WB comparison

One point of criticism in this forum is usually the Auto White Balance of the SD9. Somebody just reminded us in another thread that the SD10 also does not a good job.

Well somebody could pay a little attention to this and tell me what he/she sees and reads here:

http://www.dpreview.com/reviews/canoneos10d/page19.asp

compared to this:

http://www.dpreview.com/reviews/sigmasd9/page16.asp

maybe someone has the time to compare that a bit more in depth, for example putting the samples next to each other and also include the S2pro and D100....

--
'A polar bear is a rectangular bear after a coordinate transform'

Regards from Old Europe,

Dominic

http://www.pbase.com/sigmasd9/dominic_gross

Very interesting. Your links seem to show that the SD-9 auto WB
is decent, even under incandescent, much better than the 10D.
I'm surprised by that.

But, when Chunsum posted this:
http://www.pbase.com/chunsum/sd10_iso_wb_test

I understood that he took those pix using Auto WB with
incandecsent lighting. The X3F output from Spp showed an extremely
yellow cast. SO, I concluded that the SD-10 Auto was no good
under incandescent.

But, perhaps I misunderstood what Chunsum did, or maybe he did
something wrong. He did not explain his testing conditions very
clearly.

I can't wait to see how Phil's review turns out on this.

ERic

Dominic Groß said:

One point of criticism in this forum is usually the Auto White
Balance of the SD9. Somebody just reminded us in another thread
that the SD10 also does not a good job.

Well somebody could pay a little attention to this and tell me what
he/she sees and reads here:

http://www.dpreview.com/reviews/canoneos10d/page19.asp

compared to this:

http://www.dpreview.com/reviews/sigmasd9/page16.asp

maybe someone has the time to compare that a bit more in depth, for
example putting the samples next to each other and also include the
S2pro and D100....

--
'A polar bear is a rectangular bear after a coordinate transform'

Regards from Old Europe,

Dominic

http://www.pbase.com/sigmasd9/dominic_gross

Click to expand...

Sorry about the lack of informatin on my test Eric.
I should have been clear on the parameters.

all the shots are taken with camera Auto WB.
other WB setting you see are done in SPP2.
the main light source is a 80W Incandescent Bulb.
However, the lights on the trees does change in color.

Here's my take on it. if you take a look @ the iso 100 shot,
I can tell you that it is fairly accurate. However @ iso 200 and up

there is this yellow cast. I dunno if it is because of the ISO or simply because the shots are under exposed.

I'll take a few more shots on the same tree in the next few days with different in camera WB and I'll bracket the shots.

if anyone has a suggestion on what else I can do, please let me know.

Chunsum.

Anon Y Mous said:

But, when Chunsum posted this:
http://www.pbase.com/chunsum/sd10_iso_wb_test

I understood that he took those pix using Auto WB with
incandecsent lighting. The X3F output from Spp showed an extremely
yellow cast. SO, I concluded that the SD-10 Auto was no good
under incandescent.

But, perhaps I misunderstood what Chunsum did, or maybe he did
something wrong. He did not explain his testing conditions very
clearly.

I can't wait to see how Phil's review turns out on this.

ERic

Dominic Groß said:

One point of criticism in this forum is usually the Auto White
Balance of the SD9. Somebody just reminded us in another thread
that the SD10 also does not a good job.

Well somebody could pay a little attention to this and tell me what
he/she sees and reads here:

http://www.dpreview.com/reviews/canoneos10d/page19.asp

compared to this:

http://www.dpreview.com/reviews/sigmasd9/page16.asp

maybe someone has the time to compare that a bit more in depth, for
example putting the samples next to each other and also include the
S2pro and D100....

--
'A polar bear is a rectangular bear after a coordinate transform'

Regards from Old Europe,

Dominic

http://www.pbase.com/sigmasd9/dominic_gross

Click to expand...

Click to expand...

--

SD10 in my hands,there are lots of learning ahead.

Dominic is right when he says there has been a lot of criticism of the SD9's WB performance. I've only been an owner of the SD9 for a few weeks, now, (previously using a Minolta Dimage7, which has good AWB performance, IMO..), and I have been pleasantly surprised at the AWB of the SD9. I was expecting to have trouble getting neutral greys, etc, but I am amazed at the job the SD9 does on auto most of the time. I have found SPP2 to do a great job of fine tuning the color on "auto", too, and particularly easy to manually tweak color casts by using the color wheel and watching the CMY figures. Generally speaking, I think anyone used to old fashioned color printing, (such as myself), would also like the SD9 and SPP combination.

PS. Phil's tests, which you provided links to, make quite a dramatic comparison in favour of the SD9, BTW!

Anon Y Mous said:

But, when Chunsum posted this:
http://www.pbase.com/chunsum/sd10_iso_wb_test

I understood that he took those pix using Auto WB with
incandecsent lighting. The X3F output from Spp showed an extremely
yellow cast. SO, I concluded that the SD-10 Auto was no good
under incandescent.

But, perhaps I misunderstood what Chunsum did, or maybe he did
something wrong. He did not explain his testing conditions very
clearly.

I can't wait to see how Phil's review turns out on this.

ERic

Dominic Groß said:

One point of criticism in this forum is usually the Auto White
Balance of the SD9. Somebody just reminded us in another thread
that the SD10 also does not a good job.

Well somebody could pay a little attention to this and tell me what
he/she sees and reads here:

http://www.dpreview.com/reviews/canoneos10d/page19.asp

compared to this:

http://www.dpreview.com/reviews/sigmasd9/page16.asp

maybe someone has the time to compare that a bit more in depth, for
example putting the samples next to each other and also include the
S2pro and D100....

--
'A polar bear is a rectangular bear after a coordinate transform'

Regards from Old Europe,

Dominic

http://www.pbase.com/sigmasd9/dominic_gross

Click to expand...

Click to expand...

Ok, so the walls really are peach colored in real life from the iso 100
shot. So the iso 100 shots are good with auto WB, but then the auto
WB goes crazy at iso 200 and above?

Phils review of the SD-9, those jpgs testing the WB were for
iso 100 (as shown in the exif).

So , I'm glad that I was wrong about the WB in iso 100.
If the higher iso have poor auto WB, then that sounds like another
firmware update.

Eric

Chunsum J Choi said:

Sorry about the lack of informatin on my test Eric.
I should have been clear on the parameters.

all the shots are taken with camera Auto WB.
other WB setting you see are done in SPP2.
the main light source is a 80W Incandescent Bulb.
However, the lights on the trees does change in color.

Here's my take on it. if you take a look @ the iso 100 shot,
I can tell you that it is fairly accurate. However @ iso 200 and up
there is this yellow cast. I dunno if it is because of the ISO or
simply because the shots are under exposed.

I'll take a few more shots on the same tree in the next few days
with different in camera WB and I'll bracket the shots.

if anyone has a suggestion on what else I can do, please let me know.

Chunsum.

Click to expand...

Bear in mind that the target "image" used by Phil has a lot of white. Which might make it easier to guess the chromaticity of the illuminant for certain "auto" algorithms and not for others.

Conversely, algorithms that work well with lots of white may or may not work well with a natural scene.

Just a thought.

• kc

yeah, the color of the wall is call Valentine. it's white with a hint

peachy-pink, so my wife tell me. I'm glad that true because the paint was expensive and took me a week to get 3 coats done.

Anon Y Mous said:

Phils review of the SD-9, those jpgs testing the WB were for
iso 100 (as shown in the exif).

So , I'm glad that I was wrong about the WB in iso 100.
If the higher iso have poor auto WB, then that sounds like another
firmware update.

Eric

Chunsum J Choi said:

Sorry about the lack of informatin on my test Eric.
I should have been clear on the parameters.

all the shots are taken with camera Auto WB.
other WB setting you see are done in SPP2.
the main light source is a 80W Incandescent Bulb.
However, the lights on the trees does change in color.

Here's my take on it. if you take a look @ the iso 100 shot,
I can tell you that it is fairly accurate. However @ iso 200 and up
there is this yellow cast. I dunno if it is because of the ISO or
simply because the shots are under exposed.

I'll take a few more shots on the same tree in the next few days
with different in camera WB and I'll bracket the shots.

if anyone has a suggestion on what else I can do, please let me know.

Chunsum.

Click to expand...

Click to expand...

--

SD10 in my hands,there are lots of learning ahead.

From these pics it does seem that SD9 AWB works better. However, In the rightmost column (custom WB) I see a lot more hue variation depending on the light source than in 10D test. It's not like 10D doesn't care, but at least it doesn't render deep red as orange.

This is most noticeable on red and green patches.

P.S. Yes, my monitor is calibrated.