Three new SD9 pics - all for now

  • Thread starter Thread starter Phil Askey
  • Start date Start date
The Canon D60 comes, in their package #1, with a 28-200mm lens.
And the Sigma comes with a 28-135mm lens, in package #1.
So it's not exactly comparing apples to apples, but just the same,
for the Sigma the package price is $1989. For the Canon D60 the
package price is $2689.

A $700 difference is a lot of money.
28-200mm lens: $450
28-135mm lens: $179

The D60 deal also includes 1 year extended warranty ($60).

So if we assume that an extra BP-511 costs about the same 4 AA + charger + case, then the D60 "bundle" has $280 more stuff in it.

So the price delta is more like $420 or almost exactly the body price diff.

--
Erik
 
Weird guy, Karl you're certainly on a mission to discredit yourself.
Probably not. Foliage looks noisy as it is. :-)
Yes, that is why it is hard to tell much about the camera quality
from foliage.
The SD9 may be the ideal outdoor, sunshiny, woodsy, daytime camera.
Heheheheh...
Direct Sun shine is probably one of the most demanding conditions
for photography, particularly digitals and the evidence so far is
that the SD9 has problems with bright light. The SD9 seems to
have problems with bright to dark contrast and gets purple blowouts
and then there are bright saturated Reds/Oranges turning white
(such as the picture of the traffic cone turning white in the horse
track shot).

It would seem that the SD9 would do best on low contrast days like
overcast conditions.

I have a lot of problems with the SD9 as a general use camera based
on my experience with a D30. I think there are way too many
conditions where the SD9 would not be usable. Such as:

Sunny days - blown out colors and purple blowouts

Night time - blown out lights and noisey at higher ISO

Sports - Even kids sports are best shot at 1/500th or faster
shutter speed to cut down motion blur. On sunny days you have to
worry about blow outs (this is an issue for any digital and even
film, but seems particularly bad on the SD9). Most longer zooms
which are often used to capture sports are F5.6 (at the long end)
and on cloudy days, late afternoon or early morning you will have
to be at ISO400 to ISO800 to keep the shutter speed up. The
performance of the SD9 at high ISOs it not good compared to other
cameras. So I can't see shooting kids sports except on bright
overcast (thin clouds).

Flash - I often shoot at ISO400 with flash to reduce the harshness
of the flash and to take advantage of available light (to keep it
from looking like people are standing in the dark). This would
not be an issue with studio lighting, but with typical candid shots
it is a big issue.

So the way I see it, an SD9 would sit in the closet waiting for the
"right" conditions. Most amatuers need a general purpose camera
and not one for just taking still lifes on overcast days.

Karl
I'm just hoping it can do at least okay in the dark just to at
least keep in the running.
It is really hard to know without seeing several side by side shots
of the exact same set up. The SD9 appears to be noisy in the
Green channel, but this may not be noticable in foliage.

--
Karl
--

Ulysses
Repository of Some of My Stuff
http://www.imagestation.com/album/pictures.html?id=4291269101

I'm an uncle!!!

--
Karl
--
Phil Askey
Editor / Owner, dpreview.com
 
Good for you, Karl

The King has just appointed you the court jester.
Probably not. Foliage looks noisy as it is. :-)
Yes, that is why it is hard to tell much about the camera quality
from foliage.
The SD9 may be the ideal outdoor, sunshiny, woodsy, daytime camera.
Heheheheh...
Direct Sun shine is probably one of the most demanding conditions
for photography, particularly digitals and the evidence so far is
that the SD9 has problems with bright light. The SD9 seems to
have problems with bright to dark contrast and gets purple blowouts
and then there are bright saturated Reds/Oranges turning white
(such as the picture of the traffic cone turning white in the horse
track shot).

It would seem that the SD9 would do best on low contrast days like
overcast conditions.

I have a lot of problems with the SD9 as a general use camera based
on my experience with a D30. I think there are way too many
conditions where the SD9 would not be usable. Such as:

Sunny days - blown out colors and purple blowouts

Night time - blown out lights and noisey at higher ISO

Sports - Even kids sports are best shot at 1/500th or faster
shutter speed to cut down motion blur. On sunny days you have to
worry about blow outs (this is an issue for any digital and even
film, but seems particularly bad on the SD9). Most longer zooms
which are often used to capture sports are F5.6 (at the long end)
and on cloudy days, late afternoon or early morning you will have
to be at ISO400 to ISO800 to keep the shutter speed up. The
performance of the SD9 at high ISOs it not good compared to other
cameras. So I can't see shooting kids sports except on bright
overcast (thin clouds).

Flash - I often shoot at ISO400 with flash to reduce the harshness
of the flash and to take advantage of available light (to keep it
from looking like people are standing in the dark). This would
not be an issue with studio lighting, but with typical candid shots
it is a big issue.

So the way I see it, an SD9 would sit in the closet waiting for the
"right" conditions. Most amatuers need a general purpose camera
and not one for just taking still lifes on overcast days.

Karl
I'm just hoping it can do at least okay in the dark just to at
least keep in the running.
It is really hard to know without seeing several side by side shots
of the exact same set up. The SD9 appears to be noisy in the
Green channel, but this may not be noticable in foliage.

--
Karl
--

Ulysses
Repository of Some of My Stuff
http://www.imagestation.com/album/pictures.html?id=4291269101

I'm an uncle!!!

--
Karl
--
Phil Askey
Editor / Owner, dpreview.com
 
Direct Sun shine is probably one of the most demanding conditions
for photography, particularly digitals and the evidence so far is
that the SD9 has problems with bright light. The SD9 seems to
have problems with bright to dark contrast and gets purple blowouts
and then there are bright saturated Reds/Oranges turning white
(such as the picture of the traffic cone turning white in the horse
track shot).
But in general, overall, that horse track picture isn't really so bad.
A large part of the picture is fine. But what about the cones were they totally lost the color? Let say you are out shooting soccer in the Sun. Are you going to accept the teams wearing red uniform blowing out all the color?
And Phil's "Tower Bridge" picture is fine too. And, in fact, during
the last 5 weeks I have seen plenty of pictures taken on sunny
days with the SD9 that were perfectly fine, pretty much.
If the lighting is even and not contrasty, the SD9 seems to be ok. The problem is that there are a lot of situations where people take pictures where the lighting is not the best.
Another thing I just noticed is over at State Street Direct, their
"package number 1" prices. For myself I've never owned any
SLR,
Not owning/having used a Digital SLR seems to be a common for many of the SD9/Foveon fans. If I had a choice between the current Foveon sensor in a Canon mount (since I have Canon lenses -- thus ignoring the "Simga issues") or the D30 (no less the D60) sensor, I would still rather have the D30. I can't remember EVER loosing a shot due to the Bayer pattern. The SD9 would have better resolution than the D30, but the limitations of when I could use the SD9 would be terrible.
so I would probably get one of their package deals... if I
were in the buying mood.
I think this is one of the big issues. To get a camera and a "cheap" consumer grade lens you are talking $2,000 plus batteries and flash/microdrives. This is a big jump up from a $500-800 point and shoot. Plus a consumer can get a 50F1.8 for about $85 where the closes thing in the Sigma line is a 50F2.8 Macro that costs about $200 more.
The Canon D60 comes, in their package #1, with a 28-200mm lens.
And the Sigma comes with a 28-135mm lens, in package #1.
So it's not exactly comparing apples to apples, but just the same,
for the Sigma the package price is $1989. For the Canon D60 the
package price is $2689.
As Eric pointed out, this is a lot of apples to oranges and backing out the extras gets you back to about a $420 difference. Ignoring the shell game of the packages, the diffrence is about 20% for the D60 body.

Frankly, I think Canon and Nikon need the equivalent of a "Digital Rebel G." A DSLR that comes in at about $1000. I expect to see this in the next year to hook people on their various lens systems.

--
Karl
 
But in general, overall, that horse track picture isn't really so bad.
A large part of the picture is fine. But what about the cones were
they totally lost the color? Let say you are out shooting soccer
in the Sun. Are you going to accept the teams wearing red uniform
blowing out all the color?
I would prefer to see an actual picture of kids (or adults) playing
soccer in the sun, with one team wearing red uniforms, and their
uniforms being all blown out, before I accept that would happen.

Maybe you right, but I am still doubtful that the situation of "red
in sunlight" is entirely hopeless for the Foveon sensor.
As Eric pointed out, this is a lot of apples to oranges and backing
out the extras gets you back to about a $420 difference. Ignoring
the shell game of the packages, the diffrence is about 20% for the
D60 body.
Part of me would like to get a D60, because I own a Canon G2, and
so I look at the D60 and I already know about what 75% or more
of the controls do.
Frankly, I think Canon and Nikon need the equivalent of a "Digital
Rebel G." A DSLR that comes in at about $1000. I expect to see
this in the next year to hook people on their various lens systems.
It would be nice to buy my first SLR. I would enjoy that. Regular
film photography was too expensive, on a per image basis, to be
fun for me. But this digital stuff is great. And I get to see the
images on my PC as soon as I get home. I enjoy digital photography
much more than film, so I am finally ready for a SLR, I guess.

Based on how much I use my G2, I get the impression it would
be a smart purchase for me.
 
Paul,

I'm not sure exactly what your are refering to. In the shot of the
leaf the DoF was very small.

There does seem to be a problem with the X3 sensor with colors
separating as the radius increases. Many on this forum try to
write it off as lens aberations and while some of it may be due to
lens aberations, I think there is another effect caused by the
angle of the light hitting the X3 pixels.

As the image radiates out from the center, the light hits the
pixels at ever steeper angles.
Karl,

What't shallowest angle (largest deviation from normal) that light strikes the sensor in the S9? We're talking less than 30 degrees, right? Prob a tad less. At 30 degrees, the light is still 86% perpendicular to the sensor. What's the aspect ratio of the wells (lateral:depth)? How much lateral bleed do you expect? If the aspect ratio is 5:1, then you'd expect less than 3% lateral bleed (worst case). The CA that you pointed out in past posts represents more than 3% of the light at a given pixel being detected at the adjacent pixel.

I truely believe that you've made some rather simplistic errors in your analysis of the CA.

I'm not a "Bayer Basher" as you like to call them. I think the Fovion will have it's own issues (color fidelity, dynamic range, light sensitivity perhaps?), but for you to consistently deny that the sensor offers significant advantages when the evidence is plain to see makes you look increasingly foolish.
 
Phil,

Just curious, what specifically do you think I said that will proven to be untrue?

The SD9 has proven to have problems with bright saturated Red turning yellow or white. That pretty well means leaving out a lot of sports uniforms.

It is noisy with certain colors in flat areas (see for example IR's color test patterns).

The purple blow outs and problems with lights at night are even acknoledged by many Foveon fans.

The lack of performance at higher ISO (or even having ISO's above 400) is a big drawback in a DSLR of this price class.

I left off the problem with "Dust". Every SD9 shot I have seen with sky in it has dark spots even at F8. It looks like your camera has relatively little dust compared to the other SD9's we have seen. I would suggest you do a dust check shot before and after shooting. I think eveyone would like to know what is causing these spots (a lot could be told by taking shots of the sky at various apertures to isolate the distance of the dust from the sensor).

My only really "controversial" item is that the X3 sensor likely has a radial color problem based on the exit angle of the light from the lens. I would really like to see a crop of a scan with a film body using the same lens using that 20-40mm zoom at F8. I'll be curous to see if the other wide angle lenses have similar problems with the X3/SD9. You probably will not see it with the 50F2.8 but might with some prime that is wider. I'm fairly confident that I will be proven right on this one.

I think when you have fully evaluated the camera under a variety of shooting conditions you are going to have to come to the conclusion that it has a limited range of application when compared against the other DSLRs available today (particularly the D60, D100 and S2).

It MAY be that you will find that when the light is good and the colors are not extreme (particulary Red) that the camera produces very good images. but so do the other DSLRs. In my experience, I end up shooting a lot of pictures where I can't control the lighting as much as I would like and can't keep certain colors out of my pictures.

Maybe the SD9 will make it as a "specialty camera" for taking still lifes. But I can't see it competing with the other DSLRs as a general purpose camera for shooting in a variety of conditions. Imagine trying to shoot Soccer/Football will a long telephoto in bright sun and cloudy conditions. In bright sun you have to worry about blowing out colors, in cloudy conditions you are limited by the ISO to get the shutter speed you want. And forget indoor sports with ISO400.

I would be really suprised, based on your other reviews, if you will in the end be able to come out and whole heartedly recommend this camera over the others.

Karl
Weird guy, Karl you're certainly on a mission to discredit yourself.
--
Karl
 
Karl, I understand what you're saying.

But I have to say I haven't seen the SD9 images to be quite as limiting as you describe. That's not to say that it doesn't have a "sweet spot" for its best exposures. But where that spot is, we have yet to see from a good review and good pictures.

For example, on some of the Imaging Resource images of a while back, I wasn't all that crazy about their overcast shots. Not bad, but definitely not at the camera's best.

At any rate, it seems that much remains to be seen yet.
It would seem that the SD9 would do best on low contrast days like
overcast conditions.
--

Ulysses
Repository of Some of My Stuff
http://www.imagestation.com/album/pictures.html?id=4291269101

I'm an uncle!!!

 
Karl,

With all due respect, you seem to have found your form again. For a while I thought you had lost it.

Now the issue is full sunlight, where I think you will find that the Foveon shines. With clean single-pixel resolution, my chances of getting a good picture under these conditions are far greater than they would be under conditions admittedly very difficult for this sensor (twilight with brake lights seems to have become one of your more recent Idefixes).

The issue of the blue sky is also an interesting twist. Have you considered that this might actually be where the Foveon is vastly superior to the other sensors. I don't know; I am not a technician. However, in theory, a sensor concept that relies on pixels being guessed for interim values will naturally push them all into one narrow spectrum when that color predominates in the area, whereas a sensor concept that does not, won't. Could this also have to do with the extensive amount of noise floating around in your perception of these images?

I am not sure what kind of monitor you're using, nor do I know what kind of printer you have. However, on my SGI (1600x1024; calibrated), these look pretty good -- consistently. At A3 from my Epson 2100 they are stunning. What matters more in the end?

You were doing pretty well for a while here, even grudingly admitting at Uly's urging that there might be a minor point or two that is good about the sensor. But you have regressed into your "totally worthless" mode"

Before this thread drops to SD9-bashing vs. Karl-bashing thematics, why not give it another try? List two good reasons why someone coming from the prosumer ranks should consider buying a Sigma SD9 with a couple of good prime lenses and perhaps a fine zoom. Money's not the primary concern here. Just the goods.

Laurence
 
[...]
A camera like the canon 1ds has very impressive resolution,
but its images still look softer when viewed at the same
DPI - by looking the images on ones monitor for example.
No doubt you get get a better print from the 1ds because
of its higher intrinsic resolution, but the SD9 appears to
have an advantage in producing a higher quality 'pixel'
in the first place.
I don't mean to be blunt, but if the print is better, there is less
noise, and the operation of the camera is better, how is this an
advantage? OK, price versus the 1Ds certainly.
Currently I see no advantage. Resolution is resolution - it
amounts to the same thing. For example, I doubt that
the 3.5 MP X3 is any better then a 6 MP bayer sensor -
probably worse in the majority of shooting situations.

You know, in some ways a bayer sensor is closer to
how the human eye works. My understanding is there
are three (or more) types of light sensitive cells in the fovea,
each one sensitive (more or less) to one primary color.
So the eye does not 'see' a full color point until after
the brain has processed the input from these cells,
much like the processing that must be done with a
bayer sensor.

I don't have any serious problem with the bayer design -
a little intelligence (the demosaicing algorithm) goes a
long way. In general I think it's better to do things in
software if possible - it's more flexible than having to
attempt to change the hardware (sensor).
In the future this ought to be more of an advantage, when the
mosaic sensors start having to get obnoxiously small pixel spacing
to increase MP (past 16MP I think this will be an issue), and
Foveon-type sensors can then step in. A 16MP full-frame Foveon
would be pretty cool.
The future is where I think the Foveon will really shine.
Assuming they can improve the sensitivity and get out some
of the other bugs. Consider:
  • On a per pixel basis the foveon is superior. When things
finally settle down and there are some standard resolutions
and sizes for sensors the foveon will be hard to beat. A full-
frame 16 MP foveon is going to beat a full-frame 16 MP bayer
sensor, clearly.
  • Assuming they can manufactor the foveon using a
smaller process (than the current .18 micron ) things will
improve. The 'fill factor' (the part of each pixel devoted
to control circuitry and not sensitive to light) will be smaller,
and the overall sensitivity of the design will improve.

Of course, none of this means the foveon will survive.
They may have missed their window of oppurtunity.
Time will tell.

-Eric
It's funny since I am always seeing people say, "the final print is
the only measure," and so far I've actually disagreed somewhat --
otherwise I'd be using a 4x5 film camera, but there are host of
reasons why I prefer 35mm digital. Hmm -- I could argue that the
D30 has better quality per "pixel" than an 8x10 film camera. Of
course that 8x10 has just a few more pixels.
-Eric
 
Thanks Phil, for taking the extra time to post these images. I am impressed. I have counted no less that 8 spots (in the sky) of the London bridge Pic. What's going on. They appear on every camera I have seen images from to date.
??????????????????
Live, Laugh and Love
 
Weird guy, Karl you're certainly on a mission to discredit yourself.
Phil, so you are saying that the SD9 does not blow out bright colors?
My gut feeling tells me the RAW conversion software is responsible for at least some of the blowing out of non-diffused reflected parts (e.g., specular reflections and light sources) because the dynamic range in those cases are much larger than that from diffused reflective surfaces. Non-specular reflections don't have much dynamic range; for example, a good photographic print only has about 45 dB or so, or about 8 bits, worth of DR.

The reason for some of the blow-outs COULD be because the sensor itself really does have a pretty large dynamic range but it is not being handled properly in the conversion software. Any good scientist would wait for more information before jumping to conclusions about the sensor itself.

This is just pure conjecture on my part, and I am waiting to see if Phil can tame some of those images by playing with the knobs in the conversion software.

If you make a print in the darkroom with all sorts of blown highlights, would you blame a perfect negative? Sure, it could be due to a bad negative, but we don't know yet, do we?

The SD-9, with just RAW output is giving you all the flexibility, and thus rope, to hang yourself with. But that is why I like it, reminds me of Dektol and D76 once again! :-) And like darkroom work, there is no such thing as a "default setting" that works for all occasions, which seems to be what most people, with the exception of Phil now, are using. I believe Jefrrey has stated that he did not see blown reds in the RAW image, but he did not confirm after the initial message. The baton (literally) now goes to Sam (thank you so, so much by the way, Jeffrey-san, if you are reading).

Just my humble two-bits.
  • kc
 
As I get to see more samples I'm beginning to get a feeling for the
Foveon "look". Like everyone else I check out the obvious things
like the spires on Tower Bridge, and I'm impressed by the existence
of real edges (look at the cab of the crane in the SD9 shot and
compare to the D60). But more important for me are the areas of low
contrast that define surfaces. The stonework on the bridge and the
embankment of the far side of the river show the advantage (to my
eyes) of the Foveon. The D60 image is typical of the usual digital
look: broad areas of subtle tonal variation have a mushy look and
there's no way to recover the sense of "surface" in post
processing. The SD9 stone just seems stonier and more clearly
localized in space to me. It seems that each pixel knows what color
and brightness it wants to be without hedging. In film terms the
difference reminds me of the difference between using a
high-acutance (crisper, grainier) versus a fine-grain (softer,
smoother) developer. Very subjective I know...John K.

--
JLK
You are into some important observations! It's difficult to describe what it is, but it's definately there, this "something" about the SD9 images that makes it more "alive and kicking"!

Geir Rune
 
You are into some important observations! It's difficult to
describe what it is, but it's definately there, this "something"
about the SD9 images that makes it more "alive and kicking"!
Take a look at fine textures, Geir.

Everybody worries about aliasing causing Moire. That is the extreme case. Before the onset of visible Moire, alaising will already cause textures to smear.

Look at the detail in stones, fabrics, leaves, tree branches. These are the common things that we see (well, excluding the dollar bill :-). We are used to seeing certain textures in real life, when they are smeared in images, we feel they don't look natural (posterized, whatever) but often can't put a precise handle on it.

At every resolution, not just high resolution, you expect to see different textures (fractal nature of "nature" :-), otherwise your brains just don't click. It is something that is hard to measure, but easy to "feel." But when assembled together, gives the entire image your "alive and kicking" perception. Your mind is not playing any new tricks on you anymore than what it is already doing when you see a naturall scene.

Show me a machine that can differentiate between a mediocre violin and an execellent violin. Humans can tell right away.
  • kc
 
Having initially suggesting the SD9 has not a lot to offer, I have read lots of other postings and come up with the theory we must rate the image in 3 common ways.

1. Printing on inkjet, just to full numberr of pixel staight through your printer driver

2. Viewing 100% on our VDU, and there are at least 3 kind out there!

3. After sizing to 800 or 1024 pixels wide and viewing at 100%

Also some might say what about chemical printing or dyesubs!

I am betting the 5 and 6 mp bayer sensors might stand up of some of these test.

Go experiment !

--
Graham

http://www.graham.uk.net
 
My gut feeling tells me the RAW conversion software is responsible
for at least some of the blowing out of non-diffused reflected
parts
Certainly is could be the software. But then the puzzling thing is why? Compressing the dynamic range is not a hard problem - otherwise the competing cameras would usually have the same problem. (And they do, but to a much lesser extent.)
Any good
scientist would wait for more information before jumping to
conclusions about the sensor itself.
So it's fair to jump to conclusions that it's a software bug? Either jump (and justify your answer) or don't jump. Can't have it both ways ;-)
This is just pure conjecture on my part, and I am waiting to see if
Phil can tame some of those images by playing with the knobs in the
conversion software.
Yes, we all are waiting for more info. But w/o speculation from incomplete data, this forum would be empty!
If you make a print in the darkroom with all sorts of blown
highlights, would you blame a perfect negative?
Well, it's unusual for "standard settings" to make a bad print from a perfect negative. So can we be blamed for assuming the starting point is imperfect? The alternative is to assume major flaws in the standard processing. Which is more common?

You mentioned Ansel Adams in another thread. Have you read much about the extent of the manipulations he would to to both the negative and the print to get the look he wanted?

--
Erik
 
Karl, I understand what you're saying.

But I have to say I haven't seen the SD9 images to be quite as
limiting as you describe. That's not to say that it doesn't have a
"sweet spot" for its best exposures. But where that spot is, we
have yet to see from a good review and good pictures.
I don't know about you, but I rarely get the lighting I want. If the kids are playing in the sun or shade, that is what I get. I can try and position myself based on the lighting and use different lenses and focal lengths and F-numbers, but often the situation dictates what you can do. If they are wearing bright red uniforms, I can't ask them to change into a neutral color. I'm not afraid to shoot at night with my D30 as well.

If it is overcast and I need a long focal length, then I know I am going to be at F5.6 and to keep my shutter speed up I am going to be over ISO200.

To paraphrase what somebody else said, "photography is all about making the best set of compromises."

I'm sure there will be situations where you can shoot and shoot and not have a problem with the SD9. I don't think most people that are going to have over $2,000 into lenses and the body will really be happy with a camera that they cannot use in a wider range of shooting situations. Half the fun of the D30 has been going out and taking shots that I would have never tried with a Film camera.
At any rate, it seems that much remains to be seen yet.
On that I would agree. I would really like to try a few experiments myself. I know people find test patterns and test picture boring, but they let you see more clearly what is going on. A "real" picture is so busy with information that it makes it impossible to get beyond subjective criteria (in which case we need to do double blind tests to take out the biases).

Besides, I'm still waiting to be proven right about the radial color issue with the sensor :-).
It would seem that the SD9 would do best on low contrast days like
overcast conditions.
--

Ulysses
Repository of Some of My Stuff
http://www.imagestation.com/album/pictures.html?id=4291269101

I'm an uncle!!!

--
Karl
 
When I look at the SD9 (daytime) pictures, it looks to me as if this camera captures the texture of the items better. Comparing the sky in these pictures to some of my film scanned images, I think they look very similar. Pictures from other bayer cameras seem to loose the texture in the sky.

Am I comparing the foveon "noise" to film "grain", or does the sky actually have texture. Any ideas.

(I can not post the samples until I get home).
 

Keyboard shortcuts

Back
Top