SD9 UV / haze-2a filter test (1)

Started Aug 12, 2003 | Discussions
Yi Ding Senior Member • Posts: 1,371
SD9 UV / haze-2a filter test (1)

There hasn't been any sunny day in Columbus since I received my 50mm lens and the haze-2a filter for it, so I only did some preliminary tests.

In the first row, gray reference was picked on the gray card. Although the lighting condition of the sky is not really the same as that of the card, I think the use of a gray card is valid, because the sky is usually only the background of a photo and we want the white balance to be optimized for the main objects. The real sky color in my impression was somewhere between the UV filter image and the haze-2a filter image (maybe closer to the latter, but I'm not very sure). It seems that the haze-2a filter, which looks a little yellowish, cuts off some visible shortwave blue. The sky color in the first two images are vivid and pleasing, just like the effect of a polarizer! If the sky was indeed deeply blue, then it would probably appear unnatural in the first two images.

The lamp in second row used a blue-coated bulb that was supposed to give out some UV light. They were all with auto white balance; choosing the gray reference on the white wall doesn't noticably change the color. There is not much difference between the image with and without the UV filter, while the haze-2a filter really did a good job. What I saw with my eye was nearly the same as the third image.

The most significant benefit of the haze-2a filter is that it suppresses haze and purple fringes very effectively. This is evident in the third row (cloud picked as white reference). The haze-2a filter also made the image a little warmer and the green color more vivid.

In summary, it seems the X3 sensor in SD9 does intensify the blue color of the sky, which sometimes is a good thing. The haze-2a filter probably reduces the sky color, but this is very easy to adjust on the computer (for example, using the magic wand tool in photoshop to select the sky). Overall, the use of a haze-2a filter on SD9 seems to be pretty advantageous. However, why didn't Sigma use haze-2a filter glass to make the dust protector? Is there any other major disadvantage of the haze-2a filter? I'll do more tests, but any input will be greatly appreciated.

Yi

OP Yi Ding Senior Member • Posts: 1,371
sorry, the test photos are here:

ding1 wrote:

There hasn't been any sunny day in Columbus since I received my
50mm lens and the haze-2a filter for it, so I only did some
preliminary tests.

In the first row, gray reference was picked on the gray card.
Although the lighting condition of the sky is not really the same
as that of the card, I think the use of a gray card is valid,
because the sky is usually only the background of a photo and we
want the white balance to be optimized for the main objects. The
real sky color in my impression was somewhere between the UV filter
image and the haze-2a filter image (maybe closer to the latter, but
I'm not very sure). It seems that the haze-2a filter, which looks
a little yellowish, cuts off some visible shortwave blue. The sky
color in the first two images are vivid and pleasing, just like the
effect of a polarizer! If the sky was indeed deeply blue, then it
would probably appear unnatural in the first two images.

The lamp in second row used a blue-coated bulb that was supposed to
give out some UV light. They were all with auto white balance;
choosing the gray reference on the white wall doesn't noticably
change the color. There is not much difference between the image
with and without the UV filter, while the haze-2a filter really did
a good job. What I saw with my eye was nearly the same as the
third image.

The most significant benefit of the haze-2a filter is that it
suppresses haze and purple fringes very effectively. This is
evident in the third row (cloud picked as white reference). The
haze-2a filter also made the image a little warmer and the green
color more vivid.

In summary, it seems the X3 sensor in SD9 does intensify the blue
color of the sky, which sometimes is a good thing. The haze-2a
filter probably reduces the sky color, but this is very easy to
adjust on the computer (for example, using the magic wand tool in
photoshop to select the sky). Overall, the use of a haze-2a filter
on SD9 seems to be pretty advantageous. However, why didn't Sigma
use haze-2a filter glass to make the dust protector? Is there any
other major disadvantage of the haze-2a filter? I'll do more
tests, but any input will be greatly appreciated.

Yi

Laurence Matson
Laurence Matson Forum Pro • Posts: 11,988
Re: SD9 UV / haze-2a filter test (1)

Yi,

That is a great test. Hum. I have always enjoyed our blue skies, but I may have to see if I can scratch up one of these things. I am very impressed with the purple fringe results.

Laurence

ding1 wrote:

There hasn't been any sunny day in Columbus since I received my
50mm lens and the haze-2a filter for it, so I only did some
preliminary tests.

In the first row, gray reference was picked on the gray card.
Although the lighting condition of the sky is not really the same
as that of the card, I think the use of a gray card is valid,
because the sky is usually only the background of a photo and we
want the white balance to be optimized for the main objects. The
real sky color in my impression was somewhere between the UV filter
image and the haze-2a filter image (maybe closer to the latter, but
I'm not very sure). It seems that the haze-2a filter, which looks
a little yellowish, cuts off some visible shortwave blue. The sky
color in the first two images are vivid and pleasing, just like the
effect of a polarizer! If the sky was indeed deeply blue, then it
would probably appear unnatural in the first two images.

The lamp in second row used a blue-coated bulb that was supposed to
give out some UV light. They were all with auto white balance;
choosing the gray reference on the white wall doesn't noticably
change the color. There is not much difference between the image
with and without the UV filter, while the haze-2a filter really did
a good job. What I saw with my eye was nearly the same as the
third image.

The most significant benefit of the haze-2a filter is that it
suppresses haze and purple fringes very effectively. This is
evident in the third row (cloud picked as white reference). The
haze-2a filter also made the image a little warmer and the green
color more vivid.

In summary, it seems the X3 sensor in SD9 does intensify the blue
color of the sky, which sometimes is a good thing. The haze-2a
filter probably reduces the sky color, but this is very easy to
adjust on the computer (for example, using the magic wand tool in
photoshop to select the sky). Overall, the use of a haze-2a filter
on SD9 seems to be pretty advantageous. However, why didn't Sigma
use haze-2a filter glass to make the dust protector? Is there any
other major disadvantage of the haze-2a filter? I'll do more
tests, but any input will be greatly appreciated.

Yi

sunny1267 Contributing Member • Posts: 571
Re: sorry, the test photos are here:

Very useful test, thank you so much!

I would love to see how the haze-2a affects skin tones (if at all) and macro flower shots, if you have an opportunity....

Sunny

ding1 wrote:

ding1 wrote:

There hasn't been any sunny day in Columbus since I received my
50mm lens and the haze-2a filter for it, so I only did some
preliminary tests.

In the first row, gray reference was picked on the gray card.
Although the lighting condition of the sky is not really the same
as that of the card, I think the use of a gray card is valid,
because the sky is usually only the background of a photo and we
want the white balance to be optimized for the main objects. The
real sky color in my impression was somewhere between the UV filter
image and the haze-2a filter image (maybe closer to the latter, but
I'm not very sure). It seems that the haze-2a filter, which looks
a little yellowish, cuts off some visible shortwave blue. The sky
color in the first two images are vivid and pleasing, just like the
effect of a polarizer! If the sky was indeed deeply blue, then it
would probably appear unnatural in the first two images.

The lamp in second row used a blue-coated bulb that was supposed to
give out some UV light. They were all with auto white balance;
choosing the gray reference on the white wall doesn't noticably
change the color. There is not much difference between the image
with and without the UV filter, while the haze-2a filter really did
a good job. What I saw with my eye was nearly the same as the
third image.

The most significant benefit of the haze-2a filter is that it
suppresses haze and purple fringes very effectively. This is
evident in the third row (cloud picked as white reference). The
haze-2a filter also made the image a little warmer and the green
color more vivid.

In summary, it seems the X3 sensor in SD9 does intensify the blue
color of the sky, which sometimes is a good thing. The haze-2a
filter probably reduces the sky color, but this is very easy to
adjust on the computer (for example, using the magic wand tool in
photoshop to select the sky). Overall, the use of a haze-2a filter
on SD9 seems to be pretty advantageous. However, why didn't Sigma
use haze-2a filter glass to make the dust protector? Is there any
other major disadvantage of the haze-2a filter? I'll do more
tests, but any input will be greatly appreciated.

Yi

DonP2 Senior Member • Posts: 1,212
Re: SD9 UV / haze-2a filter test (1)

Thank you, thank you, there is a God after all.

This is exactly what I have been tell you guys all along. Somehow, the pics while impressing me with the film-like nature ( I am spoiled by HDTV ), their colors lack something. And as I said in one of the previous thread, even the action pics look "dead".

I am sorry, a lot of you will be angry. But something we must face to find remedies.

OP Yi Ding Senior Member • Posts: 1,371
color is easily adjustable

It is so easy to adjust color in SPP or other software, and in the pursuit of satisfatory colors, one does not need to restrict oneself to the "correct" white balance. You've posted a lot of 10D photos whose color you liked, but do you know if the color was adjusted on the computer?

On the other hand, most users of other cameras simply use the JPEG output instead of the raw file, and it makes sense to compare different camera's JPEG output, but this doesn't apply to SD9. It is the SD9 users' burden to produce a JPEG or TIFF image out of a raw file, and this is a very personal and subjective process. Suppose you see an SD9 photo with unsatisfactory color to you, I believe in most cases you will be able to achieve satisfactory color if you have the opportunity to use Sigma Photo Pro to process the raw file.

BTW, I think all digital cameras have the purple fringe problem to some extent. This is harder to adjust on the computer, and a haze-2a filter really helps, but the same thing applies to 10D, D100, or whatever.

Yi

DonP wrote:

Thank you, thank you, there is a God after all.

This is exactly what I have been tell you guys all along. Somehow,
the pics while impressing me with the film-like nature ( I am
spoiled by HDTV ), their colors lack something. And as I said in
one of the previous thread, even the action pics look "dead".

I am sorry, a lot of you will be angry. But something we must face
to find remedies.

DonP2 Senior Member • Posts: 1,212
Re: color is easily adjustable

ding1 wrote:

It is so easy to adjust color in SPP or other software, and in the
pursuit of satisfatory colors, one does not need to restrict
oneself to the "correct" white balance. You've posted a lot of 10D
photos whose color you liked, but do you know if the color was
adjusted on the computer?

On the other hand, most users of other cameras simply use the JPEG
output instead of the raw file, and it makes sense to compare
different camera's JPEG output, but this doesn't apply to SD9. It
is the SD9 users' burden to produce a JPEG or TIFF image out of a
raw file, and this is a very personal and subjective process.
Suppose you see an SD9 photo with unsatisfactory color to you, I
believe in most cases you will be able to achieve satisfactory
color if you have the opportunity to use Sigma Photo Pro to process
the raw file.

BTW, I think all digital cameras have the purple fringe problem to
some extent. This is harder to adjust on the computer, and a
haze-2a filter really helps, but the same thing applies to 10D,
D100, or whatever.

Yi

Yi,

I am so glad you mentioned that it is harder to adjust on the computer. (Yes it's true, that you can adjust the overall color to some extent with RAW files, and with Photoshop and another 80 hours, you probably can manipulate any colors in ONE picture) .

However, the Foveon sensor is more sensitive and it makes the problem mentioned a lot more pronounced than other cameras, which affect the colors of an SD9 pic severely.

I mentioned on the other thread about the colors of E1, and it's true that those pics aren't that great and even the Olympus fans are bashing that camera to its death before relased. But it's the color reproduction I was talking about, and yes, you are starting to see people here attacking E1 with lack of sharpness and blurriness and megapixel etc. I hope they would concentrate on the excellent colors.

docmaas
docmaas Veteran Member • Posts: 7,023
Re: color is easily adjustable

DonP wrote:

ding1 wrote:

It is so easy to adjust color in SPP or other software, and in the
pursuit of satisfatory colors, one does not need to restrict
oneself to the "correct" white balance. You've posted a lot of 10D
photos whose color you liked, but do you know if the color was
adjusted on the computer?

On the other hand, most users of other cameras simply use the JPEG
output instead of the raw file, and it makes sense to compare
different camera's JPEG output, but this doesn't apply to SD9. It
is the SD9 users' burden to produce a JPEG or TIFF image out of a
raw file, and this is a very personal and subjective process.
Suppose you see an SD9 photo with unsatisfactory color to you, I
believe in most cases you will be able to achieve satisfactory
color if you have the opportunity to use Sigma Photo Pro to process
the raw file.

BTW, I think all digital cameras have the purple fringe problem to
some extent. This is harder to adjust on the computer, and a
haze-2a filter really helps, but the same thing applies to 10D,
D100, or whatever.

Yi

Yi,

Hi Don,

I think you have a valid point here. Foveon appears to me to be much more sensitive to raw light and overexposes much more easily which in turn often results in a "thin" quality to its color.

However, it's not really a matter of adjusting one color as much as it is changing the quality of the light overall IMHO. One way to do this is with filters. It might take a lot of time in photoshop but there are applications like filtersim where it is a pushbutton operation.

All photography demands operator intervention -- that's why it's an art. If it only took a good camera to take a good picture there wouldn't be any really good photographers because all photographers with appropriate equipment would be great! In other words it's not the camera that is responsible for good pictures but the photographers and photography doesn't end with composition it involves a lot of labwork too whether using chemicals or pixels.

Mike

I am so glad you mentioned that it is harder to adjust on the
computer. (Yes it's true, that you can adjust the overall color to
some extent with RAW files, and with Photoshop and another 80
hours, you probably can manipulate any colors in ONE picture) .

However, the Foveon sensor is more sensitive and it makes the
problem mentioned a lot more pronounced than other cameras, which
affect the colors of an SD9 pic severely.

I mentioned on the other thread about the colors of E1, and it's
true that those pics aren't that great and even the Olympus fans
are bashing that camera to its death before relased. But it's the
color reproduction I was talking about, and yes, you are starting
to see people here attacking E1 with lack of sharpness and
blurriness and megapixel etc. I hope they would concentrate on the
excellent colors.

Brian Slater Regular Member • Posts: 408
Re: color is easily adjustable

DonP wrote:

BTW, I think all digital cameras have the purple fringe problem to
some extent. This is harder to adjust on the computer, and a
haze-2a filter really helps, but the same thing applies to 10D,
D100, or whatever.

Yi

I mentioned on the other thread about the colors of E1, and it's
true that those pics aren't that great and even the Olympus fans
are bashing that camera to its death before relased. But it's the
color reproduction I was talking about, and yes, you are starting
to see people here attacking E1 with lack of sharpness and
blurriness and megapixel etc. I hope they would concentrate on the
excellent colors.

Are you still going to get an E1 now that you have seen some images?
http://forums.dpreview.com/forums/read.asp?forum=1022&message=5572115

 Brian Slater's gear list:Brian Slater's gear list
Canon EOS 6D Canon EOS M6 Canon EF 17-40mm f/4.0L USM Canon EF 70-200mm F4L USM Canon EF-S 60mm f/2.8 Macro USM +4 more
OP Yi Ding Senior Member • Posts: 1,371
Re: color is easily adjustable

DonP wrote:

However, the Foveon sensor is more sensitive and it makes the
problem mentioned a lot more pronounced than other cameras, which
affect the colors of an SD9 pic severely.

You are talking about the purple fringe, I believe. Actually if you browse though the SD9 pictures on pbase, you won't find much evidence of purple fringe, because this mostly happens under very high contrast lighting conditions (like the backlit samples I provided). What's more, if you downsize a photo, the small fringe may disappear! I used to have an Olympus c2100 that also showed purple fringe under similar conditions. I have no idea about 10D or D100, but almost all their sample photos are downsized and purple fringes, if any, have been hidden.

I mentioned on the other thread about the colors of E1, and it's
true that those pics aren't that great and even the Olympus fans
are bashing that camera to its death before relased. But it's the
color reproduction I was talking about, and yes, you are starting
to see people here attacking E1 with lack of sharpness and
blurriness and megapixel etc. I hope they would concentrate on the
excellent colors.

I don't know if those E-1 sample photos have been post-processed to enhance the colors, but let's assume that the colors was the same as the original JPEG output, then this is surely an advantage. All SD9 users have the burden of image processing using Sigma Photo Pro. This can be fun, but you may not like it. On the other hand, other cameras' users have the burden of downsizing their pictures for better "sharpness".

Do you really believe SD9 couldn't produce excellent colors? I think I do have some flower photos whose colors rival those of the E-1 photos. I'll try to find and post some later.

Another issue is color accuracy. Sometimes the seemingly vivid colors is not the truth, and when accuracy is prefered, the colors might not look that good. Without the constraint of color accuracy, vivid color is very easy to achieve using any software.

BTW, when the term "color reproduction" implies accuracy, but I think what you really want are visually pleasing photos that are not necessarily accurate in color reproduction, right?

Yi

OP Yi Ding Senior Member • Posts: 1,371
skin tone and flower

I did some limited tests yesterday but haven't reached a conclusion.

For example, for some colorful plants under morning sunlight, when the white balance was set to auto, the images without any filter and with a UV filter wouldn't need any further adjustment in colors, but the one with a haze-2a filter showed a slight yellow cast that had to be compensated with blue. When the WB was set to sunlight, all three showed some yellow cast that is smallest without any filter and biggest with the haze-2a filter. Using the color wheel, the three images can be made indistinguishable.

Another red flower under slightly cloudy sky didn't show any noticeably difference under the three filter settings (without filter, with UV filter, and with haze-2a filter).

About skin tone, the friend who posed for me requested I should not post any of his images online, so I can't provide any samples. I asked him to hold a gray card, but strangely, picking the reference on the card resulted in very bad color, which may be because the test was done an hour before sunset. Using auto white balance, I can't tell any difference in skin tone, but my eyes might not be sensitive to small changes.

I'd suggest you get a haze-2a filter for one of your lenses and carry out some tests as well.

Yi

sunny1267 wrote:

Very useful test, thank you so much!

I would love to see how the haze-2a affects skin tones (if at all)
and macro flower shots, if you have an opportunity....

Sunny

kalahari Senior Member • Posts: 1,099
purple fringe -- 10D

You can get purple fringe on 10D under high contrast situations. Here is an example:

ding1 wrote:

DonP wrote:

However, the Foveon sensor is more sensitive and it makes the
problem mentioned a lot more pronounced than other cameras, which
affect the colors of an SD9 pic severely.

You are talking about the purple fringe, I believe. Actually if
you browse though the SD9 pictures on pbase, you won't find much
evidence of purple fringe, because this mostly happens under very
high contrast lighting conditions (like the backlit samples I
provided). What's more, if you downsize a photo, the small fringe
may disappear! I used to have an Olympus c2100 that also showed
purple fringe under similar conditions. I have no idea about 10D
or D100, but almost all their sample photos are downsized and
purple fringes, if any, have been hidden.

I mentioned on the other thread about the colors of E1, and it's
true that those pics aren't that great and even the Olympus fans
are bashing that camera to its death before relased. But it's the
color reproduction I was talking about, and yes, you are starting
to see people here attacking E1 with lack of sharpness and
blurriness and megapixel etc. I hope they would concentrate on the
excellent colors.

I don't know if those E-1 sample photos have been post-processed to
enhance the colors, but let's assume that the colors was the same
as the original JPEG output, then this is surely an advantage. All
SD9 users have the burden of image processing using Sigma Photo
Pro. This can be fun, but you may not like it. On the other hand,
other cameras' users have the burden of downsizing their pictures
for better "sharpness".

Do you really believe SD9 couldn't produce excellent colors? I
think I do have some flower photos whose colors rival those of the
E-1 photos. I'll try to find and post some later.

Another issue is color accuracy. Sometimes the seemingly vivid
colors is not the truth, and when accuracy is prefered, the colors
might not look that good. Without the constraint of color
accuracy, vivid color is very easy to achieve using any software.

BTW, when the term "color reproduction" implies accuracy, but I
think what you really want are visually pleasing photos that are
not necessarily accurate in color reproduction, right?

Yi

 kalahari's gear list:kalahari's gear list
Canon EOS RP
Keyboard shortcuts:
FForum MMy threads