Critical focus and 120 film bulge

[kmac]

4 hours ago, conrad_hoffman said:

slow lenses

I guess that's why very early Kodak roll film cameras had no "pressure" plates. Instead, they just had two lengths of springy thin sheet metal which applied pressure on the outer edges of the backing paper. Who knows what bows or contortions the film was subjected to without a plate ? Perhaps leaving it to the slow lenses to fix things with long depth of fields. With no film available any longer for those old cameras, I use sheet film inserted into fabricated black cardboard holders that I made up.

                                             

Interestingly, the two rollers in these old cameras are about twenty thou (0.020") above the level of the film gate. I haven't measured the exact height but there is a sizeable gap when I placed a straight edge across the rollers. It may have been an additional, but necessary attempt to keep the film flat. The path of the film would have been: going over the first roller, then bending down to the film gate, then traveling along the film gate under the pressure of the springy rails, then bending up to the second roller, then down to the take-up spool.

Whenever I check the focus on the old Kodaks with a piece of translucent paper stretched across the film gate, the focus is always pretty sharp due to the greater depth of field, provided the pointer is exactly opposite the settings on the focus scale. The widest apertures are f7.7 or f6.3 depending on the camera. Anyway, we have a "Folding Camera" board now, so I'll post in there about old folders in the future, it's just interesting how designers went about keeping roll film flat, it must have been a head ache for them at times, versus production costs.

 

 

[rodeo_joe1]

21 hours ago, Lucas Lee said:

Will this mean that the film will rest against the pressure plate given time to “relax”? 

Probably not. As I said, even processed film still has the same tendency to curve inwards to its emulsion side. It takes quite long storage inside a negative sleeve in a folder to flatten it, but as soon as it's released from its prison it starts to acquire a curl again. 

I spent years in darkrooms wet-printing, and in all that time I never saw a piece of (thin base) film that wanted to curl away from its emulsion side. It's always, however slightly, concave on the emulsion side and convex on the base side. 

Given that tendency, I'd say the chances of the film going flat of its own accord in a loose channel are zero. My best guess is that its natural curve away from the lens is simply restrained by the 'pressure' plate. Resulting in it touching the plate in the centre region of the frame, and bowing away from the plate at the sides in order to follow the film path.

OTOH, the Mamiya 645 film-path bends the film tightly back over rollers, against its natural curl and immediately before the gate. There are also sprung pin-rollers on the emulsion side of the film at either end of the guide rails. Without microscopic examination of a film within the camera, I surmise that this does indeed keep the film flatter in the gate. However there's still that 0.5mm gap for the film to do whatever it likes. 

Like I said at the very start - it's film! It's flexible and has an emulsion thickness. So ultimate accuracy and precision of the image-plane isn't its forte. It's a fairly approximate medium. 

Edited January 28, 2023 by rodeo_joe1

[blopin]

6 hours ago, rodeo_joe1 said:

Probably not. As I said, even processed film still has the same tendency to curve inwards to its emulsion side. It takes quite long storage inside a negative sleeve in a folder to flatten it, but as soon as it's released from its prison it starts to acquire a curl again. 

Thanks rodeo, your explanation was not lost on me there. By relax I meant curl. As the film loses longitudinal tension it will start to curve, possibly causing the middle to contact and flatten itself on the locating plate as the outer edges push on the rollers. 

[kmac]

12 hours ago, Lucas Lee said:

Thanks rodeo, your explanation was not lost on me there. By relax I meant curl. As the film loses longitudinal tension it will start to curve, possibly causing the middle to contact and flatten itself on the locating plate as the outer edges push on the rollers. 

That's why the plate is there, to keep the film within the given space, designed into the camera, so that curly films are "restrained" (Joe's description, and a good one) in place to be within the shortest depth of field of f2.8. The "space" is the film + backing paper + clearance of about three to four thousands of an inch. A film might not be dead flat within that given space, but it should still be within the shortest depth of field ... and what would that be in reality, 12inches ? Look up the DOF chart for a Rolleiflex, has your Rollei got a chart on it's back door ?

I beg to differ about roll film having cross-sectional curve when it is initially wound off it's spool. I haven't come across that yet, only after processing have I seen it, the emulsion changes and causes that cross sectional curve, but it's not there when you're taking pics. Perhaps with long expired film there might be a little, but not with fresh film. To my mind, the only curve you should be concerned about is longitudinal curve, caused by the film rolled up on it's spool, the film "sets" with that curve in it and has "memory" of that curve after it leaves the spool, hence the plate to keep it from moving away from the film gate.

If, as you say, your Rollei has rollers that are below the level of the film rails, this would be a case of the film running firmly on the film rails and the few thou clearance provided for free travel, would be at the back of the backing paper, or in the case of 220 film, at the back of the film. When you get your ground glass, it seems with your Rollei, it will have to be held down firmly on the film rails, without a paper spacer as discussed before, because if it's true that the rollers are below the film rail level, then the film will run along the rails, and not above them. I would double check the rollers with a good straight edge and try slipping a thin piece of paper between the straight edge and the rollers, and make sure of it.

Edited January 29, 2023 by kmac

[rodeo_joe1]

On 1/29/2023 at 5:29 AM, kmac said:

The "space" is the film + backing paper + clearance of about three to four thousands of an inch.

From my measurements of the Yashica-mat and Mamiya (~0.5mm) and Rollei's spec of 0.45mm, the free space is clearly a fair bit more than 3 or 4 thou. It's between 0.2 to 0.25 mm, or about 8 to 10 thou. 

FWIW: Depth-of-field tables won't give an accurate assessment of film-plane displacement blur, which is a depth-of-focus issue. There's a subtle difference between depth-of-field and depth-of-focus, whereby D-o-Field is very dependent on subject distance, while D-o-Focus is not. A film-plane displacement will indeed cause a shift of focus, but the degree of blur can't be given a fixed field depth because it varies with the subject distance. 

As you say, the tolerable blur is given by the f-number * whatever acceptable circle-of-confusion is chosen. Conventionally it's about 0.053mm for the 6x6cm format. Which gives a film-plane tolerance of +/- 0.148 mm @ f/2.8, well less than the (at least) 0.2mm of slop that many cameras are machined to!

On 1/29/2023 at 5:29 AM, kmac said:

I beg to differ about roll film having cross-sectional curve when it is initially wound off it's spool.

Who mentioned a cross-wise curve? And does it matter? A displacement from flat is relevant in whatever axis. However, the squashing of a naturally curling film by the pressure-plate (the pressure coming from the film itself) will constrain the film to the plate-distance, minus the thickness of film+backing paper, over a considerable area of the frame. So it's not a problem if the focal plane is calibrated to that distance. 

It could only be an issue if, for some weird reason the film decided to bulge toward the lens. - Hasselblad, please take note.

Edited January 30, 2023 by rodeo_joe1

[conrad_hoffman]

Ah, for a process camera with a vacuum film back!

[rodeo_joe1]

10 hours ago, conrad_hoffman said:

Ah, for a process camera with a vacuum film back!

And ear-defenders to go with it! 

Before H&S was such an issue, I was required to use a Littlejohn copy camera with a huge vacuum platen - 24" by 30" or some similar ridiculous size. When the pump was turned on the noise of air being sucked into dozens of tiny holes was pretty deafening in the small darkroom that formed the camera chamber. Great memories... I think! 

[blopin]

Thanks for clearing up any dissonance from my intended reasoning rodeo, not to discredit kmac for that matter. I meant the bend along the axis that curls around the spool.

If the silver acetate sits firmly on the guide rails, it will rub on the sharp corners on the ends of the rails and may induce friction/scratching/or even tearing. Though the rails are above the rollers and may exacerbate the likelihood of this possible thesis, I don’t think this is the case. I simply think that the natural curve of the film helps it shy away from the sharp edge of the guide rails. This same leaf-spring tension also allows it to rest on the locating plate. 

23 hours ago, rodeo_joe1 said:

As you say, the tolerable blur is given by the f-number * whatever acceptable circle-of-confusion is chosen. Conventionally it's about 0.053mm for the 6x6cm format. Which gives a film-plane tolerance of +/- 0.148 mm @ f/2.8, well less than the (at least) 0.2mm of slop that many cameras are machined to!

However I need help understanding the last sentence of this paragraph. What’s the method to deduce film plane tolerance from circle of confusion? 

[rodeo_joe1]

11 hours ago, Lucas Lee said:

However I need help understanding the last sentence of this paragraph. What’s the method to deduce film plane tolerance from circle of confusion? 

Both depth-of-field and depth-of-focus make assumptions about lens behaviour and human vision:

1. That real lenses behave like a 'thin lens' model - having zero distance between front and rear nodes. 

2. That a lens has no aberrations, especially spherical aberration, and projects a perfect cone of light from each infinitely small point on the subject plane to an equally small point in the image plane.

3. That diffraction of light plays no part in image sharpness

4. That human vision is consistent in tolerating a certain limit of blur, called a 'circle-of-confusion', which in turn is a cross-section of the aforesaid perfect cone of projected light. 

If we take all those assumptions as being representative of real-world lens behaviour (ahem!) then we can formulate the point of intersection of an imaging-plane (i.e. Film or digital sensor) with the projected light cone at other planes of focus. Planes resulting from a change of subject distance - depth-of-field - or a misplacement of the film or sensor from the true image-plane - depth-of-focus. 

Depth-of-focus is easier to calculate because it's symmetrical about the point of focus. So, if we set a limit to the acceptable cone diameter - the circle of confusion - and we know the lens aperture number ( = focal-length/physical lens diameter) then we can work out how far from true focus the film or sensor is allowed to be. To a first approximation this simply = C-o-C * f-number.

Plus or minus depending on whether the true focus falls in front of, or behind, the imaging plane. 

Edited January 31, 2023 by rodeo_joe1

[glen_h]

On 1/26/2023 at 4:50 AM, rodeo_joe1 said:

(snip)

However, unprocessed film is a little thicker than when it's processed,

(snip)

In an optics lab in college, we got to make white light holograms.

Traditional holograms have a thin (enough) emulsion, that you get a 2D interference pattern.

White light holograms are made with a thick emulsion, such that you get a 3D pattern.

Using a red (HeNe) laser, the interference pattern has a scale based on the 632.8nm wavelength.

 

But when developed, the image comes out green.

 

The emulsion shrinks in thickness, such that the spacing is now a green wavelength.

You want a (close to) point source to view, but otherwise you get a green hologram.

[kmac]

Here's a good link on the subject of Circle of Confusion ...

https://en.wikipedia.org/wiki/Circle_of_confusion

However, the question in the opening post is related more to a mechanical issue than an optical one: "How much does the film bow in the camera, and how can it be compensated for ?"

A simple answer could be to make sure the film is within the short depth of field of f2.8 ... this, I take it, is what the OP is going to do by using the the ground glass. The image on the ground glass can be checked for sharpness over the whole frame. If any blurriness is picked up on any part of the image, then the bow in the film will be too much for that short depth of field, and f4 will fix it with it's slightly longer depth of field. But this won't happen in the case of a Rolleiflex if it's in good condition, let's be honest about it. Do they get that bad that the film bows excessively and f2.8 is no longer of use ?

[blopin]

3 hours ago, kmac said:

A simple answer could be to make sure the film is within the short depth of field of f2.8 ... this, I take it, is what the OP is going to do by using the the ground glass. The image on the ground glass can be checked for sharpness over the whole frame. If any blurriness is picked up on any part of the image, then the bow in the film will be too much for that short depth of field, and f4 will fix it with it's slightly longer depth of field. 

Hi kmac, I just wish to verify center sharpness. But loss in sharpness at the corners arising from lens alignment issues puts me at the deep end of amateur camera repair, and I’ll need to get it professionally serviced instead. 

3 hours ago, kmac said:

But this won't happen in the case of a Rolleiflex if it's in good condition, let's be honest about it. Do they get that bad that the film bows excessively and f2.8 is no longer of use ?

I wasn’t saying that film-bow results in overt loss of sharpness and an unusable image, just that this effect may be more pronounced on enlargements. Rodeo did bring up the hasselblad flatness issue, another thread on this website ceded into mudslinging between impassioned hassy loyalists and those too poor to afford one (just joking, no offence intended) 

Moving the faceplate of my Rollei 1mm takes about a 17mm circumference rotation of the focusing knob. Cut to the focal plane distance tolerance of around +/- 0.15mm at f/2.8 as rodeo mentioned regarding circle-of-confusion, which is a generous 2.55mm of focusing knob movement. A paper shim or two might not seem like a lot, but at 1/10 a millimetre thick this would mean a disparity of 1.7-3.4mm on the focusing knob! That could be the difference between infinity and 20 meters. 

[rodeo_joe1]

2 hours ago, Lucas Lee said:

But loss in sharpness at the corners arising from lens alignment issues puts me at the deep end of amateur camera repair, and I’ll need to get it professionally serviced instead. 

All lenses have inferior sharpness wide open and away from the centre of the frame, often becoming noticeably blurred in the corners. That's assuming they don't suffer from curvature of field, which is another big assumption. 

Worrying about corner sharpness at f/2.8 is a wild-goose chase. 

[rodeo_joe1]

Here are some comparisons of centre versus corner resolution with Planar-type 50mm lenses on the 35mm format. Expect an 80mm lens on 6x6cm to behave a bit worse, because the corners are 3 degrees further off-axis.

New style 7 element West German Zeiss Planar @ f/2.8, centre - 

And corner - 

CZ Jena Pancolar - an old style 6 element Planar copy @ f/2.8, centre - 

And corner - 

And the corner @ f/4 - 

Now re-focused for best corner definition; showing that there's quite a bit of field curvature - 

BTW, these are 100% crops from a 60 megapixel digital camera. On film you'd be hard-pressed to resolve the 100 lppmm bars at all.... but if someone wants to give it a try, go ahead! 

Edited February 1, 2023 by rodeo_joe1

[blopin]

I just got the digitized scans back and wow, are they soft. I must have missed focus on every single image except those racked to infinity. 

Albeit a low resolution scan, here’s an image of a vinyl record sleeve for The Cars that I used as a makeshift test chart. I used a 7x loupe and spent a gross amount  of time and concentration on attaining the most precise focus possible (not willing to disclose my age but I have a very young pair of eyes). Then shot at 1/125, f/2.8 on Portra 160. The point of focus was on Ric Ocasek’s name. 

I don’t have the negatives right now but I plan to go over them thoroughly under the loupe. But from viewing the scans there appears to be no perceptible blur or indistinctness. 

[Dustin McAmera]

11 minutes ago, blopin said:

vinyl record sleeve for The Cars

11 minutes ago, blopin said:

not willing to disclose my age but

 

That ship sailed! 😜

[blopin]

4 minutes ago, Dustin McAmera said:

 

That ship sailed! 😜

I’m young with an affinity for 80s riffs, I must have been a bohemian in the last life 🫠 

[rodeo_joe1]

All I can say is that I'd expect my FF (36x24mm sensor size) digital camera to resolve the half-tone dots on that record sleeve with its 28-75mm zoom lens at 50mm and stopped down to f/4, using digitally magnified focus-peaking and my aged eyes. 

Without disclosing my age, I'm old enough to severely lower my expectations if ever I use a film camera.

Having said that, I would expect the lettering to be reasonably crisply rendered in the centre of the frame and in about a 1" radius circle around it. 

The sharpness of the negatives will be the real acid test. I really wouldn't trust commercial scans unless costing a small fortune from a pro lab. 

IME you should expect between 60 to 80 lppmm to be clearly resolved centrally on 100 ISO T-max 120 film with a Planar lens @ f/2.8. Any faster film, and with an older style emulsion - all bets are off. 

Edited February 17, 2023 by rodeo_joe1

[rodeo_joe1]

On 2/11/2023 at 5:44 PM, blopin said:

Then shot at 1/125, f/2.8 on Portra 160. The point of focus was on Ric Ocasek’s name. 

Addendum, just read the above. 

Did you use a tripod? 

And Portra 160!? Why? 

Use a slow B&W film and you might have a chance of seeing exactly what the camera and lens can do. Tripod mounted and with electronic flash for illumination of course.