flow".
But I've got my eye on the Foveon sensor and if a higher-megapixel
model comes out next year I'm almost certainly going to pick up one
because I just LOVE those true-color images. To get that kind of
fidelity from Bayer you have to downsample to 50% in each
direction, for 1/4 the number of pixels.
--
my favorite work:
http://www.pbase.com/sdaconsulting/featured_art&page=1
Click to expand...
 
Mathew is and was right though. The Foveon produces great resolution, but has a hard time getting the colors correct. Of course I'm basing this on the SD9, not tne SD10 which remains to be proven.

I'd love foveon to become the standard for digital imaging to come when it is ready (but I want mine in a Canon DSLR). Bayer is the standard right now.
Ogilvy said:
I told you you should have gotten an sd9 for your landscape work
but you said back then that the colors weren't accurate, nice to
see you've seen the light.
Click to expand...
--



http://www.pbase.com/domotang
 
Stephen,

I appreciate your note, but this kind of comment just cannot stand.
Stephen Reed said:
The Foveon produces great
resolution, but has a hard time getting the colors correct. Of
course I'm basing this on the SD9, not tne SD10 which remains to be
proven.
Click to expand...
The correct statement is that some users have a hard time getting the colors correct. There are few images which back up your statement, even including the tortuous efforts for some to dissect this problem. There are thousands supporting mine.

The SD9 gets great color results. No one ever said it was always easy to learn how to do this, but it is absolutely possible.

--
Laurence Φ€ 08 LL

http://www.pbase.com/lmatson/sd9_images
http://www.pbase.com/sigmasd9/root
http://www.pbase.com/cameras/sigma/sd9
http://www.beachbriss.com (eternal test site)
 
It's the pixel-level color resolution that I am referring to, not the exact color match. No Bayer camera can come close to the pixel-level accuracy of colors the Foveon can.

As for the exact tonalities, yes the SD9 has had a bit of an issue with that, but I suspect these are software issues to a large extent or can be worked around.

-
my favorite work: http://www.pbase.com/sdaconsulting/featured_art&page=1
 
I know the colors “look” good, but some come strait from the imager inaccurate under certain circumstances. There is a difference between colors “looking good” and being accurate. At this stage this is not a software issue, although software can help cover up the artifacts.

If these were not issues, I can assure you that Nikon and Olympus (possibly Canon) would have chosen the Foveon over the traditional Bayer imager (and they DO have the ability to license the Foveon technology should they choose). We haven’t even witnessed any firms considering it for astronomy imaging applications yet.

I can’t concure with you on this one, the problems with the Foveon are too obvious at this stage. On a side note I do believe that the Foveon is the next standard for imaging devices when it’s issues which go unseen (or merely denied) by some are corrected.
Laurence Matson said:
I appreciate your note, but this kind of comment just cannot stand.
Stephen Reed said:
The Foveon produces great
resolution, but has a hard time getting the colors correct. Of
course I'm basing this on the SD9, not tne SD10 which remains to be
proven.
Click to expand...
The correct statement is that some users have a hard time getting
the colors correct. There are few images which back up your
statement, even including the tortuous efforts for some to dissect
this problem. There are thousands supporting mine.

The SD9 gets great color results. No one ever said it was always
easy to learn how to do this, but it is absolutely possible.

--
Laurence Φ€ 08 LL

http://www.pbase.com/lmatson/sd9_images
http://www.pbase.com/sigmasd9/root
http://www.pbase.com/cameras/sigma/sd9
http://www.beachbriss.com (eternal test site)
Click to expand...
--



http://www.pbase.com/domotang
 
Stephen Reed said:
I know the colors “look” good, but some come strait from the imager
inaccurate under certain circumstances. There is a difference
between colors “looking good” and being accurate. At this stage
this is not a software issue, although software can help cover up
the artifacts.
Click to expand...
Please spend some time understanding how the SD9 (and SD10) work. No colors come "strai(gh)t from the imager". These cameras demand a Raw workflow, and hence the interaction of a user. If you mean that "accurate" colors do not result from the user CHOOSING to allow the software to use "auto" settings to come up with a result, it is still the responsibility of the user to make the decision that "auto" is what is wanted.
Inaccurate color is the fault of the user.
Stephen Reed said:
If these were not issues, I can assure you that Nikon and Oly
(possibly Canon) would have chosen the Foveon over the traditional
Bayer imager (and they DO have the ability to license the Foveon
technology should they choose).
Click to expand...
Whatever the reason these manufacturers have not chosen Foveon, it is not because of any inability of the sensor to record color.
Stephen Reed said:
We haven’t even witnessed any firms
considering it for astronomy imaging applications yet.
Click to expand...
Astronomy imaging does not seem to be the major priority for most photographers, nor for manufacturers involved in the production of consumer DSLR cameras. The SD10 should definitely be an improvement over the SD9 for this kind of work though.
Stephen Reed said:
I can’t concure with you on this one, the problems with the Foveon
are too obvious at this stage.
Click to expand...
I must be stupid; they are are not too obvious to me.
 
That's really insulting to call me a liar. Please desist.

I was talking about pixel-level color fidelity. It seems obvious that Bayer sensors are not going to have pixel-level color resolution since they only sample red and blue at 1/4 of the photosites. By downsampling 50% color resolution should be very similar to a native foveon sensor.

Obviously overall resolution is a different matter. But I care a lot about color resolution for my landscape images as well as overall resolution. The fine color resolution really makes a difference for getting foliage shots to look natural.

--
my favorite work: http://www.pbase.com/sdaconsulting/featured_art&page=1
 
Laurence Matson said:
I appreciate your note, but this kind of comment just cannot stand.
Stephen Reed said:
The Foveon produces great
resolution, but has a hard time getting the colors correct. Of
course I'm basing this on the SD9, not tne SD10 which remains to be
proven.
Click to expand...
The correct statement is that some users have a hard time getting
the colors correct. There are few images which back up your
statement, even including the tortuous efforts for some to dissect
this problem. There are thousands supporting mine.
Click to expand...
I see Sigma photos with lousy colors all the time. Quasi-constantly.

Look at this image of Byrce Canyon from a Nikon D100:

http://forums.dpreview.com/forums/read.asp?forum=1021&message=6489995

And then go find that picture of Bryce Canyon in the Sigma pbase.com
pages ... the one where Bryce Canyon is all yellow.

Or go back to Rick Decker's "SD9 vs. 10D image" post from a few days
ago. In the 10D image the old car is black, with a standard black
paint job from the factory. In the SD9 photo the car is dark purple.

Or that yellow house in Normandy you and I discussed 2 weeks
ago.

I am really not looking to come into the Sigma forum and bash the
SD9, but when you say there's no "color problem", I really have
trouble stopping myself from speaking up.
Laurence Matson said:
The SD9 gets great color results. No one ever said it was always
easy to learn how to do this, but it is absolutely possible.
Click to expand...
The Nikon D100 and Canon 10D both can also give great color
results ... and there's very little to learn.

===
Sorry, Laurence ... I don't want to argue. And if you love the SD9
that's great. But we disagree about the color issue.
 
Stephen Waits said:
Matthew Cromer said:
That's really insulting to call me a liar. Please desist.
Click to expand...
Fine, I apologize and retract - I will still say that you are
walking a fine line.
Click to expand...
Indeed, a fine line. If you look at his original statement that you took issue with, "To get that kind of fidelity from Bayer you have to downsample to 50% in each direction, for 1/4 the number of pixels." you can perhaps see a sense in which it is not a lie, not an opinion, but actually some kind of truth.

If by "that kind of fidelity" he meant freedom from the kinds of artifacts that come from undersampling red and blue, or he meant the look that you get from sampling all three color planes at the same pixel pitch as the output file, then his statement would be literally true, would it not?

You jumped to the conclusion: "This implies that a 3Mp X3 sensor is equivalent to a 12Mp CFA/'Bayer' sensor which is a lie."

So, if he had claimed that kind of equivalence, I would agree with you that it is incorrect, or an insufficiently supportable opinion, or maybe even a lie if I thought he knew better. But he didn't say the thing that you said was untrue; that was you putting words in his mouth.

Walking fine lines is a specialty of mine, as you've probably noticed. My point is to pay close attention to what people are saying, and to try to understand what's behind it, whether facts, opinions, distortions, or whatever. If you think some of what I say are opinions and distortions, too, you could be right, and I'd like to hear about it.

j
 
Brian Slater said:
Please spend some time understanding how the SD9 (and SD10) work.
No colors come "strai(gh)t from the imager". These cameras demand a
Raw workflow, and hence the interaction of a user. If you mean that
"accurate" colors do not result from the user CHOOSING to allow the
software to use "auto" settings to come up with a result, it is
still the responsibility of the user to make the decision that
"auto" is what is wanted.
Inaccurate color is the fault of the user.
Click to expand...
So when the folks at Imaging Resource write:

"Color rendering has also seen substantial improvement, to the point
that I'd now rate the SD10's color as 'very good to excellent,'
something I'd not have said about the SD9"

that just goes to show that they don't fully understand the issue
either ?

I can't really be sure, but I suspect the "auto" levels the Sigma
software comes up with are as much as a problem as the camera
itself. And then the users try to do better, and make it even worse.
 
David,

Your saying that the camera does not put out good color just requires one image with good color to refute it. I am pretty sure you could find one.

And it is good you mention that "yellow" house, because after correcting it, it looked better in your own words.

http://forums.dpreview.com/forums/read.asp?forum=1000&message=6393136

The Bryce Canyon shot, I would trust, having been there, knowing the guy who took that shot, and knowing the vagaries of early sunlight.

I stand by what I have said often: You can get good color out of an SD9 consistently. When it is not there, I would check on what I had done wrong first. Also, I usually trust the operator who was there.

The two images of Bryce from the D100 seem quite washed out if my memory serves me, but that may have been the light that day. I plan to go there in a few months myself.

I took quite a few shots that got dumped on for color. That is because few people know anything about the Föhn. Those that did said right away that they were right.

I also took a pile in France, Spain and Andorra. Those colors are accurate for that light.

I am glad the D100 and 10D are so easy to use. I think the SD9 is too, but I have used it.

This reminds me again of this passage:

http://forums.dpreview.com/forums/read.asp?forum=1027&message=6492803

Nice book too.

--
Laurence Φ€ 08 LL

http://www.pbase.com/lmatson/sd9_images
http://www.pbase.com/sigmasd9/root
http://www.pbase.com/cameras/sigma/sd9
http://www.beachbriss.com (eternal test site)
 
David,

Now you seem to be agreeing with me:
David Mazeau said:
I can't really be sure, but I suspect the "auto" levels the Sigma
software comes up with are as much as a problem as the camera
itself. And then the users try to do better, and make it even worse.
Click to expand...
This snap review on Dave's part was not a best effort in many respects. To say that the color is better, is his opinion, but it does not say it was wrong in the SD9, at least if we are to take into consideration what you wrote above.

--
Laurence Φ€ 08 LL

http://www.pbase.com/lmatson/sd9_images
http://www.pbase.com/sigmasd9/root
http://www.pbase.com/cameras/sigma/sd9
http://www.beachbriss.com (eternal test site)
 
David Mazeau said:
Brian Slater said:
Please spend some time understanding how the SD9 (and SD10) work.
No colors come "strai(gh)t from the imager". These cameras demand a
Raw workflow, and hence the interaction of a user. If you mean that
"accurate" colors do not result from the user CHOOSING to allow the
software to use "auto" settings to come up with a result, it is
still the responsibility of the user to make the decision that
"auto" is what is wanted.
Inaccurate color is the fault of the user.
Click to expand...
So when the folks at Imaging Resource write:

"Color rendering has also seen substantial improvement, to the point
that I'd now rate the SD10's color as 'very good to excellent,'
something I'd not have said about the SD9"

that just goes to show that they don't fully understand the issue
either ?
Click to expand...
Two points. Firstly, Imaging Resource has already published unreliable information about the SD10. Secondly, I did not say that they don't fully understand anything. On the other hand, I do not know the details of how IR reached their conclusion, whether they used custom white balance, whether they used auto in SPP etc. My point was in reply to the statement that "this is not a software issue".
David Mazeau said:
I can't really be sure, but I suspect the "auto" levels the Sigma
software comes up with are as much as a problem as the camera
itself. And then the users try to do better, and make it even worse.
Click to expand...
Users would not do worse than auto if they can use a mouse and have two functioning eyes.
 
Laurence Matson said:
And it is good you mention that "yellow" house, because after
correcting it, it looked better in your own words.

http://forums.dpreview.com/forums/read.asp?forum=1000&message=6393136
Click to expand...
Well... my saying "better" just goes to show that I don't really have
much interest in going after these issues for too, too long. We
can just disagree a bit, and then leave it for anybody that is
interested to click on the links to the images, and look at the images
themselves. And form their own opinions.
Laurence Matson said:
The Bryce Canyon shot, I would trust, having been there, knowing
the guy who took that shot, and knowing the vagaries of early
sunlight.

I stand by what I have said often: You can get good color out of an
SD9 consistently. When it is not there, I would check on what I had
done wrong first. Also, I usually trust the operator who was there.

The two images of Bryce from the D100 seem quite washed out if my
memory serves me, but that may have been the light that day.
Click to expand...
That's probably why it is best to not argue about this too long. I
can X, and you can say Y ... but it's pretty hard to get anything
definite.

But others can just look at the images themselves, and form their
own opinions.

Here's the SD9 Bryce Canyon image we are talking about.

http://www.pbase.com/image/14165229
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top