[iClick]

I was out taking pictures last night at an event and I was using my K200D with an old SMC-Pentax-A 1:1.4 50mm lens. I shot this pic at 800 ISO, 1/45 at f/1.4, +.5 exposure bias . I'm getting a lot of shots where the lights in the background are out of focus as I want but are hard edged and not soft edged like I want. They are too distracting in my opinion. Is this an example of bad bokeh? What can I do to soften them up, in camera, not in pp?

Thanks for your comments and suggestions.

Attached Images

 

[kxr4trids]

One solution would be to take a quick jog backwards and affix a longer lens to the camera.

[jboyde]

Its not bad bokeh more than its an example of how OOF areas look with a standard 50mm compared to a telephoto lens (eg 105mm).
For you to get what youre looking for out of a 50mm your subject (sign) would have to be pretty close while the background (lights) have to be further out. The more telephoto your lens is, the more the effect gets magnified. I know someone on this board had a link to examples of the same portrait subject shot at different focal lengths and how the relation of the degree of the OOF areas changed. If you had, say, a 135 mm lens then your lights probably would have been a mesh of gray, but then again I don't know how far away they were.
I hope I'm not over complicating this. Bokeh being good or bad is subjective since people like to go for the "creamy" background, which would be hard to get with a background of small points of light. But honestly, I dont find it distracting in this shot.

[Ira]

No such thing as good bokeh versus bad.

It all depends on the shot.

And the people who look at your image, who don't even know what the hell bokeh is.

[JonPB]

I agree that there is no "good" or "bad" bokeh in the absolute sense. And it doesn't matter whether viewers know what "bokeh" is, it matters whether visual elements detract from conveying the intended image. I agree that the bokeh in this image is distracting.

One issue is that there's no transition between in-focus and out-of-focus. Harsh bokeh (when hard-edged out-of-focus points are considered negative) can be partially mitigated by having an object that transitions from in-focus to out-of-focus, so the eye can "read" what the bokeh represents. Here, we have in-focus and quite out-of-focus, but because the out-of-focus is mostly bright points with hard outlines, it doesn't look like soft, out-of-focus blur. I think my brain sees the hard edges and tries to figure out whether it is in-focus or not. We need some indicator that it is truly out of focus.

That said, the lens determines the bokeh. Usually, stopping down a little will make the blurry areas smoother, so you could experiment at f/1.7 or f/2.0.

With this image, you can see that the bokeh is smooth at the center of the image, but develops a lopsided halo toward the corners. Knowing that, you can place the out-of-focus elements in the center, and the in-focus elements at the corners; or, at least avoid bright and out-of-focus points of light at the corners.

Photozone.de does a good job of looking at the bokeh in their lens reviews. Even if you're not going to purchase a new lens, you can browse through their reviews and get a feel for different kinds of bokeh.

[wlachan]

It is bad bokeh alright, and what makes it bad is the bright edge. With certain OOF backgrounds, this kind of bokeh will render them like double images and makes you dizzy. It is the characteristic of the lens and nothing you can do. Unfortunately, this kind of bokeh is rather typical with Pentax lenses. Back then Minolta lenses were known to produce awesome bokeh, and so are Canon lenses. Nikkors are also known to produce harsh bokeh. However, each lens is different regardless of brand. Even some Leica were demonstrated to produce very ugly bokeh.

[KxBlaze]

Bokeh is definitely up to personal preference but I do think that bokeh is too "hard" as you already stated. I have the M 50 1.7 and I love its bokeh but I have never done a test to see the difference from 1.7 to 3-4 and would be curious to see if I can notice if it's harder/softer.

[Just1MoreDave]

Bokeh is definitely up to personal preference but I do think that bokeh is too "hard" as you already stated. I have the M 50 1.7 and I love its bokeh but I have never done a test to see the difference from 1.7 to 3-4 and would be curious to see if I can notice if it's harder/softer.

With OOF highlights like the example, you'd probably see distracting hexagons instead.

I guess this example is "bad" but that's why I like this lens.

[orly_andico]

The sharp edges are due to over-correction. Also notice that off-center the OOF highlights are oblong "cats eye."

You can get rid of both by stopping down to around f/4. The problem is that the highlights will become polygonal (depending on how many blades on your aperture iris).

If you look at the blur circle "on edge," your image would show a sharp edge, with a "hole" or "pit" in the center (plotting brightness versus radius of the blur circle). The ideal OOF highlight would be Gaussian (e.g. brightest at the center, fading off gradually as you approach the edge). Unfortunately, if you have Gaussian OOF, you'll also have tons of spherical aberration. Most people care about the in-focus image, not the out-of-focus image, hence most modern lenses, which are very sharp, have hard-edge OOF highlights.

Almost all 50mm lenses are Planar designs (except Leica, apparently..) and would show the above characteristics, although the polygonal issue can be fixed by more blades, e.g. the dirt-cheap Russian Helios-44 has a 15-blade iris.

The Zeiss Sonnar which has less correction (e.g. softer) and is slower (f/2 vs f/1.4) than Planar designs, has less hard-edge OOF because it has some residual uncorrected spherical aberration.

A couple Pentax lenses which are legendary for their bokeh are (1) the rare 58mm f/2.4 Takumar, which is a Sonnar and not a Planar (58mm is the shortest you can make a Sonnar for SLR, the 50mm Sonnars are for rangefinders). And (2) the Super-Macro-Takumar 100mm f/4 which is a Heliar design.

Generally all/most 135mm's are Sonnar derivatives and have more pleasing bokeh than the 50mm's. Although the Pentax-M 135/3.5 is one of the less stellar examples. Personally the CZ Jena (M42) 135/3.5 and Pentacon 135/2.8 have some of the nicest bokeh around, as well as the cheap Russian Jupiter-137.

Minolta made a 135mm f/2.8 "STF" with a secondary circular aperture to give really gorgeous OOF highlights. Sony has reissued this lens as 135mm f/2.8 [T/4.5] STF and positions it as a movie lens (hence the T-stop rating). Ditto for the Minolta 85mm f/1.4, they took pains to make the aperture iris blades curved to give a circular opening at most/all apertures.

[orly_andico]

Got a bit off-topic there. Quite a number of lenses with pleasing bokeh but if you want to try things out without spending a lot, I'd suggest the following, from short to long..

1) Mir-1v / Mir-1B, 37mm f/2.8 (Zeiss Flektogon copy), many blades

2) Helios-44, 58mm f/2 (I got mine for $2!! plus postage of course, and it was beat up) -- Planar/Biotar copy, but many many blades

3) Jupiter-9 85mm f/2, Sonnar copy, "legendary" but very busy bokeh wide-open, stop down to f/2.8 to clean things up; many blades also

4) Jupiter-37 135mm f/3.5, also a Sonnar, also many blades

5) Pentacon GDR 135mm f/2.8, not such what it is, some versions have few blades and some have many, the many-bladed ones are called the "bokeh monster," I have the few-bladed version though

6) CZ Jena 135mm f/3.5 Sonnar, few blades



All of the above can be snagged for ridiculous prices on ebay if you're patient. Sad to say, I sold off all my Russian glass already.

[RioRico]

@orly,
I have copies of those first four and they're indeed bokeh wizards. Mir-1BV: M42, 8 iris blades. Helios-44M: M42, 8 blades. Jupiter-9: M39, 15 blades. Jupiter-11: M39, 11 blades. Number of blades alone doesn't guarantee admirable bokeh. My Schneider Betavaron 50-125 enlarger zoom has only 5 blades forming a strong starry diaphragm, but creamy smooth bokeh -- maybe its 800g of glass has something to do with that! Ah, but those Russians are often no longer cheap. The suppliers seem to have a cartel with high starting prices. Bother...

On a nearby thread on preset lenses, it's noted that many 2-ring presets have 10 or more blades. And some common lenses have interesting variants. My small aluminum Exakta-mount CZJ Tessar 50/2.8 has 12 blades where the large M42 version has only 5. And it *is* a bokeh monster! Although not quite in the same league as an old Vivitar Tele 200/3.5 with 18 blades.

obTopic: Those hard-edged specular highlights can be dramatic when exploited correctly, but distracting otherwise. Blobs of any geometry can be used as design elements. And I hold the view that bokeh can and should be tailored to one's purpose, to provide needed contrast. Sharp detailed subjects may demand creamy smooth bokeh. Creamy smooth subjects may demand harsh jagged bokeh. Et cetera.

[orly_andico]

The Taylor-Hobson Cooke formulations are also considered to have very desirable bokeh; and are primarily used as cine lenses.

[newarts]

........
The ideal OOF highlight would be Gaussian (e.g. brightest at the center, fading off gradually as you approach the edge). Unfortunately, if you have Gaussian OOF, you'll also have tons of spherical aberration. Most people care about the in-focus image, not the out-of-focus image, hence most modern lenses, which are very sharp, have hard-edge OOF highlights.........

A couple Pentax lenses which are legendary for their bokeh are (1) the rare 58mm f/2.4 Takumar, which is a Sonnar and not a Planar (58mm is the shortest you can make a Sonnar for SLR, the 50mm Sonnars are for rangefinders). And (2) the Super-Macro-Takumar 100mm f/4 which is a Heliar design..........

Generally all/most 135mm's are Sonnar derivatives and have more pleasing bokeh than the 50mm's. Although the Pentax-M 135/3.5 is one of the less stellar examples. Personally the CZ Jena (M42) 135/3.5 and Pentacon 135/2.8 have some of the nicest bokeh around, as well as the cheap Russian Jupiter-137.......

Orly,

Thank you for the informative posts.

Is it true that a sharp lens must always have hard edged Bokeh or is it simply a general rule?

Is the sharp-edged aperture shape seen on most out-of-focus light points simply mechanical vignetting of a wider blur function or are things more complex than that?

Dave

[Ira]

Here's the nutty bokeh of the Helios 44-2, which many people say is ridiculous. I call it psychedelic:

[orly_andico]

Orly,

Thank you for the informative posts.

Is it true that a sharp lens must always have hard edged Bokeh or is it simply a general rule?

Is the sharp-edged aperture shape seen on most out-of-focus light points simply mechanical vignetting of a wider blur function or are things more complex than that?

Dave

I'm no optical expert

The hard-edge OOF highlights are indicative of overcorrection for spherical aberration. If you correct all the residual SA in focus, then the lens will be overcorrected for out-of-focus areas. Since most people like maximum correction for the in focus image (heh!) then generally yes the OOF image will have hard edges.

But there are exceptions... in general simpler lenses have less correction but can remain sharp (at the center) e.g. Tessar formulations.

Have no idea about the vignetting etc. But do consider that when you stop down, you're chopping off the edges of the lens.. and lens designers spend most of their time correcting for off-axis aberrations. Almost any lens (even a singlet) is sharp at its center. So my guess is, when you stop down, you're only taking the center where the correction is already very good (and where correction/overcorrection is not needed as much) which is maybe why the hard edges become less apparent.

That said I've never seen Gaussian bokeh, at best what I've seen is neutral bokeh (e.g. the edges are not brighter than the center).