[52mm]

Please excuse me if I am preaching to the choir and beating a dead horse here, but I recently got some generic screw-in lens hoods in for my bunch of old manual glass and decided to do a bit of experimenting. Anyone can do this - a look into the back of the lens with a naked eye at various angles to a compact light source (with a hand for shading / light seal) will reveal stray light coming from places it's not supposed to (which is outside the aperture). If you can see it, chances are so will the sensor, leading to reduced contrast if it's severe enough. (The exact level still is up for debate.) We're generally talking angles well outside the field of view here, up to 30-60° off-axis. Doing it this way seems a lot easier and more insightful than staring at images taken under potentially less than ideally controlled and constant conditions.

My conclusions:

1. For best contrast under adverse lighting, use a hood as long as possible without introducing vignetting. (Surprise!) That also means different hoods for crop vs. FF. Even if you can't resist the flair of a lens, you should be able to resist its flare.
2. Performance improves when stopping down, sometimes dramatically, depending on how good internal blackening is. (Lenses with manual aperture tend to prove more insightful than modern-day all-automatic ones as you have a manual readout.)
3. There are some major performance differences in lenses at times. My worst 1.4/55 from the mid-'70s was 3 stops worse or needed a hood twice as long (into corner vignetting territory above f/2.8 for we are talking plain circular hoods) to become bulletproof at the same aperture as my best (if several years newer) 1.4/50s - or even its 1.7/50 cousin of same vintage for that matter. Somebody once called this 55 a "vampire lens" - do not subject to daylight, especially wide open. This trait seems to be quite common on early fast lenses. My f/1.7s varied a lot less in general, none were straight terrible. The best result, incidentally, was turned in by an M 50/1.7, which would be decent even without a hood (and oddly enough, somewhat better than its A series counterpart).
4. Integrated sliding lens hoods on 135s (as they were common around 1980) are a nice touch but generally way too short to do much of anything aside from keeping the worst reflections out of the front lens. We're talking something like 15 mm of effective extension when you could easily do with 100+ on crop assuming this is not getting too unwieldy for your tastes (I combined two hoods with filter threads for 70 mm total). Results on 135s seem less convincing in general, most of mine were rather meh. The K 135/3.5 did reasonably well, though the star of the show surprisingly was the CZJ "Zebra" Sonnar from the early '70s, complete with single coating and shiny metal aperture blades (if good internal blackening otherwise)! (Unsurprisingly, however, you don't actually want the sun in the frame with this one.)

(Writeup with all results at 135 and 50-55, respectively. Note that I am generally referring to crop, not FF.)

Now I've got a question left: There are some nifty websites that let you design print-out lens hoods, but you have to enter the measurements yourself - how do you actually arrive at the dimensions necessary? Is there a way to calculate them? Or is that way too lens-dependent, making it generally infeasible for anyone but the lens manufacturer and requiring an emprical approach instead?

---

Last edited by 52mm; 11-03-2017 at 10:52 PM.

[Alex645]

Now I've got a question left: There are some nifty websites that let you design print-out lens hoods, but you have to enter the measurements yourself - how do you actually arrive at the dimensions necessary? Is there a way to calculate them? Or is that way too lens-dependent, making it generally infeasible for anyone but the lens manufacturer and requiring an emprical approach instead?

I would take a practical approach and create a crude mockup to measure. If the front of the lens rotates when focusing or zooming, then you're going to end up with a cylinder shaped hood. If it doesn't rotate, you can make a petal shaped hood.

I'd use a heavy weight black paper and make it into a cylinder for the lens. With it taped to the front of the lens, shoot a white surface at your largest aperture (f/1.4, f/2, f/2.8, f/4, etc). It should be long enough to see vignetting on both the corners and center edges. Trim the length of the paper until you do not see vignetting in the white test shots in the corners. For a petal shape, the max length is determined by when the vignetting is gone in the center of the edges, and the min length in the corners when vignetting no longer darkens the corner.

This would be the most effective for prime lenses. For zooms, you'd have to do the test and design the hood with the shortest focal length or widest angle set.

Personally I think that's a lot of work for a hood and would only do this if I had a lens with some rare or extreme diameter for which no hoods are made or available.

[rawr]

Designing your own hoods does indeed seem a lot of work, esp. when eBay is full of cheap hoods of every dimension, material and form (I have about 20 of them).

[Nicolas06]

To be honest I recognize the usefullness of hoods but I hate most of them. I like small and light lenses.

My findings is that the lens design and coating is more important that the exact hood of having a hood at all. The DAltd comes with bundled hood that are small/unobtrusive but even if you forget to put them on you get more punchy image and no flare even in difficult condition than many lenses not optimized to resist flare with the best hood ever.

Older zooms especially tend to be the worst as the coating are not that great and the hood can only cover the widest setting but overall the more glass you have the more the lens tend to be sensitive to flare.

[photoptimist]

Now I've got a question left: There are some nifty websites that let you design print-out lens hoods, but you have to enter the measurements yourself - how do you actually arrive at the dimensions necessary? Is there a way to calculate them? Or is that way too lens-dependent, making it generally infeasible for anyone but the lens manufacturer and requiring an emprical approach instead?

For rectilinear lenses, if you know:

1. the sensor format (including allowances for IBIS movement);
2. the minimum focal length;
3. the maximum aperture;
4. the filter diameter; and
5. the distance between the entrance pupil of the lens and the filter ring

then you can design the perfect circular or petal hood*.

The first four numbers are easy to get. The last can be found by techniques used in the panoramic photography community for avoiding parallax which stitching tiled images together (NN4 Guide | Finding the Entrance Pupil of a Lens). With those numbers, the hood geometry is "simple" in being defined by some 3-D computations of the geometry of the light rays and the lens.

*Note: petal hoods are vastly superior because a circular hood lets in twice as much flaring light as a petal hood and it's too easy to have a bright flaring light source that's not visible in the viewfinder but that a circular hood is failing to block.

[BigMackCam]

Designing the perfect hood for a lens sounds like an interesting and fun project But, it's more work than I'd be comfortable with for the gains returned.

I shoot a lot of old manual Soviet glass on both full-frame and APS-C. In most cases, regardless of sensor format, I use generic metal screw-in hoods designed for wide, normal or telephoto lenses on full-frame. Often-times, they're not specific to the individual focal length of the lens, so they're far from optimal, but they help considerably... and when necessary, I can always position my hand to block out additional light.

One improvement I've made on a number of these hoods is to line the inner surface with anti-reflective black felt. Easily and cheaply done, and it makes a big difference in certain situations.

[photoptimist]

There's a quick-and-dirty test to see if the hood you've picked is too narrower or too wide:

1. Set the lens with hood to: infinity focus, the widest zoom, & widest aperture
2. Hold the camera backwards at arms length with the viewfinder pointing up at the sky (notice that you can see the lit focus screen in the center of the lens).
3. If you carefully rotate the camera diagonally, you can actually find the edge and then corner of the focus screen as seen through the lens.
4. If the hood blocks your view of the corner, there will be vignetting and if you get the corner and the edge of the hood is far away, then the hood is too wide.

[52mm]

That's a very clever trick, actually. I presume it would only work up to about f/2.8 though, limited by the viewfinder? I can just about still make out the corner with the DA 35 f/2.4 on the K-50.

Expanding on the idea, I proceeded to roll up a piece of ordinary white paper (80 g/m²), make a tube to go around the lens and shine a smallish (LED) flashlight through the viewfinder. In a darkened room, it does give some decent-ish petal contours, more so on the inside than the outside. Seems workable enough for some pencil marks in any case. Cool.

EDIT: So for this combination of camera and lens I am getting:
Min distance 8-8.5 cm from mount
Side petals 11-11.5 cm
Long petals 15 cm
with a tube diameter of about 59 mm.
The lens itself is about 4.5 cm long, so from the front you'd be getting:
Min distance 3.5-4 cm
Side petals 6.5-7 cm
Long petals 10.5 cm
So my 35mm long, 58 mm dia. circular screw-in hood clears the corner with not much room to spare (seems right - I could already see that earlier). Those petals are quite impressively long, so I'm not surprised a circular hood is far from ideal.

Now with M 50 f/1.7. Boy, this is getting fuzzy.
Min distance ~9.5 cm from mount
Long petals 18 cm
Second try:
Min distance 9-11 cm
Long petals 20.5 cm
Lens length ~3.5 cm, so subtract that.

Now since I know that 70 mm long, 62 mm dia. in front of the lens already is a bit too tight, I guess minimum distance from the front should be around 6 cm, long petals maybe 12 then? This is getting pretty long.

Another data point for an Auto Revuenon (Chinon) 50mm f/1.4:
Min distance ~9 cm from mount
Long petals ~20 cm
Lens length ~3.8 cm

EDIT^2: I am thinking that instead of going with a super long petal hood, one may be better off complementing a circular hood with a rectangular window thingy, like what the 21 Limited has... This seems to work best on shorter lenses, since at short distances things are round... on a 135 I have to go maybe 20 cm away to see a decent rectangle (so that's why tele lens hoods are round, huh?), but a 35 plus 35 mm extension would be OK.