Reading mode: Light Dark

New SD9 Images at Chasseur d'Images ?

Go to last post
Not to fuel a fire but each of the new $2K DSLRs has had issues.. D60 and AF along with the need to develop a good Photoshop post processing work flow ( not anegative jsut a feature ), the D100 and it's Coolpix digital "look" along with very soft out of camera jpgs ( just to name 1 design feature that isn't perfect ), the S2 and moire issues at times, and the SD9 with a few of it's on. Trick is to pick the camera which best fits YOUR individual needs and enjoy. This is why owners at times defend their camera to the end but forget it may NOT be the best camera for someone else. They will all take great shots given the understanding of the camera's limitations.. just some will excel in one area or another. it's been a great year... maybe that's why there's so much debate on the SD9 and maybe also because the hopes/expectations were so high. Not to mention the fact many just spent $2000 on a camera body that perhaps isn't making great out of camera pictures and then they see Phil's sample SD9 shots and go WOW.....not bad for using "crappy glass" and a cheap DSLR body.... hmmmmm now about the NEED for L glass :)

PS Crappy glass was a bit if sarcasm on my part due to the sometimes snobbish behavior people demonstrate when talking about Sigma/Tokina/Tamron, etc 3rd party glass :) Remember, not everyone has $1500 for a single lens but some still have the love of photography and simply do not want a P&S camera! Maybe the SD9 will fit their needs perfectly!
Erik Magnuson said:
Joe Rivera said:
Some of the photos are just stunning, and the sharpness and clarity
of color boggles my mind. It makes me wonder how a camera can span
performance from that kind of quality to seemingly shocking
glitches.
Click to expand...
Just for a minute, let's assume that the observed problems are not
fixable before production. So you have a camera that produces
stunning shots most of the time but has a few obnoxious quirks.
What do you do? Wait until it's perfect? (but you are already
behind schedule.) Release it anyway and trust the users to
workaround the limitations?

To be honest, I think I could live with "blown-out red light
problem" and the CA enhancement, and even the noise profile. There
are some other reasons why the SD-9 is probably not for me, but
even though I pick on them, these are not product-fatal flaws. Of
course, expectations were so high, any flaws may seem that much
more disappointing. I agree that they could be totally unacceptable
to some people. No camera (or technology) is perfect -- X3 and the
SD-9 are different enough that it will take a while to learn what
the real limitations are.
Member said:
Is it possible we are looking at photos from camera units that were
not meant to see the public light of day?
Click to expand...
It's not likely.

--
Erik
Click to expand...
 
Member said:
Member said:
I don't know what
caused the ugly artifacts on the racing car picture, could it be a
result of too much tweaking of the parameters in the RAW converter?
Click to expand...
You might be able to argue "not enough" tweaking. Both me and
Babelfish think that they used the "default" parameters for
everything.
Click to expand...
"Toutes les photos ont été "développées" avec le logiciel Sigma Photo Pro sans aucune modification des paramètres standard, avec enregistrement demandé en Jpeg."

All the photos have been "develolped" using the Sigma Photo Pro program without any modification to the default parameters, output to Jpeg.
 
Dolan Halbrook said:
"Toutes les photos ont été "développées" avec le logiciel Sigma
Photo Pro sans aucune modification des paramètres standard, avec
enregistrement demandé en Jpeg."

All the photos have been "develolped" using the Sigma Photo Pro
program without any modification to the default parameters, output
to Jpeg.
Click to expand...
Thanks for clearing that up, I cannot read french!

Well, if these "overflow artifacts" cannot be tweaked/adjusted to the better with the RAW converter, then (hopefully) there are other ways around to eliminate or minimize the errors. Perhaps it will turn out better at ISO 200 or 400? Perhaps a faster shutter speed, and then adjust the exposure in the RAW converter? I don't know, but I hope Phil will bring up this issue in his review, he should be able to test and see if there are reasonable work arounds to this problem.

Geir Rune
 

Ladehaug said:
The "moire" (?) or artifact in this crop is most probably amplified
by a "line shift" since photo of the resolution chart is not
straight (not aligned horizontally).
Click to expand...
I suspect it will always be prone towards that aliasing artifact, as those lines are converging arcs; they'll never be straight.
 
Peter G said:
Well,

This crops up again here in a daylit sky at ISO 400, with 1/1500
exposure. So I find myself in dissagreement with those who say
noise is not an issue.

http://www.numeritest.com/Foveon/Images/IMG04913.jpg
Click to expand...
I am not sure what horizontal pattern noise you see. I have inspected the image, and I can't see any patterns in the noise. I only see evenly distributed noise. The noise also seems to be more correct chromatically, which is not always the case with bayer images.

I want to state my opinion and disagree with you, as I think the image you are referring to is pretty good to be a ISO 400 image. I have found that the noise is very small grained (high resolution) compared to what comes from a bayer sensor, making it very simple to manipulate in Photoshop (that is; if someone should have a desire to remove or minimize the noise, which can be a good idea before downsampling/upsampling the image).

Two examples:





To the left are original crops, and to the right is the manipulated version. Both magnified 200%. What I did was simply applying a Gaussian blur (radius 0.5), followed by a USM 0.5, 150%. These parameters can be fine tuned to whatever you want. The point is, there is no way I can do this with a bayer image, at least not with my ISO 100/200 pictures from my Canon S40 (I should not even tell that there is a ISO400 on the S40.. it is extremly noisy!). The ISO noise in a typical bayer image contains larger dots (blurred), making it almost impossible to remove or minimize with the method I just demonstrated.

Just my opinion!

Geir Rune
 
Michael Long said:

Michael Long said:
The "moire" (?) or artifact in this crop is most probably amplified
by a "line shift" since photo of the resolution chart is not
straight (not aligned horizontally).
Click to expand...
I suspect it will always be prone towards that aliasing artifact,
as those lines are converging arcs; they'll never be straight.
Click to expand...
Yes, I am pretty sure this will be the case. Not much to do about it, unless an AA filter is applied to the camera, which will only result in a more blurred image. Well, I think we'll need an optimized 15-20Mp Foveon sensor (at least) to make things closer to perfection. A fantasy today, but who knows what will be in 5-10 years :-)
 
Its a fairly large grained banding, obviously more subtle than in low light shots. It is more obvious in the sky upper left quarter. If you don't see it try rotating the image 90 degrees and it may become more evident. This is still pre-production, so its no biggie.

I only see it in the specific shot and only in the sky:
http://www.numeritest.com/Foveon/Images/IMG04913.jpg
Ladehaug said:
Peter G said:
Well,

This crops up again here in a daylit sky at ISO 400, with 1/1500
exposure. So I find myself in dissagreement with those who say
noise is not an issue.

http://www.numeritest.com/Foveon/Images/IMG04913.jpg
Click to expand...
I am not sure what horizontal pattern noise you see. I have
inspected the image, and I can't see any patterns in the noise. I
only see evenly distributed noise. The noise also seems to be more
correct chromatically, which is not always the case with bayer
images.

I want to state my opinion and disagree with you, as I think the
image you are referring to is pretty good to be a ISO 400 image. I
have found that the noise is very small grained (high resolution)
compared to what comes from a bayer sensor, making it very simple
to manipulate in Photoshop (that is; if someone should have a
desire to remove or minimize the noise, which can be a good idea
before downsampling/upsampling the image).

Two examples:





To the left are original crops, and to the right is the manipulated
version. Both magnified 200%. What I did was simply applying a
Gaussian blur (radius 0.5), followed by a USM 0.5, 150%. These
parameters can be fine tuned to whatever you want. The point is,
there is no way I can do this with a bayer image, at least not with
my ISO 100/200 pictures from my Canon S40 (I should not even tell
that there is a ISO400 on the S40.. it is extremly noisy!). The ISO
noise in a typical bayer image contains larger dots (blurred),
making it almost impossible to remove or minimize with the method I
just demonstrated.

Just my opinion!

Geir Rune
Click to expand...
 
Ladehaug said:
To the left are original crops, and to the right is the manipulated
version. Both magnified 200%. What I did was simply applying a
Gaussian blur (radius 0.5), followed by a USM 0.5, 150%. These
parameters can be fine tuned to whatever you want. The point is,
there is no way I can do this with a bayer image, at least not with
my ISO 100/200 pictures from my Canon S40 (I should not even tell
that there is a ISO400 on the S40.. it is extremly noisy!). The ISO
noise in a typical bayer image contains larger dots (blurred),
making it almost impossible to remove or minimize with the method I
just demonstrated.
Click to expand...
I think what you (and many others that have only uses cameras with the very small sensors) are confusing at least some of issues related to the size of the sensor with the Bayer issue. There is a huge difference in the size of the sensors between the "fixed lens" cameras and the DSLRs with intechangable lenses.

Your S40 retailed for about $800 a year ago (probably worth about $500 today). It uses a sensor with pixels about 10 times smaller in area than a D60/D100/S2, the DSLR bodies in the $2,200 price range that are the more direct competitors to the $1,800 SD9 body. Even a 2 year old D30 will blow the doors off the S40 or any of the other P&S cameras today in terms of lower noise and overall image quality.

What I really object to in your statement was "The ISO noise in a typical bayer image contains larger dots." What is it that you are calling the "typical bayer?" As the diagram (taken from the old D30 review) shows, a DSLR sensor is about 10 times the AREA of a Points and Shoot sensor. This larger area translates into capturing more light per pixel and thus lower noise. Note the small purple area in the center, that is the size of the typical P&S sensor and the Red area is the size of a DSLR sensor.



--
Karl
 
Peter G said:
Its a fairly large grained banding, obviously more subtle than in
low light shots. It is more obvious in the sky upper left quarter.
If you don't see it try rotating the image 90 degrees and it may
become more evident. This is still pre-production, so its no biggie.
Click to expand...
I cannot see this banding on the original (on my monitor). However, after adjusting the contrast, some slight banding does indeed show up.



The banding seems to appear all over the sky. I also inspected the channels, and the banding is most visible in the red channel, also appearing in the green channel, but not visible in the blue channel.

Speaking for myself, this doesn't bother me since I had to tweak the image to the limits in order to see it.

Geir Rune
 
Stuffy said:
Your technical justification for moire may be correct, but since
the sensor is billed as being less susceptible to moire (remember
the photo of the tweed cloth?), I'm left wondering what the real
story is.
Click to expand...
It's less susceptible to color moire. There's nothing you can do with a sensor with square pixels to make it less susceptible to B&W moire, apart from adding an AA filter. However, most people would rather have occasional moire than have softened pictures 100% of the time.

IMO, with Bayer pattern sensors, when most people talk about moire, they're referring to color moire because it's more frequent, more noticeable, and more distracting. It's also harder to deal with in software.

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/
 
I didn't do anything, it was clear as soon as I brought the image up, my monitor is fairly high contrast though.

Also I may be attuned to this phenomena.
Ladehaug said:
Peter G said:
Its a fairly large grained banding, obviously more subtle than in
low light shots. It is more obvious in the sky upper left quarter.
If you don't see it try rotating the image 90 degrees and it may
become more evident. This is still pre-production, so its no biggie.
Click to expand...
I cannot see this banding on the original (on my monitor). However,
after adjusting the contrast, some slight banding does indeed show
up.



The banding seems to appear all over the sky. I also inspected the
channels, and the banding is most visible in the red channel, also
appearing in the green channel, but not visible in the blue channel.

Speaking for myself, this doesn't bother me since I had to tweak
the image to the limits in order to see it.

Geir Rune
Click to expand...
 
Ron Parr said:
IMO, with Bayer pattern sensors, when most people talk about moire,
they're referring to color moire because it's more frequent, more
noticeable, and more distracting. It's also harder to deal with in
software.
Click to expand...
Harder to deal with in software compared to what? Monochrome moire? Chromatic aberration? There are pretty good filters for both color moire and CA, if applied selectively. I'm not sure why you think one is harder than the other - they are pretty similar.

--
Erik
 
Erik Magnuson said:
Ron Parr said:
IMO, with Bayer pattern sensors, when most people talk about moire,
they're referring to color moire because it's more frequent, more
noticeable, and more distracting. It's also harder to deal with in
software.
Click to expand...
Harder to deal with in software compared to what? Monochrome moire?
Chromatic aberration? There are pretty good filters for both color
moire and CA, if applied selectively. I'm not sure why you think
one is harder than the other - they are pretty similar.
Click to expand...
It's realtively easy to blur away B&W moire b/c the result of blurring makes a direct tradeoff of sharpness for moire.

The problem with color moire is that it contains color artifacts resulting from the lack of data on the true colors of the pixels. You can try blurring or other tricks, but you just can't avoid getting the colors wrong. The software false color filter for the D30/D60 helps a little, but doesn't really fix the problem.

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/
 
Yeah, Babelfish isn't bad, but i've seen it hose quite a few translations, so I thought i'd try to help out. Next time I'll see how the fish does and try to correct that.
JKP said:
Ladehaug said:
Thanks for clearing that up, I cannot read french!
Click to expand...
Copy the link, go to http://www.altavista.com , select Tools...translate and
paste the link into 'Translate a Web page'. You can choose from
French, Chinese, Japanese, German and other languages.

Gives usually good enough translation so one can figure out what
they are
talking about.

-jkp-
Click to expand...
--
--- http://www.reneeanddolan.net ---
 
Karl Guttag said:
I think what you (and many others that have only uses cameras with
the very small sensors) are confusing at least some of issues
related to the size of the sensor with the Bayer issue. There is
a huge difference in the size of the sensors between the "fixed
lens" cameras and the DSLRs with intechangable lenses.
Your S40 retailed for about $800 a year ago (probably worth about
$500 today). It uses a sensor with pixels about 10 times smaller
in area than a D60/D100/S2, the DSLR bodies in the $2,200 price
range that are the more direct competitors to the $1,800 SD9 body.
Even a 2 year old D30 will blow the doors off the S40 or any of the
other P&S cameras today in terms of lower noise and overall image
quality.
Click to expand...
Well, this information is not new to me. I am well aware of the sensor sizes used in various cameras and what their pros and cons are. May be I owe an outdated S40, but what this camera is worth today or tomorrow is not so important to me. This camera is a keeper. It is compact, solid, easy to bring along, with full manual settings available. The memories inprinted on all those photos I have taken with this camera has been worth the investment. With the S40 I have really started to enjoy the art of photography. I've taken a lot of excellent pictures with the S40, and I am often surprised how well the pictures can turn out, as long as I do the right things in the given situation. But it has its weaknesses, and some of these weaknesses are connected to the artifacts caused by the bayer interpolation. In a year or so I'm planning to invest in a DSLR. Most of the Sigma SD9 pictures I have seen have qualities that I would like to see in my future DSLR. Looking at digital cameras (pictures/samples/reviews etc.) for years, I can see that the Foveon sensor is something new. And it looks very promising. That's why I'm spending time at this forum!
Karl Guttag said:
What I really object to in your statement was "The ISO noise in a
typical bayer image contains larger dots." What is it that you are
calling the "typical bayer?"
Click to expand...
I've seen a lot of digital photos/samples/reviews of digital cameras, and at higher ISOs, they all seem to have some characteristics that can be characterized as typical for a bayer sensor. The bayer images have noise (grain) characteristics that are not similar to what I see from film. Not to wonder, since the sensor data originates from a mosaic pattern that needs to be interpolated. The bayer interpolation is blurring the noise, and this is clearly visible as larger blurred dots/areas. As you know, the Foveon sensor data does not need to be interpolated, and consequently the noise is left as it is without being blurred. That was my point.

I've put together this (not so scientificly) comparison:



(200% magnified crop)

Same comparison, boosted brightness (+50) and contrast (+25):



Well, every people are entitled to have their opinion about this, but if you ask me, I am not put back by the noise in the SD9 images.

Geir Rune
 
What's really funny is that Karl seems to be quite critical of this sensor...and he posts pictures where 5-pixel letters (original, not down-sampled) are clearly legible on the Fuji box...this will never be the case with a Bayer-pattern sensor. Karl...if you are reading, please admit that the Fovion, if nothing else, produces images that are markedly sharper (on a per-pixel basis) than the images that come from Bayer-pattern cameras. I appreciate your criticism and you've made some good input regarding sensor-related issues. However, you appear to me to have a strong bias. Are you not impressed with the sharpness of the Fovion images?
Zoltan Csuka said:
Karl Guttag said:
richard peters said:
Have these images been discussed? Two sets at 100, 200, 400 ISO .

http://www.numeritest.com/Sigma-SD9.htm

Richard
Click to expand...
Nice Find.

A few observations:

First some crops from the Fruit and Film Box picture.

Taken at ISO100 50F5.6 (don’t know the lens)

http://www.fototime.com/ {2373E0C8-BE1F-457D-8F51-3EA0FB22CB55} picture.JPG

Now taken at ISO400 50F5.6

http://www.fototime.com/ {10069C16-1D10-4998-B927-24067611C569} picture.JPG

Look at the “Kodak” in “Kodak Professional,” on the lower right the
Red is fading away. If you can on your monitor scroll so you can
see the ISO100 and ISO400 "Kodaks" at the same time. Then look at
the Green in the Fuji box, it has “posterized” or gone blotchy (as
has been seen in other SD9 pictures). Also look at the white
lettering in the Kodak box such as the words “ultra color” you will
see the Red bleeding into the white where it did not in the ISO100
shot.

Now for the Messe building shot:

At ISO 100 and 70F5.6:

Look at the purple blow out in the left side of the window and the
right side of the white window frame (it might help to use
DPREVIEW's 200% option).

http://www.fototime.com/ {B98699DF-21AE-4C27-B8AA-639282F23CC0} picture.JPG

Now for the same shot at ISO400:

http://www.fototime.com/ {FF5ACE74-B56C-41D2-8193-80547533DE2A} picture.JPG

The purple “blowout” from the ISO100 is gone! But now the bricks
to the left of the window have a green pattern and purple pattern
in them.

I will agree that it is still too early to call for sure, but at
least these pictures seem to be much better than the IR ones for
making some observations -- It would be nice if they had taken the
exact same shots with a D60 for comparison.

Oh yeah, how about them colors in the SD9's B&W test patterns? Ok
cheap shot, but there is something strange going on. Folks, it is
not chroma aberrations but something going on with the sensor or
the software (see also the Messe ISO100 picture above).

Karl

--
Karl
Click to expand...
--
Best wishes,
Zoli
Click to expand...
--
http://www.pbase.com/mango
 
stuffy,

Thanks for the reply.
Stuffy said:
since
the sensor is billed as being less susceptible to moire (remember
the photo of the tweed cloth?), I'm left wondering what the real
story is.
Click to expand...
To get the same resolution with a Bayer-sensor camera, you need more pixels. This means that Moire won't appear until farther down on the res chart. If you compare the fovion with the Canon D30, for example, there is no comparison. The D30 shows color moire, then solid gray where the Fovion still clearly resolves features. All the Moire on the fovion sensor appears far beyond the resolving limit of an equivalent pixel Bayer sensor. If you are concerned about moire, apply a 1.5 pixel blur to the Sigma res chart. You won't see any Moire, and the resolution will be about what you get with the Canon D30.
Stuffy said:
Some aspects/portions of the Sigma res chart are as good as, or
even clearly suprerior to, the D60. Other aspects/portions are
vastly inferior. The Sigma chart looks a whole lot better when I
desaturate it.
Click to expand...
Down-sample the D60 to match the Sigma. Then upsample the Sigma to match the D60. I said that the sigma rivals (not beats) the D60.
Stuffy said:
So, RESOLUTION ASIDE, if you think the Sigma chart is as good as
the D60 chart, then yes, we are in strong disagreement. And since
Phil's evaluation of the chart almost always emphasizes resolution
above all else, we may never have an impartial third-hand judgement.
Click to expand...
I don't see how we can put resolution aside if we are using resolution charts as the image subject. I will agree that Moire will show up at lower resolution on the 3MP Sigma than on the 6MP Canon.

I have no intention of buying a(nother) SLR camera. I have a Canon S30 as my sole digicam, so I'm not a serious photographer, and I don't know jack about anything. However, I am interested in digital photography. All other things being equal, It is my sincere belief that, were my S30 to be equipped with an equal MP Fovion sensor instead of the Bayer sensor, the camera would produce images that would be FAR SUPERIOR . That's all.

Fovion type sensors ultimately have the ability to sample more photons/area than Bayer-type sensors. This is their ultimate advantage. At the end of the day, all other things being equal (which they never are!!!) Fovion-type sensors will always produce better results at any given MP number. My S30 was my first and last Bayer-type digicamera. That said by a man who is absolutely thrilled with the performance of his camera. However, it's obvious that a fovion-equipped camera would be that much better.

cheers.
 
Ladehaug said:
I cannot see this banding on the original (on my monitor). However,
after adjusting the contrast, some slight banding does indeed show
up.
Click to expand...
There have been some threads in the Canon SLR forum on horz banding as well in respect to the 1Ds and many other cameras. I recall Phil has pointed it out as a jpeg artifact.

I myself not have paid any attention to how it looks like within the Canon SLR forum. Do you guys know if this is the same a different pattern?

--
jc
Sony F707
http://www.reefkeepers.org/gallery/f707
http://www.reeftec.com/gallery
 
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