Improving the D7 (and D7i?)

Bill Cook · May 16, 2002

The D7 either has no hot mirror filter (to eliminate IR responce) or a very poor one. This lack can cause some real difficulties. One is illustrated with the pair of pictures below

The picture was taken at a high school play that my granddaughter was in. The top picture was processed through DIVU converting it to sRGB and size reduced and saved for the web in Photoshop 7.0

The one below is the same picture with a minus 20 cyan --- red correction in both the shadow and the mid tones (eg shifted toward the cyan). This is the maximum amount of correction that could be applied without clipping. The young ladies dress was black. The reason it appears brown in the top picture is because many synthetic fibers (like leaves) have a very high IR reflection.

Having observed this I obtained a 67mm hot mirror from Amazon (they have a clearance at $40 each - three left) and a 49mm to 67mm step up ring. I now have the hot mirror on my camera. Though I have not had another chance to take pictures in the theater (I only got the filter on Monday) I have notice two other very interesting effects. 1) The camera focuses in sub 1 second time consistently even with the lens in the 200mm focal length equivalent regular or macro. 2) The images seem to have sharper edge (and perhaps over all sharper) appearance.

This would make sense because most lenses do not bring visible light into focus at the same point as IR. Indeed they can be rather far apart. That being the case since the D7 uses contrast focusing on the CCD the image probably seems a bit fuzzy (or seeing double). Of course since the IR is making a significant contribution to the exposure and it is not quite in focus with the visiible light it makes the image look a bit less sharp.

In any event you might want to try it.

Bill Cook

newsposter · May 16, 2002

What kind of fstop hit are you taking?

I tried the Tiffen a while back and the hit was about 1 full stop.

Bryan Biggers · May 16, 2002

When I tested a Tiffen hot mirror filter, I could find no discernable effect on the D7 under vs. lighting conditions. It will be interesting to see how the one that Bill got works. Bryan

newsposter said:
What kind of fstop hit are you taking?

I tried the Tiffen a while back and the hit was about 1 full stop.

Jawed49152 · May 16, 2002

Bill great work.

Your observations about focus performance are truly stunning. Dare I say it, it should have been obvious - pity no one seems to have realised this in the past (or I missed it...)

Damn this means the D7's autofocus was hobbled by a quite MAJOR design error by Minolta. That really should be very embarrassing for the design department.

What menu item in Photoshop (?) did you use to achieve your cyan adjustment? Also did you adjust cyan in the red channel, or cyan in all channels (maybe a redundant question)?

Will be fascinated to see some in-depth comparisons, especially with regard to image sharpness. That, too, constitutes a major blunder.

Jawed

Bill Cook said:
The D7 either has no hot mirror filter (to eliminate IR responce)
or a very poor one. This lack can cause some real difficulties. One
is illustrated with the pair of pictures below

The picture was taken at a high school play that my granddaughter
was in. The top picture was processed through DIVU converting it to
sRGB and size reduced and saved for the web in Photoshop 7.0

The one below is the same picture with a minus 20 cyan --- red
correction in both the shadow and the mid tones (eg shifted toward
the cyan). This is the maximum amount of correction that could be
applied without clipping. The young ladies dress was black. The
reason it appears brown in the top picture is because many
synthetic fibers (like leaves) have a very high IR reflection.

Having observed this I obtained a 67mm hot mirror from Amazon (they
have a clearance at $40 each - three left) and a 49mm to 67mm step
up ring. I now have the hot mirror on my camera. Though I have not
had another chance to take pictures in the theater (I only got the
filter on Monday) I have notice two other very interesting effects.
1) The camera focuses in sub 1 second time consistently even with
the lens in the 200mm focal length equivalent regular or macro. 2)
The images seem to have sharper edge (and perhaps over all sharper)
appearance.

This would make sense because most lenses do not bring visible
light into focus at the same point as IR. Indeed they can be rather
far apart. That being the case since the D7 uses contrast focusing
on the CCD the image probably seems a bit fuzzy (or seeing double).
Of course since the IR is making a significant contribution to the
exposure and it is not quite in focus with the visiible light it
makes the image look a bit less sharp.

In any event you might want to try it.

Bill Cook

newsposter · May 16, 2002

Bill, can you tell what firmware you have in your D7?

newsposter · May 16, 2002

With just a little research, it appears that Kodak recommends the use of an IR filter (not mirror) for their ITO sensor cameras. This incudes the high-end SLR type as well as a few of the low-end. Kodak at one point even had their own branded filters as for-sale accessories.

Donald Hutton · May 16, 2002

These D7'ers are a bunch of great thinkers, that's for sure - I am very intrigued with developments on this train of thought - please keep us informed.

Jawed said:
Your observations about focus performance are truly stunning. Dare
I say it, it should have been obvious - pity no one seems to have
realised this in the past (or I missed it...)

Damn this means the D7's autofocus was hobbled by a quite MAJOR
design error by Minolta. That really should be very embarrassing
for the design department.

What menu item in Photoshop (?) did you use to achieve your cyan
adjustment? Also did you adjust cyan in the red channel, or cyan
in all channels (maybe a redundant question)?

Will be fascinated to see some in-depth comparisons, especially
with regard to image sharpness. That, too, constitutes a major
blunder.

Jawed

Bill Cook said:

The D7 either has no hot mirror filter (to eliminate IR responce)
or a very poor one. This lack can cause some real difficulties. One
is illustrated with the pair of pictures below

The picture was taken at a high school play that my granddaughter
was in. The top picture was processed through DIVU converting it to
sRGB and size reduced and saved for the web in Photoshop 7.0

The one below is the same picture with a minus 20 cyan --- red
correction in both the shadow and the mid tones (eg shifted toward
the cyan). This is the maximum amount of correction that could be
applied without clipping. The young ladies dress was black. The
reason it appears brown in the top picture is because many
synthetic fibers (like leaves) have a very high IR reflection.

Having observed this I obtained a 67mm hot mirror from Amazon (they
have a clearance at $40 each - three left) and a 49mm to 67mm step
up ring. I now have the hot mirror on my camera. Though I have not
had another chance to take pictures in the theater (I only got the
filter on Monday) I have notice two other very interesting effects.
1) The camera focuses in sub 1 second time consistently even with
the lens in the 200mm focal length equivalent regular or macro. 2)
The images seem to have sharper edge (and perhaps over all sharper)
appearance.

This would make sense because most lenses do not bring visible
light into focus at the same point as IR. Indeed they can be rather
far apart. That being the case since the D7 uses contrast focusing
on the CCD the image probably seems a bit fuzzy (or seeing double).
Of course since the IR is making a significant contribution to the
exposure and it is not quite in focus with the visiible light it
makes the image look a bit less sharp.

In any event you might want to try it.

Bill Cook

Click to expand...

PeteGanzel · May 16, 2002

Bill:

Some great observations. The D-7 has an apochromatic optic. I'm not sure if this extends to the IR range though. If not, the difference in focus

is very tiny at the wide setting (7mm) and similar to a 50mm normal lens on a 35mm camera (most are not apochromats). The adjustment would be setting the manual focus to about 20m to get infinity focus. This is within the range of the depth of focus of the lens, especially if stopped down.

Of course if you did this, the focus would not be optimum in the middle of the visible spectrum. The near red IR sensitivity would show up as "color fringing" if the lens were not apochromatic in this range, although I'm not really sure how the camera interprets invisible radiation as visible to us. I really don't see how all of this can significantly affect autofocus unless you were using an IR filter and excluding all visible light.

There's no question about the difficulty with black clothing though and it appears to be related to IR sensitivity.

Thanks and Good Luck

Pete

Bill Cook said:
The D7 either has no hot mirror filter (to eliminate IR responce)
or a very poor one. This lack can cause some real difficulties. One
is illustrated with the pair of pictures below

The picture was taken at a high school play that my granddaughter
was in. The top picture was processed through DIVU converting it to
sRGB and size reduced and saved for the web in Photoshop 7.0

The one below is the same picture with a minus 20 cyan --- red
correction in both the shadow and the mid tones (eg shifted toward
the cyan). This is the maximum amount of correction that could be
applied without clipping. The young ladies dress was black. The
reason it appears brown in the top picture is because many
synthetic fibers (like leaves) have a very high IR reflection.

Having observed this I obtained a 67mm hot mirror from Amazon (they
have a clearance at $40 each - three left) and a 49mm to 67mm step
up ring. I now have the hot mirror on my camera. Though I have not
had another chance to take pictures in the theater (I only got the
filter on Monday) I have notice two other very interesting effects.
1) The camera focuses in sub 1 second time consistently even with
the lens in the 200mm focal length equivalent regular or macro. 2)
The images seem to have sharper edge (and perhaps over all sharper)
appearance.

This would make sense because most lenses do not bring visible
light into focus at the same point as IR. Indeed they can be rather
far apart. That being the case since the D7 uses contrast focusing
on the CCD the image probably seems a bit fuzzy (or seeing double).
Of course since the IR is making a significant contribution to the
exposure and it is not quite in focus with the visiible light it
makes the image look a bit less sharp.

In any event you might want to try it.

Bill Cook

Scott Martinez · May 16, 2002

I am thinking you might have had a wrong white balance setting?
Thoughts?
-Scott

Bill Cook said:
The D7 either has no hot mirror filter (to eliminate IR responce)
or a very poor one. This lack can cause some real difficulties. One
is illustrated with the pair of pictures below

The picture was taken at a high school play that my granddaughter
was in. The top picture was processed through DIVU converting it to
sRGB and size reduced and saved for the web in Photoshop 7.0

The one below is the same picture with a minus 20 cyan --- red
correction in both the shadow and the mid tones (eg shifted toward
the cyan). This is the maximum amount of correction that could be
applied without clipping. The young ladies dress was black. The
reason it appears brown in the top picture is because many
synthetic fibers (like leaves) have a very high IR reflection.

Having observed this I obtained a 67mm hot mirror from Amazon (they
have a clearance at $40 each - three left) and a 49mm to 67mm step
up ring. I now have the hot mirror on my camera. Though I have not
had another chance to take pictures in the theater (I only got the
filter on Monday) I have notice two other very interesting effects.
1) The camera focuses in sub 1 second time consistently even with
the lens in the 200mm focal length equivalent regular or macro. 2)
The images seem to have sharper edge (and perhaps over all sharper)
appearance.

This would make sense because most lenses do not bring visible
light into focus at the same point as IR. Indeed they can be rather
far apart. That being the case since the D7 uses contrast focusing
on the CCD the image probably seems a bit fuzzy (or seeing double).
Of course since the IR is making a significant contribution to the
exposure and it is not quite in focus with the visiible light it
makes the image look a bit less sharp.

In any event you might want to try it.

Bill Cook

--
-Scott

PeteGanzel · May 16, 2002

Scott:

I thought this too, but check out the difference between the dress and shoes. Both probably appeared very black to the eye at the time.

Good Luck,

Pete

Scott Martinez said:
Bill Cook said:

The D7 either has no hot mirror filter (to eliminate IR responce)
or a very poor one. This lack can cause some real difficulties. One
is illustrated with the pair of pictures below

The picture was taken at a high school play that my granddaughter
was in. The top picture was processed through DIVU converting it to
sRGB and size reduced and saved for the web in Photoshop 7.0

The one below is the same picture with a minus 20 cyan --- red
correction in both the shadow and the mid tones (eg shifted toward
the cyan). This is the maximum amount of correction that could be
applied without clipping. The young ladies dress was black. The
reason it appears brown in the top picture is because many
synthetic fibers (like leaves) have a very high IR reflection.

Having observed this I obtained a 67mm hot mirror from Amazon (they
have a clearance at $40 each - three left) and a 49mm to 67mm step
up ring. I now have the hot mirror on my camera. Though I have not
had another chance to take pictures in the theater (I only got the
filter on Monday) I have notice two other very interesting effects.
1) The camera focuses in sub 1 second time consistently even with
the lens in the 200mm focal length equivalent regular or macro. 2)
The images seem to have sharper edge (and perhaps over all sharper)
appearance.

This would make sense because most lenses do not bring visible
light into focus at the same point as IR. Indeed they can be rather
far apart. That being the case since the D7 uses contrast focusing
on the CCD the image probably seems a bit fuzzy (or seeing double).
Of course since the IR is making a significant contribution to the
exposure and it is not quite in focus with the visiible light it
makes the image look a bit less sharp.

In any event you might want to try it.

Bill Cook

Click to expand...

--
-Scott

Scott Martinez · May 17, 2002

Yeah I see that now. As a matter of fact I went back and looked at pictures from a high school dance I shot and see the blacks are many different shades depending on the reflectivity. I wondered about that but was really trying to get decent skin tones on my prints at the time and did not give it too much thought.

Two samples shot under constant quartz lighting. White balanced manual with a white card.

-Scott

PeteGanzel said:
Scott:

I thought this too, but check out the difference between the dress
and shoes. Both probably appeared very black to the eye at the
time.

Good Luck,

Pete

Bill Cook · May 17, 2002

I don't think that is the problem. Three reasons. First I took roughly half the shots using the built in tungsten white balance and the remainder with a custom white balance from the white side of a Kodak 18% grey card. All of the shots with the black clothing should the same problem. Second all of theskin tones look reasonably correct. Third clothing which is truly black (and the dress was) should be black independent of white balance.

Finally I did a web search on the subject and the IR reflection of synthetic fibers and the problems photographing them are well documented.

Bill Cook

Scott Martinez said:
I am thinking you might have had a wrong white balance setting?
Thoughts?
-Scott

Bill Cook · May 17, 2002

A good question, to which there is no easy answer. I am using a Tiffen Standard Hot Mirror. When I put the camera in P mode and take the same image (a group of dense trees with open shy above) with the camera on a tripod (so the framing doesn't change) without the filter the exposure was 1/750 at f8.0. With the filter the exposure was 1/500 at f8.0. But the image, particularly the sky looked some what lighter. When I did a manual exposure at 1/750 at f8.0 the images were virtually the same density. My conclusion is that the IR affects the exposure calculation. I have since adjusted the camera for a -0.6 exposure correction (I had been using a -0.3) and the images seem to have comparable density. I will be doing more extensive sensiometric comparisons but it will have to wait until next week as my granddaughter is graduating from High School this weekend and I intend to do the grandpa thing. (Can you say Party?)

Bill Cook

newsposter said:
What kind of fstop hit are you taking?

I tried the Tiffen a while back and the hit was about 1 full stop.

Bill Cook · May 17, 2002

I have tried both the V121 and the V122 and have not noticed any real difference in this test. I have access to the V122 from my dealer largely because of the concerns about CF failure under V121. I have to say in this shop anyway we have not seen a CF which fails under V121 that does not also fail under V122. But our sampling is much to small for statistical comparisons.

newsposter said:
Bill, can you tell what firmware you have in your D7?

Bill Cook · May 17, 2002

Jawed said:
Bill great work.

Your observations about focus performance are truly stunning. Dare
I say it, it should have been obvious - pity no one seems to have
realised this in the past (or I missed it...)

Damn this means the D7's autofocus was hobbled by a quite MAJOR
design error by Minolta. That really should be very embarrassing
for the design department.

Not necessarily. The decison to include a hot mirror or not is not an easy decision. For one thing they are not cheap (a 67mm normally costs about $140) and although they certainly could use a smaller one internally it still would add significant cost. Also the inclusion of such filter would effectively preclude the use of the camera for IR photography and there are those out there that think that the mark of a good camera is the ability to do IR work. Admiitedly IR photos can be very beautiful. Finally the hot mirror is something that you can add after the fact (removing it would be a near impossible task). It should be noted that many pro digital SLRs don't have one either and Kodak recommends there consistent use except when deliberately trying to do IR phtography. Is it a design flaw? Guess it depends on whose Ox you are willing to gore. It would have been nice if Minolta had noted the lack in a foot note to their documentation though.

Jawed49152 said:
What menu item in Photoshop (?) did you use to achieve your cyan
adjustment? Also did you adjust cyan in the red channel, or cyan
in all channels (maybe a redundant question)?

Image/Adjustment/Color Balance in the Mac version anyway.I only adjusted the cyan/red slider. I don't know whether it affects only the Red channel but that is my assumption.

Jawed49152 said:
Will be fascinated to see some in-depth comparisons, especially
with regard to image sharpness. That, too, constitutes a major
blunder.

Next week. I am busy being grandpa this weekend.

Bill Cook

Bill Cook · May 17, 2002

I know of no lens design that is truly apochromatic from IR through UV. I suspect that such a lens would be very expensive indeed. All prosumer digital cameras show some color fringing at high contrast areas against the sky, such as a tree limb or power line. Admittedly the Minoltas show some of the least but it still is observable (check Phils comparisons). Your theory is as good as any in explaining them. Even so the apparent change in edge sharpness could also be explained in the same way.

In any event I was looking for or expecting any improvement in focus speed or sharpness I was only trying to get a black dress to be black. The rest was serendipity, a bonus. I'll take it.

Bill Cook

PeteGanzel said:
Bill:

Some great observations. The D-7 has an apochromatic optic. I'm
not sure if this extends to the IR range though. If not, the
difference in focus
is very tiny at the wide setting (7mm) and similar to a 50mm
normal lens on a 35mm camera (most are not apochromats). The
adjustment would be setting the manual focus to about 20m to get
infinity focus. This is within the range of the depth of focus of
the lens, especially if stopped down.
Of course if you did this, the focus would not be optimum in the
middle of the visible spectrum. The near red IR sensitivity would
show up as "color fringing" if the lens were not apochromatic in
this range, although I'm not really sure how the camera interprets
invisible radiation as visible to us. I really don't see how all
of this can significantly affect autofocus unless you were using
an IR filter and excluding all visible light.
There's no question about the difficulty with black clothing though
and it appears to be related to IR sensitivity.

Thanks and Good Luck

Pete

Jawed49152 · May 17, 2002

Bill, all other consumer digicams seem to have effective IR filtering. I don't notice the "red clothes" problem in pix from any other digicam. Sure I stand to be corrected, but I just haven't seen them. The Dimage X has effective filtering, from the samples I've seen (or its sensor's IR sensitivity is very poor, which seems doubtful).

Bill Cook said:
would effectively preclude the use of the
camera for IR photography and there are those out there that think
that the mark of a good camera is the ability to do IR work

I'll happily round up these idiots and kill them all to do away with this irrelevance. The ability to perform IR photography, has with the D7, gotten the baby's legs mashed in the waste disposal. Your experimental evidence of the side-effects of the D7's IR sensitivity is quite damning.

If people want to do IR photography that bad, they can buy the Sony F707 which has an internal IR filter that can be moved out of the way of the sensor.

Early pro digital cameras didn't have one, I know that, as I've read some old reviews (on other websites) which talk about this explicitly. Modern cameras have one built-in, I'm reasonably sure, as their pictures don't seem to suffer these effects. It would be nice if Phil could investigate this question as so far I've not seen any comment from him on this subject in any review (anyone know better?).

One thing's for sure, pix I've seen from the E10 and the F707 do not exhibit this problem. Minolta has supposedly improved IR filtering with the D7i, I certainly hope it solves the red clothes problem.

Thanks for the Color Balance tip. Can't decide whether I like the result with outdoor pictures - I'm inclined towards a moderate amount (-5). Curiously the result looks more like what I consider typical of other digicams.

Thanks for your help,
Jawed

Mountain · May 17, 2002

Question from novice:

Does that mean that I can put the Hot Mirror filter all the time instead of UV filter? (Trying to take advantage of sub-1 secod focusing and coorect colors at the same time!)

Mountain

newsposter · May 17, 2002

they run 5x to 10x the price of a decent UV filter.

I would screw one on in specific situations rather than all the time.

Remember though that the discussion we're having here is based on the considered observations of a single competent user.

Before spending $50 to $100 on a hot mirror filter I would wait for some more users to chime in.

Unless you really want to be one of the theory testers.

Bill Cook · May 17, 2002

In addition I would add that hot mirrors have extremely delicate front coatings. If I decided to leave it on all the time I will likely decide to put some sort of protecting (UV or Skylight) filter in front of it.

Bill Cook

newsposter said:
they run 5x to 10x the price of a decent UV filter.

I would screw one on in specific situations rather than all the time.

Remember though that the discussion we're having here is based on
the considered observations of a single competent user.

Before spending $50 to $100 on a hot mirror filter I would wait for
some more users to chime in.

Unless you really want to be one of the theory testers.

Improving the D7 (and D7i?)

Well-known member

Leading Member

Veteran Member

Senior Member

Leading Member

Leading Member

Leading Member

Senior Member

Active member

Senior Member

Active member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Senior Member

Active member

Leading Member

Well-known member

Similar threads

Keyboard shortcuts