[newarts]

the big issue with respect to hoods, regardless of rectangular or petal hoods is the ability to gaurantee regestration to the sensor, to avoid unwanted vignetting at the corners.

Clearly

also note that rectangular hoods, while better than circular hoods are still not optimum.

as someone stated earlier the shape of a petal hood is defined by the field of view at that specific point along the sensor . if you are shooting very close to the sun, the difference could still be quite noticebale.

I'm not clear on what "at that specific point along the sensor" means; I need to think about it a bit and would appreciate an analytic reference.

I'm not clear on what basis optimum is meant. I think I can make a rectangular prism hood the same depth as the total depth of the petaled hood that excludes more unwanted light.

My basic argument follows. Let's look along the diagonal of the petaled hood - the upper hood corners shown would be the lowest points on a tulip hood's profile:


Clearly I can exclude more peripheral light by increasing the hood's depth at the cross section shown. Regardless of the hood shape outside this cross-section the fraction of peripheral light will be decreased if I increase the hood depth of this particular cross section.

If you hand me a petaled hood, I can decrease the total peripheral light entering it by decreasing the depth of the grooves defining the petals (I've got to take care that I do this at an angle that doesn't cause vignetting, but it can be done.)

Dave

[pacerr]

Definitions of tulips and "optimum" lengths and angles notwithstanding, there's one characteristic of hoods I value most -- and haven't seen addressed here -- compact storage.

With a little help from an origami fold pattern, a square or rectangular hood design allows it to be folded flat for convenient carry.

The rectangular "hole" where it fits to the lens is left open and tabs are left on the sides there to temporarily attach the device with a rubber band or tape. The short dimension of the sides of the "hole" is equal to the diameter of the lens which defines overall size.

This isn't going to be the smallest possible, "optimum" size of hood for a given lens but since it folds flat when not in use and can more or less match the rectangular proportions of the sensor, it's very efficient and can even be made somewhat oversize without penalty when not in use.

I recall seeing a metal, collapsible hood back in the dark ages and I just replicated that design as a proof of concept with a piece of typing paper. These would be so cheap to make from stiff paper/poster board they could be considered one-time-use items and would be perfect on the trail.

H2

This topic has departed from Robin's OT. Might it be usefully revived as an article?

[newarts]

Definitions of tulips and "optimum" lengths and angles notwithstanding, there's one characteristic of hoods I value most -- and haven't seen addressed here -- compact storage. ....................
I recall seeing a metal, collapsible hood back in the dark ages and I just replicated that design as a proof of concept with a piece of typing paper. ............

I agree! I'd love to see your proof of concept!

[Ron Kruger]

It is pretty simple geometry.

One type lens hood is a cone with the same angular field of view as the lens so it cannot cause vignetting.

If big end is one focal length from the small end, the big end diameter should be 2x the filter ring diameter to avoid vignetting.

I think this basic design should work....the larger the ratio of height to focal length the better.


EDIT: there is no theoretical limit on how deep the hood can be, the deeper it is, the less light gets in from outside the field of view.

I'm not very good at math, but I do have an inate understanding of physics, and I usually follow the KISS program in matters like this. The diagram posted puts it simply.
Camera companies put a lot of science into their hood designs, but the end product is a compromise between functionality and size. As long as hood doesn't invade the photo plane, it could be as long and as wide as one wants, and might even do a better job than the factory hood, which must be shorter and narrower for marketing reasons. I have tought of devising a hood for my 15mm that is a few inches long and a foot or two wide. To devise it, I would use Pacer's approach.

[Lowell Goudge]

Clearly

I'm not clear on what "at that specific point along the sensor" means; I need to think about it a bit and would appreciate an analytic reference.

your diagram is not really correct, which is why you are not understanding the point here. In your diagram the lines indicating the "total light admitted into hood" is actually the correct representation of the lens , hood and field of view. Remember the image is inverted through the lens. Now, looking at those lines, consider a smaller image circle, which you would have as yoou move alonmg the top or bottom of a frame, from the corner towards the sensor center. The lines will still cross at the same point and therefore, as the image circle reduces the length of the hood would increase, this is why the petal shape evolves.

I'm not clear on what basis optimum is meant. I think I can make a rectangular prism hood the same depth as the total depth of the petaled hood that excludes more unwanted light.
Dave

note again you mis interpret the use of "Unwanted" light, but a rectangular hood could also achieve the same but not better effectiveness, just that since lenses are round it is much easier to attach a round , but petaled hood, as opposed to a rectangular pyramid, that you need to somehow, make circular.

[sewebster]

?

[Lowell Goudge]

?

Nope

consider the lens as a simple lens (for this discussion it works)

Light rays cross after the lens before the sensor.

Distance lens to sensor is the focal length.

[sewebster]

Nope

consider the lens as a simple lens (for this discussion it works)

Light rays cross after the lens before the sensor.

Distance lens to sensor is the focal length.

Ok, right.

But anyway, what level of optimization are we talking about here? Imagine if you have your subject, let's say of a painting, something flat. Now you point your camera at it, and then go and draw a rectangle on the painting of the area that can be seen on the sensor. If you now make a rectangular prism hood that extends from the rectangle you drew on the painting back to the outside of the front of the lens (not covering the front element), this would work, right? (ignoring the fact the the painting would be shaded by the hood)

So now you can just cut off portions of this rectangular prism hood to make it as short as you want.

What am I missing?

[Lowell Goudge]

Ok, right.

But anyway, what level of optimization are we talking about here? Imagine if you have your subject, let's say of a painting, something flat. Now you point your camera at it, and then go and draw a rectangle on the painting of the area that can be seen on the sensor. If you now make a rectangular prism hood that extends from the rectangle you drew on the painting back to the outside of the front of the lens (not covering the front element), this would work, right? (ignoring the fact the the painting would be shaded by the hood)

So now you can just cut off portions of this rectangular prism hood to make it as short as you want.

What am I missing?

That could work, cutting it down to the limit of the vignetting, AS LONG AS the taper is a lower angle than the field of view. otherwise the shade will not give oiptimum coverage.

I just think a little math, that lets you calculate a cylindrical hood with tulip edges neatly is a lot simpler.

[sewebster]

That could work, cutting it down to the limit of the vignetting, AS LONG AS the taper is a lower angle than the field of view. otherwise the shade will not give oiptimum coverage.

Well, my hood design (the rectangular prism) would not vignette at any length. You just cut it down for convenience. The taper is almost exactly the field of view of the lens. The only reason it isn't exact is due to the fact that the lens is a compound design with finite size etc.

I just think a little math, that lets you calculate a cylindrical hood with tulip edges neatly is a lot simpler.

I guess it depends on your concept of simple. In terms of visualization I think I disagree as the tulip is a somewhat "complicated" shape. In terms of manufacturing, probably, though if I were making my own hood I would probably just use a plain cylinder (possibly with a step up ring).

[Lowell Goudge]

Sewebster.

Just remember that I still maintain that the angle of the taper is wrong and therefore without math your design won't be optimum. You would need to start with a taper that formed a partial mask over the front of the lens to follow the FOV

The tulip design is easy to calculate if you look at the formulas in the spreadsheet hoodcalc you just need to copy that calculation for a 360 degree path around the lens

[sewebster]

Just remember that I still maintain that the angle of the taper is wrong and therefore without math your design won't be optimum. You would need to start with a taper that formed a partial mask over the front of the lens to follow the FOV

Sure, this is what I'm talking about when discussing the finite size of the lens above. In terms of actual light blocking, only the orifice on the front of the hood matters. After that (as you move towards the lens) the shape is basically only relevant in terms of size/ergonomics etc., as long as your hood does not taper _more_ than the FOV of the lens (unless you are worried about reflections off the inside of your hood). So you just need to calculate the size of the rectangular plane your lens "sees" at the desired hood length and build a hood with that orifice in the front, and then just use a convenient shape.

[newarts]

I've thought through what's required for a lens hood and how the design varies with distance from the scene as well as other parameters such as focal length and lens diameter.

Here's what I think:
1. The goal is to exclude all light that does not originate from the scene.
2. The scene is rectangular in shape and has the same aspect ratio as the sensor.

Imagine a scene that is lit from behind, so we don't have to worry about light getting to the scene to reflect from it - like a back-lit movie screen.

Based on the above the perfect hood has a rectangular open end that contacts the edges of the scene and perfectly black walls that exclude all peripheral light. The other end of the hood has a hole through which the lens sees the center of the scene on its optic axis.

Here's an interesting bit about how the perfect hood changes for close-up scenes.....
As the distance to the scene approaches infinity angle of the rectangular pyramid defining the hood's diagonal approaches the lens's nominal field of view. However, as the scene gets closer to the lens, the angle of the pyramid must get smaller if it is to remain in contact with the edges of the scene.

When the scene is positioned 2 focal lengths in front of the lens, it is exactly the same size as the rectangular sensor, so the hood has a small angle. As the scene gets closer and closer to the lens it approaches being a rectangular point of light at the lens' focal point.

Therefore as the scene approaches the lens' focal point the hood's scene end must get smaller and smaller; the macro hood has a relatively large circular opening for the lens and tapers into a rectangular cone with the scene at the tiny end.

While the decrease in optimal hood angle with close scenes is interesting it is of little consequence for normal photography. For normal photography we must expect scenes that are many focal lengths from the lens and the lens' nominal angular field of view on the sensor diagonal is appropriate for the hood to receive the light from the scene's diagonal.

The optimum hood allows in the least light - one can carve petals onto a conical or cylindrical shaped hood to exclude some light which will ultimately not strike the sensor (ie. off the scene diagonal) but I believe petals are not needed for a properly shaped rectangular hood.

Dave

[Lowell Goudge]

Dave

To put it in simple terms the
Petal is defined by the intersection of the lines that project out from the lens at any point along the angle of view of the lens for a particular sensor to the edges of the rectangular hood with the cylinder

They both offer the same exact protection since the edges follow the same plane of the rectangular permid.

Don't believe me just cut a hole in the bottom of the permid that fights tightly around a cylinder

[sewebster]

Dave, I think I agree basically exactly with you, and that's what I've been saying.

Lowell, a pyramid that fits on the outside of a cylinder (if the cylinder is a reasonable hood) would be much larger than necessary. The rectangle for the plane of view needs to fit INSIDE the end of the cylinder. The rectangular prism hood is basically like a cylinder with parts of it shaved off, reducing the size.