I'm still intrigued but I'm not getting exactly what your selling and how to use it. The PDFs read like a biochemistry article on the subject of focusing and film flatness, I need to see an instruction manual and an actual 1-2-3-4 step process of what it is I'd be using and how I'd use it.
Is there a glass sheet that I tape my 8x10 film onto that I slide into some sort of special film holder you provide?
Where you should be hosting your photos: www.SmugMug.com
The most logical solution to in camera film flatness would probably be a vacuum holder. But certainly not so easy to implement technically. There are small portable low vacuum devices used for picking up small parts and something like that could be adapted for the purposes of holding film flat. One would need to design a special film holder of course but I'm sure that could be accomplished. Probably not useful for Quickloads though.
But given the myriad of variables that degrade image quality under practical field conditions I would question the utility of being really fanatical about film flatness in the field. Sounds more like an exercise necessary for precision scientific photography.
In fact I used the vacuum hold down approach for the photo documentation of IC chips before features became so small as to not be resolvable with standard lenses. The film plate was specially machined with a few hundred tiny holes accessing the vacuum space under the faceplate. The vacuum space was connected to a small hose which was attached to a dry vane pump capable a few tens of mTorr vacuum. I sort of wish I had my design for the device now.
Nate Potter, Austin TX.
A working vacation for photographers can put them on a location where a once in aOriginally Posted by Nathan Potter
lifetime money shot can be taken.
The holy grail of 4x5 photography is a 16bit fluid scan that yields >= 400 mega pixels of an image worthy of high acclaim. The end-product is mural fine art (see my profile for amusement and muse).
In another life I also used high vacuum equipment. I've done a prototype and theIn fact I used the vacuum hold down approach for the photo documentation of IC chips before features became so small as to not be resolvable with standard lenses. The film plate was specially machined with a few hundred tiny holes accessing the vacuum space under the faceplate. The vacuum space was connected to a small hose which was attached to a dry vane pump capable a few tens of mTorr vacuum. I sort of wish I had my design for the device now.
problem encountered was that the film gets rumpled at each hole (with a hole density of about 100 and hole size of about 1/16" ... I have not tried smaller hole sizes (I do have a micro-bit set now) and higher hole densities (200 or 400 - ugh).
Rolling and taping (for film scanning) does hold film perfectly flat. But designing a
a precision holder to accept the plenum/septum is extremely problematic, expensive and time consuming.
In the end it may be less work using a well calibrated pan head with the mural being
a sequence of panels that look like windows. The experience viewing a sectional
verses a continuous display would be different. Perhaps one might be better than the other in different environments.
For sale are
(1) the tools and instructions for taking pictures worthy of a tango (or betterscanning.com flatbed) fluid scan (maybe as high as a giga pixel for 8x10)
(2) the tools and instructions to characterize your lenses (and entire system end-to-end) using the lp/mm MTF metric
There are some many derivative functions and improved capabilities that it is improbable that I could ever describe them all in a way that would satisfy everyone's questions.
Lets just summarize it by saying that the tools and documentation are very powerful
both optically and functionally.
The 1-2-3-4 alignment instruction document is in the works. Its purpose is for bringing the emulsion plane into perfect alignment with the lens image plane. This is done with feeler gauges, glass gauge blocks, stainless steel rule and honed digital calipers with the .0005" touch feature.I need to see an instruction manual and an actual 1-2-3-4 step process of what it is I'd be using and how I'd use it.
The first task is to characterize the film holders for the critical
dimension (back of camera mating surface to emulsion plane). Once critical dimension
is known a set of bracket shots is taken with plenums shimmed in .001" increments
in and on either side of the emulsion/image plane.
The focus of each shot is done with a piece of developed film you plan to use (taped
to the plenum glass in the flat emulsion/image plane at the critical dimension distance from the back of the camer), a 35x loupe and the 90, 75, 65, and 55 degree stands to cover the entire (160mm 8x10 wide angle) frame. You do not need as many stand changes for normal and near telephoto lenses.
How you bring the center, middle, side and corner of the image into fine focus is
dependent on your lens lp/mm profile characteristics, scene and your style.
The scene image will be in focus at the same time and distance as the grain in the developed films emulsion layer. These planes are aligned to one another (residing in the
same space as close as possible) and calibrated for use with a specific film thickness
and set of holders.
Once you determine which plenum shim distance works best you will shoot with that
from that point forward UNLESS the thickness of your film changes or you use a different type of film (that has a different thickness).
This instruction is a lot different from that of characterizing lens MTF lp/mm measurements (as was described in the older auction site ads).
If your film does not curl out of the box then a standard film holder should work fine.Is there a glass sheet that I tape my 8x10 film onto that I slide into some sort of special film holder you provide?
8x10 film does seem to have a problem with sagging because of its size and weight.
Rolling and Taping the film to a glass septum will hold the film down with mirror like
precision.
I'd love to provide such a beast (and would be very happy if anyone else did for that
matter).
The 8x10 is going to wait until the 4x5 version is complete.
The biggest issue with rolling and taping is that the glass-septum needs to be about the same size as the film holder to give the tape something to hold onto. This makes
it very difficult to design a film holder that provides plenty of tape room and enclose
the septum and dark side(s).
Precision pieces of sheet metal held together by machine screws (in a laminar fashion) seems to make the most sense at this point.
All this pain is a result of the film OEMs not making film that does not curl. If the fine
focusing methodology becomes generally accepted I would love nothing more than to
see flat film come fresh out of the pouch (box).
I am with Nathan on this one....
But first, i would like to applaud the OP for his tenacity in his pursuit, we need more people like this in LF :-)
Like others, I am a bit confused....but I think I finally understand the targeted goals of the products presented:
1) A method to check / correct the film plane vs. the focus plane (in our case, the gg)
2) A method to make film flat while used in a film holder, to prevent focus errors where film curl, or film buckling exist.
3) A method to place keep film flat while being scanned, and the film resides exactly on the focus plane.
Am I missing anything?
Based on this, here is a few thoughts on each of the above...
1) With most modern cameras, probably mid 80's +, makers go through extremes to keep the film plane and focus plane aligned. This is true of the better film holders as well. Often this task is confirmed and / or tweaked with lasers, which offer precision well beyond what the eye can accomplish via optics. More vintage cameras / film holders would probably benefit from the OP's system.
Keep in mind here..... if the film plane and focal plane are out of alignment, we are often saved by high f stop numbers LF uses. If LF lenses were shot at f4.0, I too would be chasing flatness down to .001".... but in non-macro photography, this issue is not so critical, such as landscapes.
Example, at f22, using .017mm for cc value (60 lp/mm), the film can vary from the focal plane by as much as +/- .015" and still acheive the desired on film resolution. This is 15x greater than .001". At f45 this "depth of focus" slop variable doubles, at f11 it's cut in half.
Also, assuming you are not photographing a flat subject (which would be a good application for the OP product), mis alignment between the film plane and focal plane greater than the amounts above, doesn't reduce resolution on the film, it simply changes the focus distance on the other side of the lens (the scene). For example, lets assume a 150mm fl lens, and you set focus at 40ft. If the film is sitting .015" further back, it will produce sharp focus at a 34ft distance. In most cases, this will rarely have any negative effects on the image recorded.... which is the beauty of LF, it really does have inherent built-in allowances for errors. Our errors of focusing are greater than this, i.e. if you tried several attempts to focus at 40ft on a ground glass, the true fl distance would vary between 35 - 45 ft... the reason is, the gg is not of sufficient resolution to enable better focus. Then toss in movement of the lens panel when locking down focus, inserting film holders, etc. Errors are everywhere.... film / gg alignment is certainly not the "weak" link in the chain. There is many weak links.
In addition, these errors are more likely to exist with lens and film standards not being square to each other, specially in older wooden vintage cameras....as wood moves over time. I would focus on these issues first.
Also a high precisioned straight edge, by a company such as starrett, combined with a high precisioned dial indicator with a roller wheel tip, at .0001" resolution, is a much simpler method IMO to pick up any errors between film plane and gg alignment. You can buy flat sheet shims in thicknesses from .001" to .1" in any increment you want, to simulate the film thickness. You can cut these shims to the film thickness you use, to make a perfect calibration between film holder and film of your choice. The shims at MSC are dirt cheap, less than $15 for a pack of multiple thicknesses. To get really picky, you need to have the focal plane lay in the middle of the grain thickness, which excludes the film base. Of course, films are close enough in thickness where this really is not a problem, unless you have a very unique application....maybe gigabyte film shot at f5.6?
2) The same numbers above, apply to film curl, or lack of film flatness. I see this being problem in two areas.... the first is, pointing the camera downward, specially with 8x10 film, the center will buckle. And secondly, roll film sitting too long in a reverse curl film holder, which there is no good solution for, other than wasting that next exposure. As for holding film flat in film holders, as mentioned previously, micro holes with a small hand pump will make the film perfectly flat, even roll film. I have this with a 6x7 RFH, and it works superb... I have not felt compelled to do this with sheet film. There has been a few makers of vacuum back film holders.... Schneider made them in 4x5 and I think Hoffman made them in 4x5 - 810.
3) Scanner film flatness. This is can be a useful tool if the scanners focal point is not on the film plane. On a flat bed scanner, I lay AN glass over my film, with Kami fluid, and all film is perfectly flat. With the consumer based scanners, this might be an issue, as I am sure QC is limited. Not sure if the betterscanning.com solution can get the film below the glass plane.
Most scanners use lenses in the f5.6 region...this again leaves sufficient Depth of Focus at the film plane, where changes of focus distances of .001" will not deliver any sharper results. The quality of the lens, the quality of the recording device, the quality of the light and electronics is what delivers a great scan. However, I am not discounting the value of confirming the lens focal plane is well aligned with film plane. I think Doug did a great job of exposing this shortcoming...
Following up on the scanner calibration I mentioned above...
On Better Scanning web site, they show the improvement in a .07" focal plane alignment correction, 70x greater than the mentioned .001" correction, and even then, it's not extreme, but still worth it IMO...
http://www.betterscanning.com/scanni...advantage.html
Thanks for the detailed post outlining concerns. Its a long one so it will take some
time to cover all the points raised. If you don't mind me asking I noticed the bglick
post of 16-Aug-2005 "cc vs. lp/mm as it relates to hyperfocal distance"
If that is not a good place for me to gather background information on your perspective feel free to let me know.
No.
No.Example, at f22, using .017mm for cc value (60 lp/mm), the film can vary from the focal plane by as much as +/- .015" and still achieve the desired on film resolution.
Let me assert that "the desired on film resolution" are the same results
that Ansel Adams got at Yosemite, with the end results being a nearly ideal
400 megapixel+ image (resulting from perfectly exposed 4x5 fluid scanned film).
Also I tend to verify, test and document everything, over, and over and over .,.
This is not the first time I've heard that the focal plane envelope of optimal
focus can vary "as much as +/- .015" especially by those who use quick loads
and ready loads. I do not recommended using them because they are known to be very soft.
I thought it was obvious to most that by looking at an unexposed piece of film
in one of those "its so easy" quick and ready things in broad daylight that the
emulsion layer's reflection and distortion at the edges reminds one of a circus
mirror and the scanned results cannot be anything but laughable.
Maybe thats one complication, you are not differentiating between focal
plane flatness, incrementally moving the plane up to a distance of +/- .015
(along the normal vector) AND nasty film curl which is helical in nature,
remembers the curve of the roll it came off of and is a good example of distortion.
I can make since of the former and see that there may exist a lens lp/mm profile
(center, middle, edge and corner of frame circumscribed by the behind lens
conic section) that at a small aperture (approaching the Rayleigh limit) or
pinhole resolve an image within an envelope approaching +/- .007 .,.
Time to put on the Test Engineering hat and figure out what is going on with
this assertion these learned scientists have made.
This perhaps (though unlikely) with the condition that the center, middle,
edge and corner of frame lp/mm values are IMMUTABLE throughout that
envelope !!! (You can-not have good focus if its constantly changing as
a function of the lenses full profile).
OOOOOOOHHHHHHH !!! AAAAAHHHHHH !!! If the pinhole (tiny itsy bitsy aperture)
assertion about the envelope is true, maybe there is some use for
the Nyquist 3d 2*root(3) here, but alias, I regress (sorry,
getting back on target). For the record my working range is usually f8
through f22, ok add an fstop to eather side for less than ideal conditions.
I apologize to the f64 and up folk, I've spent way to much time in another
life modeling knife edge photon scatter in hundreds (maybe thousands) of
photo micro graphs (and 3d beam topologies, 10, 50 and 90 percent
energy lines, etc.).
Ideally this experiment will have four scopes locked on USAF targets and
instrumenting lp/mm as we rack through the normal base envelope AND
that these supposed IMMUTABLE values are recorded on film (as observed)
and easily studied in the fluid scanned generated image (our real world
work flow).
Other constraints (not mentioned, for example, could be) excellent signal
to noise ratio and perfect point resolution, modeled as slope at a point,
correlation with knife edge photon scatter, standard deviation and
(but not limited to) other generally recognised scientific and engineering
methodologies.
Well, even if I fall flat on my face its always best to tell everyone exactly
what I'm going to do, listen to any an all suggestions and complaints,
and then execute it. Let the facts be what they may.
The resolving power of a wide angle schneider lens focused on a power line
(and insulators that hold it) at 100 yards @ f22 is impressive, but seem to be
resolved at one and only one point on the rack an pinion focusing mechanism
(and NEVER behind ground glass - yes I'm a Beattie Ultra-bright addict but only
for composition of the frame as a whole).
I've got a chart somewhere about how ground glass does bad things to sharpness
so if all of your testing has been done behind GG with ReadyLoads I'd need three or
four stiff drinks before telling you the were in focus (my wife calls be a film snob
so I apologize to those who are happy with Quickloads and GG - happiness is
a good thing).
I've tested, verified and documented this over and over again and will
continue to do so. I'd love for your assertion to be true because if I could
re-create it 4x5 photography would be much simpler but alias, I am never
that lucky (sigh).,.
I'm happy for you that there are many a math model that proves this, but
I want to see the results come out of a Heidelberg Tango or Higher end
CCD scanner. Then (what ever the results) look at and maybe come up
with a little math model or two.
Now on the off chance that your assertion is correct I hope to be running
yet another test soon !!! But this time it will be with a piece of mirror
like mounted film inside a newly fabricated film holder.
Now I've really done it, let the cat out of the bag, I hope i do not get
eaten.
I'll use three different shnieider lenses (all wide angle - hopefully one will
have this magical focal plane property under normal use). I do not plan to
go above f32 but should probably use it for completeness. I'm apprehensive
about f64 because it is so dark and there are diffraction issues.
Then I'll run the film through some calibrated wooden graphic holders,
some calibrated plastic elites and then those quick ready easy load things
(choke - anything for science).
I hope I can find a lens that will give me +/- .015 inches of sloop to play with,
but I characterize those things to death although usually not at
infinity. If I cannot reproduce your assertion I'll even go as far as testing
lp/mm with big USAF targets at infinity with a wide angle, normal
and near telephoto lenses (on a foot ball field - anything for science).
Its always fun to learn something new. If there is any way to re-create the
effect described above I will quantify, qualify, study and test it to death.
But I need to actually fabricate that super flat film holder just to be on the
safe side.
All the best,
Richard
Please supply the following information
I. A lens characterization profile (lp/mm matrix) with
A.) Center, Middle, Side and Corner of lens (x axis)
B.) f16, f22, f32, f64 (the limit of pinhole size is your choice along y axis)
II. Repeat with film (or sensor) plane offsets (along the normal
vector) of .007" , .015" , .021" (more if desired).
III. State distance to target, mm size of wide angle lens and
format (6x9, 4x5, 8x10, etc).
- Use of common window glass (shimmed for offsets) is fine.
- A 35x better pocket microscope is required for somewhat accurate measurements.
- For 4x5 90, 75 and 65 degree angled microscope stands are needed.
- For 8x10 a 55 degree angled microscope stand will be needed as well.
Best Regards,
Richard
Hi Rich
Thank you for the thorough response....
The only reason I mentioned cc, was how it related to focus plane alignment, NOT hyperfocal distance, which is a completely different issue, un related to the discussion here.
> 400 megapixel+ image (resulting from perfectly exposed 4x5 fluid scanned film).
Rich, through the years, the web is flooded with mis information about how much data film can hold vs. digital recording media. Here is the short version of what is possible with 4x5 film. First, we leave out the scanner, lets ASSUME it will grab every bit of resolution from the film, (regardless of the oversampling required) so that variable is removed. Now it comes down to lens aerial resolution at the f stop used and the MTF of the film. First, use color 4x5, f32. Assume the lens is diffraction limited at f32 (it never is, but lets give film its best chance)... 1500/32 = 47 lp/mm aerial resolution. Lets use the MTF for the color film at 65 lp/m.... a good avg. Combine in the time tested 1/R equation,
1/R = 1/(1/r1 + 1/r2)
R = System Resolution
r1 = Lens aerial resolution in lp/mm
r2 = Film MTF in lp/mm
The On Film Resolution (OFR) is 27 lp/mm. If you extract this out over a 90x120 mm film area (4x5), it will equal 32MP of resolution....this is nothing new, it has been tested by every major lens and film maker for years. There has been many tests that bring 4x5 color film very close to 39MP digital backs. Based on the f stop used for each, the scales can tilt in either direction.... as an example, if you shoot at f11 instead of f32, the OFR jumps to 84MP, a huge increase and will outperform the 39MP digital back, assuming the subject matter can withstand such small DOF range.
If you substitute B&W film, say TMax at 120 lp/mm vs. color, at f32 this value jumps to 49MP, and at f11 176 MP. Still no where near the 400MP you quote, and these are BEST CASE scenarios with a perfect camera system and diffraction limited optics, perfect alignment, etc. And as you know, at f11, the DOF is so small, you will be forced to shooting subjects that are flat, or at infinity. It seems all the digital vs. film comparisons I see, fail to make these associations.... the variables are so great, the resultant outcomes are "all over the map". That is the nature of optics, all variables must be considered in a comparison, and often the photo community wants short n sweet answers to everything. But anyway, hopefully this will bring some realistic MP values into play.
> This is not the first time I've heard that the focal plane envelope of optimal
focus can vary "as much as +/- .015"
This is based on ONE single f stop.... the depth of focus principle of the film plane has been an accepted optical principle for over 100+ years. As you know, Depth of focus is a function of f stop only, not fl. As f stop increases, so does Depth of Focus, hence the amazing viability of LF photography...
As for film curl in sheet film... I personally have never seen it in the thousands of sheets I have used... and when I insert film into my film holders, the edges are held very secure, forcing flatness, spcially in 4x5 which is small enough to prevent bulges, unless you point the camera downward. I have shot quickloads for years, and never had a problem with what you describe, maybe I got lucky batches.... Fuji understands this problem very well... But to be clear, I never shoot LF at f8 either, where any of the problems you mention could surface. So the application is significant factor here ....
> Time to put on the Test Engineering hat and figure out what is going on with
this assertion these learned scientists have made.
Grab any books on photo optics and read about Depth of Focus.... its a simple optics equation that makes perfect sense. Think of it as Depth of Field at the film plane. But more importantly, if you are shooting a scene with depth, the mis alignment issues do bring different parts of the scene into more/less focus...and with a lot of depth, its impossible to evaluate this after the exposure....unless the error was soooo great it might be obvious. The only example I can think of, is extreme curl in roll film, from sitting on a reverse curl, 180 deg roller too long.
> This perhaps (though unlikely) with the condition that the center, middle,
edge and corner of frame lp/mm values are IMMUTABLE throughout that
envelope !!! (You can-not have good focus if its constantly changing as
a function of the lenses full profile).
the examples above, assume consistent aerial resolution on the entire film circle radius. At high f numbers, such as LF...this is decent assumption...as stopping down tends to level the playing field.... but the center of any spherical optic will always be sharpest. So you can reduce the numbers I provided above by 20% to be safe....again, in most cases, we are trying to get in the ballpark.... If you look at Schneiders Website, you can get the MTF curves of all the lenses made, and see radius performance levels out at high f stops. But certainly not true at f5.6, which LF lenses are never optimized for, except the new digital lenses, which can be used for RF or digital backs.
> For the record my working range is usually f8 through f22,
LF lenses will perform very poorly at f8.... these lower f stops are used to produce enough light on the ground glass to compose the image - otherwise, they would not be offered on the lens, they are not for exposing. Most all 4x5 and 810 lenses are optimized in a tight 2 f stop range..... shorter fl's in the f16-22 range, and the mid fl lenses, in the f22-32 range, and the longer fl's in the f45 range. The abberations at f8 on 98% of LF lenses are so degrading, you often can produce a better image at f64. There is a few rare exceptions, and these are very new SSXL designs and a few others, whereas in the short fl's they can be used at f11. But regardless, at f8 - f11, you are reduced to shooting flat subjects or infinity to hold resolution..
You seem to be in agreement with focus error being a necessary evil of LF in general. And of course, the big issue not mentioned is, when you focus at f5.6 and then stop down to f45 to shoot, you have shifted the focus plane, another weak link in the chain here, specially when considering the alignment precision you are pursuing. In other words, no matter how perfect the alignment is in all aspects of the camera, the shortcoming will be the inability to focus precisely, and the inability to focus well enough at higher f stops, due to lack of light. Of course, this is not justification for the rest of the camera system to be completely out of alignment.... but it demonstrates why ultra precision film/gg alignment will often not improve the final OFR.
> If I cannot reproduce your assertion I'll even go as far as testing
lp/mm with big USAF targets at infinity with a wide angle, normal
and near telephoto lenses (on a foot ball field - anything for science).
I have tested all types of lenses for many years.... and its truly remarkable when the set-up is right, how it confirms the math and often the MTF values the makers provide. This is for both film and digital. keep in mind, a football field is extreme. You don't need targets this far from the lens. Chris Perez has done an excellent job testing lenses of all types, I am sure you have seen his web site? His results concur with everything I have ever tested.
http://www.hevanet.com/cperez/testing.html
Thank you again Chris for being such a big supporter of the LF community.
To be clear Richard, I am not knocking your desire to achieve the flatness you desire. And in some extreme shooting situations, I am sure it will be very helpful, such as f5.6 - f8 doing reproduction work of 2d subjects. But using common LF lenses, which probably 90% of them in existence, are designed to be shot at f22 - 45, their is such a large inherent slop factor, in the end...the means don't justify the ends. If you follow this forum, you will see this issue surface every few years.
I think a more obvious area of attack to improve image quality is the squareness between the front and back standards, specially with short fl lenses shot at f8 - f16. As slight un-squareness presents huge variances in the focal plane vs. the film plane. Of course, this problem is magnified in 810 vs. 45.
Comments on your 2nd post....
A good test, to reduce testing time, is to first test the lens with a digital back. I have a set-up that enables a DSLR to mount as a film holder. This gives you a "sure" flat recording plane, and a great focus mechanism, as I use after-market focusing screens sold by Katz Eye Optics - with these screens you can really nail the focus to the levels of accuracy you desire. Now you have a basis to compare film with... Digital recording is subjected to the same 1/R curse....(which is Nyquist theory in action) of course, SWA and even some WA lenses will not perform well with most digital sensors, as this still a weak link in digital sensors....
Anyway, we all are interested in your findings.....
Not that long ago, I exchanged info on line with a guy who made a DIY 4x5 vacuum holder. It didn't look that hard to do. It turns out that a rubber bulb pipette bulb has enough suction. I no longer have the contact info, but a little Googling should turn it up.
Posting Permissions
Reply With Quote
Bookmarks