PDA

View Full Version : Digital Center Filters



Jim Andrada
18-Jan-2012, 00:40
I was looking at the Schneider digital center filter list - nice idea but very pricey - and only for Schneider wide angles.

Has anyone played with making their own digital center filter for use in Photoshop? At first glance it doesn't seem to be such a difficult thing to do but first glances are often (always???) deceiving.

Sevo
18-Jan-2012, 01:23
Simply take a photograph of a perfectly evenly lit surface (at the magnification ratio you want to photograph at), scan, spot, invert and blur it, and use that as a exposure overlay...

Collas
18-Jan-2012, 01:53
In the Lens Corrections section of ACR, the Lens Vignetting settings should do the trick. I use ACR 4.6 so there may new tools to play with.

Nick

Jim Andrada
18-Jan-2012, 01:58
Yes, that's exactly what I was thinking - but I wonder if it really is that easy.

Jim Andrada
18-Jan-2012, 09:34
I guess what bothers me aout it it that I can't help thinking that any correction after the fact would have to reduce the dynamic range of the photograph whereas a true center filter would not. Thoughts???

Sevo
18-Jan-2012, 09:51
I guess what bothers me aout it it that I can't help thinking that any correction after the fact would have to reduce the dynamic range of the photograph whereas a true center filter would not. Thoughts???

Right - modern negative film tends to have more range than most output media, so it might not matter in many cases. But if you are after squeezing as much out of a image as possible, it might be wiser to combine some "wrong" centre filter with a correction of its over- or undershoot in postprocessing rather than to attempt to do it all in the post.

That could actually be a better Schneider product than a all digital centre filter. A standardized centre filter along with a digital correction filter that matches it to each lens could be considerably cheaper than individualized filters. (Schneider engineers: If you read this and build it, you owe me a freebie!)

Bob Salomon - HP Marketing
18-Jan-2012, 10:16
" A standardized centre filter "

And how would you propose that a "standard" one be made when current Rodenstock center filters range in size from 58 to 86mm diameter screw-in and some, like the 67mm ones, are available in two different center densities whose choice of use depends on the coverage of the lenses. 75 4.5 and 90mm 6.8 require a center with a 1.5x correction and the 35, 45 and 55mm 4.5 lenses require a center density of 2.5x.

You might want to read this about the newest type of center filter design:

http://www.rodenstock-photo.com/mediabase/original/Filter_brochure_Center_2010_CD_Engl_10184.pdf

Jim Andrada
18-Jan-2012, 10:52
Well,if you're scanning and printing digitally you'd really only have to "Dodge and Burn" once in PS - although each time you print you might WANT to dodge and burn differently because each time you might feel a little differently about the print.

I guess my disinclination to buy a center filter arises from a) having more than one size lens and b) really not using the extreme wide angles tremendously often. Often enough to have the lens, of course, but still a small percentage of my usage.

If I were doing a high percentage of work with the superwides, no doubt about it, I'd have the filter(s), particularly if I were using extensive rise/fall etc where I'd be working further into the corners and offset from the center of the frame.

And I might actually convince myself to get one - I'd like to have them, just a bit cost averse at the moment.

Jim Andrada
18-Jan-2012, 11:00
By the way, thanks Bob - The link you provided was very interesting.

Sevo
18-Jan-2012, 11:06
By the way, do they do these absorptive glass filters for obsolete lenses too?

Brian Ellis
18-Jan-2012, 11:08
In the Lens Corrections section of ACR, the Lens Vignetting settings should do the trick. I use ACR 4.6 so there may new tools to play with.

Nick

I perhaps don't know how to use the Lens Vignetting settings correctly but they haven't worked well for me. Most of the time I don't have perfectly even vignetting, i.e. the same amount on both sides. Usually one side (typically a sky) has more than the other or the other side has none at all. Again, it's probably me but the Lens Vignetting settings seem to work well only when both sides are equally dark.

I'm certainly open to learning how they can be used when, for example, there's only vignetting on one side. FWIW, I now use the gradient tool to deal with that situation but while it's an improvement it's not perfect.

Sevo
18-Jan-2012, 11:32
75 4.5 and 90mm 6.8 require a center with a 1.5x correction and the 35, 45 and 55mm 4.5 lenses require a center density of 2.5x.


Well, it might take a filter per size, if the small ones should not be the flat-sloped ones. But given your figures it should be possible to design a joint centre filter that is the best compromise of all slopes, and maybe even sizes, with an deviation of less than one stop even for the least well matched lens. It would have higher order falloff than without a filter, but we'd gain at least 1.5 stops of dynamic range for the worst exposed part of the image. Higher order errors are quite undesirable in the analogue domain (hence mismatching centre filters is a stupid idea if there is no digital PP involved) but digitally filtering them away is easy.

Drew Wiley
18-Jan-2012, 12:20
Merely depending on the exp "latitude" of the film, and expecting to correct in either
PS or dodge/burn conventionally is seldom an ideal situation. Underexposed or overexp
areas of a color neg might print detail, but not optimal color balance due to shifts in shadows or highlights which cannot be corrected relative to the ideally exposed areas of the same neg. One and a half or up to three stops of difference is really a lot in color work. And unless I deliberately wanted a falloff look, I'd even prefer to use a CF for black and white film when a harsh lighting ratio is encountered. I've never seen
anything resembling color fringing or image degredation with my own CF's, but the
problem might be much greater with small digital capture surfaces sensitive to non-parallel rays than with typical LF film applications.

Bob Salomon - HP Marketing
18-Jan-2012, 13:27
By the way, do they do these absorptive glass filters for obsolete lenses too?

No, only for current Rodenstock lenses, but if your old lens has the same filter thread then they may work. They do for older Rodenstock's with the same thread.

Bob Salomon - HP Marketing
18-Jan-2012, 13:28
I perhaps don't know how to use the Lens Vignetting settings correctly but they haven't worked well for me. Most of the time I don't have perfectly even vignetting, i.e. the same amount on both sides. Usually one side (typically a sky) has more than the other or the other side has none at all. Again, it's probably me but the Lens Vignetting settings seem to work well only when both sides are equally dark.

I'm certainly open to learning how they can be used when, for example, there's only vignetting on one side. FWIW, I now use the gradient tool to deal with that situation but while it's an improvement it's not perfect.

Center filters work, even with camera movements, since they are on the lens. When you do it in computer it sounds like it is not taking movements (direct or indirect) into consideration.

Bob Salomon - HP Marketing
18-Jan-2012, 13:32
Merely depending on the exp "latitude" of the film, and expecting to correct in either
PS or dodge/burn conventionally is seldom an ideal situation. Underexposed or overexp
areas of a color neg might print detail, but not optimal color balance due to shifts in shadows or highlights which cannot be corrected relative to the ideally exposed areas of the same neg. One and a half or up to three stops of difference is really a lot in color work. And unless I deliberately wanted a falloff look, I'd even prefer to use a CF for black and white film when a harsh lighting ratio is encountered. I've never seen
anything resembling color fringing or image degredation with my own CF's, but the
problem might be much greater with small digital capture surfaces sensitive to non-parallel rays than with typical LF film applications.

Go down to the Embarcadro a shoot one of the aluminum and glass builings so the building fills the long dimension of the film and use an older type center filter. As you go up the building you will see a color shift. That is what the newer ones eliminate.

If you are doing landscapes you probably will not see it. If you do room sets you might.

Jim Andrada
18-Jan-2012, 13:54
Fortunately I do almost 100% B & W. And when it's color it's almost always landscapes or abstract stuff, old buildings, etc

Collas
18-Jan-2012, 14:02
I perhaps don't know how to use the Lens Vignetting settings correctly but they haven't worked well for me. Most of the time I don't have perfectly even vignetting, i.e. the same amount on both sides. Usually one side (typically a sky) has more than the other or the other side has none at all. Again, it's probably me but the Lens Vignetting settings seem to work well only when both sides are equally dark.

I'm certainly open to learning how they can be used when, for example, there's only vignetting on one side. FWIW, I now use the gradient tool to deal with that situation but while it's an improvement it's not perfect.

I've had some success with this with a 1937 Leitz 28mm f6.3 Hektor. I used it on an Epson RD-1 and created a lens profile within ACR, which also corrects the chromatic aberration. But sometimes the vignetting suited the image. The only difficulty with the Epson is that it has no way of writing which lens has been used into the EXIF file, so it can be difficult to remember which lens took the image, although the Hektor has a unique look compared to a much more modern Zeiss 28mm Biogon that I also have.

I do use notes when taking pictures with the Tachihara, but the 35mm and digital cameras don't get the same treatment as they can rattle through film or SD cards so quickly. I wish I had the patience with 35mm that I have with 5 x 4!

Nick

Sevo
18-Jan-2012, 14:59
Center filters work, even with camera movements, since they are on the lens. When you do it in computer it sounds like it is not taking movements (direct or indirect) into consideration.

No, you'd have to write down all movements, and enter the offsets into the filter. Arguably a nuisance and unprofitable when you can still buy a matching centre filter for your lens. But a filter that only corrects the difference to a not perfectly matched filter (intended for another lens) sounds as it if could be quite useful for legacy lens owners.

Drew Wiley
18-Jan-2012, 19:21
Thanks for the tip, Bob. I'll have to try an experiment like that. I used wide-angle lenses in the past mostly for architectural work that had a relatively small degree of
enlargement, if any. Typical commercial stuff. Most of my personal work that gets
seriously enlarged is done with lenses that don't need this kind of correction. The
last few times I've used CF is with black and white film. Thanks for the diagrams.
It's certainly interesting to see this new kind of engineering.

Larry Gebhardt
19-Jan-2012, 08:33
" A standardized centre filter "

And how would you propose that a "standard" one be made when current Rodenstock center filters range in size from 58 to 86mm diameter screw-in and some, like the 67mm ones, are available in two different center densities whose choice of use depends on the coverage of the lenses. 75 4.5 and 90mm 6.8 require a center with a 1.5x correction and the 35, 45 and 55mm 4.5 lenses require a center density of 2.5x.

You might want to read this about the newest type of center filter design:

http://www.rodenstock-photo.com/mediabase/original/Filter_brochure_Center_2010_CD_Engl_10184.pdf

Bob, I have a Caltar 90mm f/6.8 (Rodenstock made I believe) and a Nikon 75mm f/4.5 SW. Both have a 67mm filter thread. Do you know if the Rodenstock filter would be the correct one for the Nikon as well? It would be great to get one to share.

rdenney
19-Jan-2012, 08:50
I have tried to make corrections using Photoshop, but with color film there are other issues. With transparency film, especially Velvia, the falloff often takes the image right out of the dynamic range. With negative films (color or black and white), the linearity of the film might not be sufficient to allow a straight correction without also requiring some correction to contrast. It's just easier to screw on the filter. I'm not sure I've ever had a black-and-white negative that could withstand a three-stop dropoff in the corners without pushing it down into a toe that would affect tonal separation, but then I always seem to be attracted to films with more of a toe and shoulder than, say, the T-grain films.

I don't use one (even with Velvia) with 90mm or longer lenses on 4x5. But I have one for use with the 47/5.6 (for use with 6x9) and 65/8 Super Angulons, which is really the old 49mm center filter for the 65/8. It clips the corners of the 47's view slightly, so I have to choose between the correction and a bit more vignetting in the corners with 6x12. With the 65/5.6 Super Angulon, I use the correct Schneider filter. I bought neither of these new or anywhere near their new price. The filter for the 65/5.6 required a lot of patience to find at a good price.

Rick "whose time for doing photography is limited" Denney

Bob Salomon - HP Marketing
19-Jan-2012, 09:00
Bob, I have a Caltar 90mm f/6.8 (Rodenstock made I believe) and a Nikon 75mm f/4.5 SW. Both have a 67mm filter thread. Do you know if the Rodenstock filter would be the correct one for the Nikon as well? It would be great to get one to share.

Larry, it would definately work with the Caltar. Since the difference in the image circles between the 75 Nikkor and the 75 Grandagon is only 5mm it should also work with your 75 as well. Unfortunately users of the Nikkors don't have much choice since Nikon elected not to supply center filters for their WA lenses even though their "remarkably even coverage" still has the same fall-off as the other wide angles from Rodenstock, Schneider and Fuji.

Bob Salomon - HP Marketing
19-Jan-2012, 09:09
Thanks for the tip, Bob. I'll have to try an experiment like that. I used wide-angle lenses in the past mostly for architectural work that had a relatively small degree of
enlargement, if any. Typical commercial stuff. Most of my personal work that gets
seriously enlarged is done with lenses that don't need this kind of correction. The
last few times I've used CF is with black and white film. Thanks for the diagrams.
It's certainly interesting to see this new kind of engineering.

Drew,

The best way to see it is on a transparency on an even light box. As you go from center to edge you can run into the shift on evenly lit buildings with an even aluminum finish.

Larry Gebhardt
19-Jan-2012, 09:33
Larry, it would definately work with the Caltar. Since the difference in the image circles between the 75 Nikkor and the 75 Grandagon is only 5mm it should also work with your 75 as well. Unfortunately users of the Nikkors don't have much choice since Nikon elected not to supply center filters for their WA lenses even though their "remarkably even coverage" still has the same fall-off as the other wide angles from Rodenstock, Schneider and Fuji.

Thanks, I wondered why I couldn't ever find a centerfilter marketed for the Nikon. Buying one filter to share is certainly nicer than needing two.

It looks like these only come with 67mm rear and 86mm front filters. Is it OK to mount this in front of a 67mm filter without it adversely affecting the function? Will it cause extra vignetting over just the one filter? I don't wish to buy a set of 86mm filters too.

Drew Wiley
19-Jan-2012, 09:55
Larry, I have used and thoroughly tested my Schneider CF's ala Super-Angulon specs
on my preferred Nikon SW series (like the 82CF on 90/4.5), and the performance is
absolutely ideal, at least as far as evenness of field and movements etc. Bob is
commenting about a new set of quality parameters; but if the general field correction is similar to the Schneider I'd imagine the result would be at least as good on a Nikkor.
There seems to be a great deal of similarity in the wides from all the major mfgs,
at least until you get to the extremely short FL's or the newer "XL" type designs.
About a stop and a half gradual correction seems to work for the typical 90's and
120's.

Ari
19-Jan-2012, 10:30
I use the Schneider digital CF with the SA 72XL; actually, I very rarely use it, I find the lens is great as-is, but the few times I've applied the filter, it balanced things out nicely and subtly.
If you've bought a new Schneider lens in the last 4-5 years (don't quote me on that timeline), it's a free download, provided you still have the box or other documentation with the serial codes.
Go here: http://www.digitalcenterfilter.com/

Bob Salomon - HP Marketing
19-Jan-2012, 10:33
Thanks, I wondered why I couldn't ever find a centerfilter marketed for the Nikon. Buying one filter to share is certainly nicer than needing two.

It looks like these only come with 67mm rear and 86mm front filters. Is it OK to mount this in front of a 67mm filter without it adversely affecting the function? Will it cause extra vignetting over just the one filter? I don't wish to buy a set of 86mm filters too.

No, the center filter must go on the lens first. Then whatever other filter you want to use goes on the front of a center filter. Otherwise you will not get the proper correction. In addition the CF requires that the lens be stopped down at least 2 stops. A CF will not do anything wide open or one stop down. For the 90mm 6.8 and a 75mm 4.5 you need the 1.5x filter, 170002. The 2.5x will not work properly on your lenses. The good news is that the 1.5x is almost 50% less then the 2.5x version!

Larry Gebhardt
19-Jan-2012, 10:56
No, the center filter must go on the lens first. Then whatever other filter you want to use goes on the front of a center filter. Otherwise you will not get the proper correction. In addition the CF requires that the lens be stopped down at least 2 stops. A CF will not do anything wide open or one stop down. For the 90mm 6.8 and a 75mm 4.5 you need the 1.5x filter, 170002. The 2.5x will not work properly on your lenses. The good news is that the 1.5x is almost 50% less then the 2.5x version!

Thanks again Bob, The 170002 is the one I was looking at. I guess I'll need to factor in the cost of a few larger black and white filters as well.