
Are optical designers designing (and re-designing) lenses today to create better bokeh at the expense of optical characteristics such as diffraction spikes (popularly known as star effects/sunstars/starbursts)? Is the trend toward creating “better” bokeh sending the lens-made diffraction spike the way of pay phones, SLR cameras, manual transmissions, and the internal combustion engine?
The photography world, especially on the Internet, is full of debates. Canon vs. Nikon. Full-frame vs. APS-C. Manual vs. Auto. DSLR vs. Mirrorless. Prime vs. Zoom. UV Filter vs. No Filter.
Is bokeh vs. diffraction spikes the latest?
Photographs © Todd Vorenkamp

What is bokeh?
“Bokeh” is a Japanese word that refers to the how out-of-focus highlights are rendered in an image. The aesthetics of bokeh are subjective. In the simplest terms, how a lens renders bokeh is a function of the optical design (lenses), shape of the aperture diaphragm, focus, depth of field, and finally, the background itself. I dig deep into the topic of bokeh in this article: Understanding Bokeh.

What are diffraction spikes?
“Diffraction spike” is the technical term for beams of light that extend outward from point light sources in photographs. As mentioned above, they are sometimes called starbursts, sunstars, star effects, sun flares, etc.
The camera renders them in photographs when light diffracts around the edges of the blades of an aperture. I dive deeper into this topic in the article 6 Tips to Create Compelling Star Effects, Sun Stars, Starbursts, Sun Flares, or Diffraction Spikes in Your Photographs.

Interestingly, your own vision can experience diffraction spikes when bright light shines through your eyelashes.
The Common Factor—Aperture Diaphragms
The overall shape and characteristics of a lens’s aperture can and will affect how the optic renders bokeh and diffraction spikes in photographs. Today’s iris-type aperture diaphragms are the lens’s most prominent factor in both optical effects.
I am not an optical engineer (or an engineer of any kind), so we are going to discuss this awesomeness in layman’s terms and avoid complex formulas and mathematical fun. If you are an optical engineer, please feel free to join the discussion in the Comments section after the article—I welcome your thoughts and input!

Aperture Diaphragms and Bokeh
The most common belief is that the more perfectly circular an aperture diaphragm is, the more beautiful the bokeh it will render. We know from my bokeh article and have already mentioned that there are other factors influencing how bokeh is rendered, but a primary player is the shape of the aperture diaphragm.
In the old days of photography, drop-in Waterhouse Plates with circular openings served to control or limit the amount of light that reached the film. Eventually, more modern lenses were equipped with adjustable iris-type aperture diaphragms.

The openings of mechanical iris-type aperture diaphragms have non-circular shapes—they form polygons. The number of sides of the polygon are determined by the number of blades of the diaphragm. Depending on the design of the diaphragm, that polygon can, and sometimes will, approximate a circle as it is opened and closed. How closely does the diaphragm opening approximate a circle instead of a polygon? That depends on the diaphragm’s design and, sometimes, the f/stop selected. Often the widest possible aperture is the most circular. Sometimes the narrowest aperture is also circular.
If you study the out-of-focus highlights of some lenses, you can see the actual polygon shape from the aperture diaphragm revealing itself in the bokeh. Although bokeh quality is subjective, the most common opinion seems to be that polygonal bokeh shapes can be harsh and distracting. I am in the minority here because I sometimes enjoy the polygonal shapes in some bokeh.


Aperture Diaphragms and Diffraction Spikes
Diffraction spikes are created when light passes the edges of the diaphragm. (Light is diffracted when it passes by an object, so a perfectly circular aperture will create diffraction spikes, but the spike on one side of the circle will overlap the spike from the opposite side of the diaphragm.) Usually, more distinct spikes are created by flatter edges of the aperture diaphragm.
Mobile-device cameras generally have circular apertures and do not create diffraction spikes. You can use portrait mode to simulate bokeh, but smartphones and tablets live in a world without sunstars!
The other profound effect that the aperture diaphragm has on diffraction spikes is that it determines the number of spikes visible. Diaphragm polygons with an even number of sides produce one spike per side of the polygon—a six-bladed aperture will produce six rays of light. Apertures with odd numbers of blades produce double the number of spikes than edges—a 7-bladed aperture will produce 14-spike sunstars.
Why does an odd number of blades create two spikes per blade while an even number of blades produce one spike each? With an even number of blades, the spikes on opposite sides overlap—creating a single spike.
Rounded Aperture Blades—Friend or Foe?
Bokeh has been one of the most discussed topics in photography (at least on the Internet) over the past few years. Likely because of this, lens manufacturers have been placing an emphasis on designing (or redesigning) lenses with an eye toward creating pleasing out-of-focus highlights—including by making the opening of the aperture diaphragm as round as possible.
There are two ways to encourage an iris-type aperture diaphragm to approximate a circle: 1) you can increase the number of blades or, 2) you can curve or round the blades. One downside of increasing the number of blades is added mechanical complexity. Therefore, many lens companies are rounding aperture blades on new lenses or when re-designing older lenses.

Rounding the blades helps create a more circular aperture opening and reduce the hard edges of the aperture polygon and, theoretically, improve the bokeh. Artie M., a frequent Explora reader, commented in the diffraction spike article saying, “A lot of lens manufacturers spent the 2010s rounding every aperture blade they could get their hands on.” A Pentaxian, he has seen the company re-design existing lenses primarily just to re-release them with rounded aperture blades (and new-and-improved coatings in many cases).
Rounded aperture blades have been in lenses for years, but, interestingly, the current rounded blade movement (or at least the marketing promoting it) seems confined to the mainstream lens manufacturing companies (Canon, Nikon, Pentax, FUJIFILM, Sony, etc) while “third party” lens companies seem content with straight blades or not promoting their curved blades.

Testing Diffraction Spikes vs. Bokeh
Artie continues, “Starbursts, run and hide! The bokeh mob is coming after you!”
The mission: To find if there is evidence that rounded blades are improving bokeh while silently killing diffraction spikes, which seem to thrive in diaphragms with straight aperture blades.
To answer this query, I decided to get a selection of my own lenses, plus some new ones, outdoors and in the studio to see what the images revealed.
The “Controlled” Experiment, Part I—Pentax 21mm f/3.2 smc vs. HD + Pentax 21mm f/2.4 HD
Wouldn’t it be great to conduct this test with optically identical lenses that have different aperture diaphragms?
Ricoh recently released a version of its popular Pentax 21mm f/3.2 pancake lens. The Pentax smc-DA FA 21mm f/3.2 AL Limited‘s replacement is the HD Pentax-DA 21mm f/3.2 AL Limited. The new “HD” lens features the same optical design—now with HD and SP coatings—and, you guessed it, rounded aperture blades.
New round iris diaphragm
A round iris diaphragm is used to enable imaging of soft round out-of-focus (bokeh) effect for illumination, shimmering on water surfaces, and other point light sources.
Here are the results:
Conclusion
While the rounded diaphragm of the new 21mm HD lens did not kill the diffraction spikes, it degraded them―they are not as sharp as they were on the 21mm smc version of the lens. Bokeh is a bit smoother with hints of diffraction spikes removed in the HD version, but 21mm lenses aren’t usually known for being bokeh machines. Flare is better controlled with the new coatings of the HD version, as well.
The “Controlled” Experiment, Part II—Pentax 31mm f/1.8 smcP vs. HD
Almost identical to the re-design of the 21mm, Ricoh’s Pentax smcP FA 31mm f/1.8 Limited’s replacement is the HD Pentax-FA 31mm f/1.8 Limited—new coatings and a rounded aperture on the “HD” version.
Pentax published the following at the time of the new HD 31mm’s release:
Round-shaped diaphragm for beautiful out-of-focus rendition
All lenses feature a completely round-shaped diaphragm to optimize the performance of their distinctive optics. This diaphragm produces a natural, beautiful bokeh (out-of-focus) effect, while minimizing the streaking effect of point light sources.
It seems that the official Pentax press release promises degraded diffraction spikes.
Here are the images.
Conclusion
Interestingly, the rounded blades of the new HD 31mm seem to have enhanced the diffraction spikes while almost imperceptibly rounding the bokeh and removing hints of the spikes surrounding the bokeh. Flare is better controlled with the new coatings of the HD version, as well.
The “Legacy” Comparisons
Taking a look at some of my lenses and photographs, we see that the Nikon AF NIKKOR 50mm f/1.8D (and its older, optically identical, manual focus Nikon NIKKOR 50mm f/1.8) has always been one of my favorite lenses for sunstars (now superseded by the 9 blades of the Nikon NIKKOR 50mm f/1.2). Here is a comparison of its straight 7-blade aperture (our spec sheet says they are rounded) to the rounded 9-blade diaphragm of the Nikon AF-S DX Zoom-NIKKOR 17-55mm f/2.8G IF-ED lens.
The Nikon AF Zoom-NIKKOR 80-200mm f/2.8D ED and my original version of the AF-S NIKKOR 70-200mm f/2.8G ED VR (Version II linked here) were both equipped with 9-bladed diaphragms. While the Internet disagrees about rounded vs. straight blades for both lenses, the degraded sunstar performance would indicate that the 70-200mm lens has rounded blades and—but still a nice diffraction spike—while its older sibling, the 80-200mm, has straight blades and prominent spikes.


Are Rounded Blades Wrecking Diffraction Spikes?
While there is anecdotal evidence around the Internet that rounded aperture diaphragms are the enemy of good diffraction spikes—and a cross-section of my own lenses/photographs agree—the test of the two Pentax 31mm versions shows that this is not the case for all optics.

Can diffraction spikes and good bokeh cohabitate in the same lens?
The Ricoh website for the HD PENTAX-D FA 21mm f/2.4ED Limited DC WR lens claims that the 8-bladed aperture diaphragm has been designed for good bokeh and good diffraction spikes.
Eight-blade circular diaphragm for beautiful light beams and bokeh effect
This lens incorporates a newly designed diaphragm unit with a circular diaphragm, which produces a beautiful bokeh (defocus) effect at open and larger apertures. This unit also features an eight-blade diaphragm mechanism to capture beautiful light beams in nightscape photography.
Bokeh vs. Diffraction Spikes
Spending days and nights thinking about this photographic debate, I have come up with a short list of miscellaneous thoughts about bokeh vs. diffraction spikes.
-
The aesthetics of bokeh are completely subjective. Is rounded bokeh better than a subtle hint of a polygon? That is up to the viewer. And, as mentioned above, the rendition of out-of-focus highlights is a function of more than just the shape of the aperture diaphragm.
-
Some photographers love diffraction spikes. Others feel disinterested in them. One could argue that they enhance some images but create distractions in others. Like bokeh, it is subjective.
-
Wide-angle lenses have never really been known for highlighting bokeh, so I am perplexed as to why wide-angle lenses would be candidates for rounded aperture blade treatment unless part of the goal is to reduce diffraction spikes.
-
Bokeh is more prominent at wider apertures and diffraction spikes are more prominent at narrower apertures. This naturally plays into benefiting both phenomena since the diaphragm shape becomes more circular as it gets wider, making the case for rounded blades more complex.
-
In the case of some lenses, rounding the aperture blades does not ruin the lens’s ability to create diffraction spikes. In the case of others, it certainly appears to have prioritized bokeh over spikes.
Some lenses certainly prove that diffraction spikes and good bokeh can coexist in the same lens, but, for fans of diffraction spikes, it looks like the trend of rounding aperture blades is making good diffraction spikes harder to find in modern lenses.
Are you a sunstar fan? Are you forever in search of perfect bokeh? Are you an optical engineer who can shed further light on this subject? We would love to hear from you. Please share your experiences with us in the Comments section, below!
13 Comments
Hi Todd!
Great to see this article! I enjoyed it. Sorry to have taken so long to get here, but please know I read every word.
It’s our EYELASHES that cause diffraction spikes in our own eyes?? That really stuns me.
Fan though I am of starbursts/sunstars/diffraction spikes, I hadn’t understood until reading your article that it was the edges of the aperture blades that caused the spikes. I had (wrongly) thought that the spikes came at the angled corners of the polygon formed by the closing blades. (Like some of the most pernicious wrong ideas, this misunderstanding lasted so long because it made sense and had predictive power.) Something I love about photography is that there is always, always new stuff to learn.
I think the FUJIFILM XF sunstar-to-bokeh-ball GIF is a great demonstration, but I sure wish the loop went a little slower. (It could be that GIF speed is dependent on my own machine, though.)
The Leica Summicron image raises all sorts of questions. You have two sunstars (at the plane of focus??) but the whole image appears out of focus near-to-far. Was any part of that shot in focus? And if not, how are the sunstars happening? That orange lens flare on the right edge of the frame is amazing.
That split-slider image you've made of the aperture blades on the two FA 31mm Limited lenses is a nice bit of community service. I had imagined the blades would have much rounder edges. The polygons look nearly identical to me. No idea why the spikes are so much sharper on the updated HD-FA version of that lens. I’m completely confounded that the FA 31mm aperture "improvement" produces such different results from the DA 21mm aperture "improvement." (As a starburst fan, I think the Pentax might indeed have improved the HD-FA 31mm, no quotation marks about it.) But both versions of the 31mm have a weird grid pattern of lens flare that reminds me of a macro photo of an LCD screen, or one of those ancient Lite Brite toys from the 1980s.
A great article on a favorite topic, Todd. Well done.
Hey Artie,
Great to hear from you and thanks again for inspiring this article!
I, too, always thought it was the corners of the blades that caused the spikes, but learned differently when writing the sunstar article years ago. I still find it hard to wrap my head around the fact that it is caused by the "flat" surfaces, and not the corners.
Re: FUJI GIF...not sure we have a way to slow that down. I don't work on the design side, so the speed might be "baked into" the code?
Re: Leica image...no part of the image was in focus as I turned the ring to its minimum focus distance for maximum bokeh simulation. It is interesting how sharp the diffraction spikes on the left remain, even way out of focus. And, yes, that flare is pretty cool! I got a new appreciation for that lens while doing this article!
Re: Pentax 31/21...I apologize for deepening the Pentax "improvement" mystery a bit! :) Based on our previous conversations, the results were very surprising.
I am glad you enjoyed the article. It was a lot of fun to assemble and think through!
Thanks for reading!
Best,
Todd
The geometry of the optical effect of an edge is simple so can make a pretty good prediction of the sunstar pattern based entirely on just the appearance of the iris opening.The diffraction spread is aligned at 90 degrees from any edge that the light encounters.
A single straight edged blade has a diffraction pattern with two very narrow "half" spikes along the perpendiculars relative to the edge. Let's call one direction "inward" and the other direction "outward" (relative to the side of the edge that is not blocked). 6 very straight edges in exact symmetry will show 6 spikes because the inward half-spike from one side exactly overlaps the outward half spike from the other. With a perfect hexagon those will be perfectly superimposed and therefore doubled up in brightness by simple addition. If that hexagonal opening is instead very asymmetrical, there can be as many as 12 spikes in total because the additive alignments didn't happen. And they will each be half as bright vs the perfect case. I still keep a lens from the early 1970's because of the perfect set of 6 narrow spikes it produces, which few modern lenses (and none of the expensive ones!) can do.
Curved spikes have a range of edge angles, therefore the spikes will fan out into that same range of angles. They are broad but not necessarily uniform in brightness as seen in images above. Because the light isn't concentrated into needle-narrow spikes that pattern is dimmer than what occurs with straight blades, and they fade out sooner as you move away from the point light source.
Rounded edge blades typically do not not produce perfect circles at every aperture setting. At most f-numbers you get polygons with bulging edges that are close to circles but not quite there.
I have seen lenses that have extremely irregular polygon shapes of the iris with some sides being long and others short. Those sunstars don't align in a regular spacing. Whether you prefer rounded or straight blades those sunstars probably won't look good to anyone.
Hey Tony,
Awesome stuff! Thank you for the explanation!
You and Daunte need to re-write the Wikipedia article on diffraction spikes so that I can look smarter when I write these articles!
Thank you, so much, for reading and taking the time to post this informative comment. What lens are you holding onto from the 70's, may I ask?
Best,
Todd
Thank you, Tony. You and Todd have explained something I'd misunderstood for years.
Unlike your 1970s keeper, many of my old lenses have irregular polygons at ƒ/11 and so on. They seem to work just fine, so I don't think it's a question of sticking or wear and tear, just an aspect of lens construction/design that the lensmakers weren't too worried about.
"Why does an odd number of blades create two spikes per blade while an even number of blades produce one spike each? With an even number of blades, the spikes on opposite sides overlap and cancel each other."
When spikes on opposite sides overlap they reinforce each other, not cancel. Same is true of diffraction from a circular aperture but the diffraction is more diffused due to the curved aperture edge.
Hi Douglas,
Thank you for that correction and the information. That makes more sense to me! I will have the text updated ASAP.
I appreciate you helping make my information more accurate! Thanks for reading!
Best,
Todd
Thanks for the great article! One thing I might add is that on many (most?) lenses, if you're shooting wide open the blades are fully out of the way and you get perfectly circular bokeh (in the center of the image at least; as you approach the edge of the image most lenses give you "footballs" because you're getting the intersection of two circles). By the same token you typically won't get diffraction spikes when wide open (of course, you still get diffraction, it's just in all directions).
"A perfectly circular aperture will create diffraction spikes, but the spike on one side of the circle will be canceled out by the spike from the opposite side of the diaphragm." I don't think cancellation is the issue. If that were true, then wouldn't the same argument apply to any aperture with an even number of blades? I think it's more that the diffraction is in all directions so very little in any one direction, and in particular, no more in one direction than another, so no "spike".
(Also, I hate to be that guy but where you said dodecahedron I think you meant dodecagon.)
Hi Christopher,
Thanks for the kind words!
You are correct: on most lenses the blades are totally out of the way at maximum aperture...so, a circle aperture is present and sunstars are not.
When I researched diffraction spikes, I was not totally sold on the "cancellation" thing either, but, when I think about it, I believe that is a simplistic and accurate way of presenting it. Thinking about it too much makes my hair hurt. :) Check out the Wikipedia article on "diffraction spikes" for a good discussion of it and feel free to circle back to continue this thread, if you would like.
I want to learn (without hair pain, if possible)!
And, you can be "that guy!" I will have the text corrected. Geometry was the ONLY math class I ever kept my head above water in...but apparently, that knowledge is fading!
Thanks for reading!
Best,
Todd
As you said, the use of diffraction spikes and bokeh in a picture, is completely subjective. Therefore, no one "right answer" can exist and all we can do is conduct a poll of sorts. You either like or dislike them, and no one can argue with you on that because what a person finds attractive is a matter of personal taste. That said, I find diffraction spikes to be irritating, and the rounder/softer the bokeh, the better. Less spiking and less noticeable bokeh (softer/rounder seems less obvious, at least to me) both lead to less distraction from the main subject of the photograph.
Hey Matthew,
Well said! I guess it is the nature of "subjective" things to create differing opinions and debate...and the Internet to give everyone a platform to enter that conversation! :)
Your votes will be included in any polling we do!
Thanks for reading!
Best,
Todd
I agree with you. I almost hate spikes. We have to accept it as unavoidable defects of the lenses the same we accept de out of focus areas in a photo. The difference is that we can (somhow, to some extent) control the second to our benefit to underline or supress some parts of the scene. As Todd said, the quality of the bokeh y 100% subjective and rounded or geometrical, both may enhance or not a photograph, depending on the object we're pointing to.
Spikes are almost out of our control and one only can include or not bright points in the composition and thats all about we can do. With a perfectly round diaphragm we wonli get infinite spikes thar are evenly distributed all over the frame and would be shown as Flare if there are enough contrast in the scene.
I hate spikes, sunstars and flare. ;-)
Hey Roberto,
Your objection to the diffraction spike is noted! May I suggest cameras with Waterhouse aperture plates featuring perfect circles? :)
I wish there were more aperture diaphragms that allowed for spikes when stopped down in mid-to-small apertures, and were circular at large/wide-open apertures to allow the photographer to get diffraction spikes when wanted while preserving the coveted round bokeh.
Thanks for reading and sharing your thoughts!
Best,
Todd