The Naked Truth About Bokeh

Started Jul 22, 2016 | Discussions
Marianne Oelund Veteran Member • Posts: 7,062
The Naked Truth About Bokeh
75

Several years ago, I acquired a handy optical device for producing high-quality pinpoint light sources, for study of lens acuity and aberrations.  I have finally gotten around to using it, but not for lens acuity studies.  Rather, I've found it is very suitable for studying the COC, or blur circle, profiles of a lens.  I couple it with a green LED light source, which further helps by narrowing the wavelength bandwidth and eliminating most chromatic aberration effects.

Looking at the blur circle formed by a lens from a pinpoint light source gives the clearest view possible of how different bokeh characteristics are formed.  It strips away the confusion and clutter, or cloak if you will, which masks the essential lens behavior in photographs, leaving it rather naked for all to see.

I will begin with two lenses which are known for pleasing bokeh:  The Nikkor 58/1.4G and 85/1.4G.  The 58G gives an especially "dreamy" look to its images when used wide open, or even at f/2, and has extraordinarily smooth background bokeh.  Looking at the pinpoint-source blur circles in the background direction, we see that the blur circle fades out gradually toward the perimeter and does not have any abrupt cutoff.

The 85/1.4G represents more typical lenses with under-corrected SA; although its blur circle does decrease in intensity toward the boundary, there is a well-defined perimeter which does not produce such a soft bokeh:

Unique 58G wide-aperture background blur circle with diffuse perimeter and bright central peak

The 58G also has a rather bright and small central peak which tends to maintain high resolution at hard image edges, while contrast falls off gradually due to the highly diffused outer area of the blur circle.  This is readily seen in images which have white lines on a black background.  The precise plane of focus at f/1.4 is difficult to decide upon; if you view these scales at different distances, you are likely to pick different optimum-focus locations for the f/1.4 scale:

The closer you look, the further back you are likely to judge the precise focus plane for the 58G at f/1.4

There is quite a bit of information in these figures, and much more could be said, but I will leave the discovery to you.  Hopefully you will find some interesting surprises.  

As I have time, I will add posts for other lenses.  The next candidate is the Sigma 50 Art, which has a reverse characteristic to the above lenses.  You will see that it is a good choice for smooth foreground bokeh.

-- hide signature --

Source credit: Prov 2:6
- Marianne

BirdieMaker Regular Member • Posts: 436
Re: The Naked Truth About Bokeh

What a post. Front booked is hardly ever talked about but definitely a factor. Coming from Canon, my 135/2 was amazing in so many ways, but I'm sure this played into it. Thanks for the work.

Valerie

-- hide signature --

Birdie maker - it's a golf thing, not BIF

 BirdieMaker's gear list:BirdieMaker's gear list
Sony RX100 III Canon PowerShot S95 Nikon D500 Nikon D5 Nikon AF-S Nikkor 14-24mm f/2.8G ED +6 more
YRUNVS
YRUNVS Contributing Member • Posts: 812
Re: The Naked Truth About Bokeh

Many thanks for this comparison! I found it both interesting and informative. I was wondering whether you planned to do similar analysis with the 135 DC (though I do understand that would be a bit of an undertaking with the different defocus settings).  In any case much appreciate your many contributions to the site.

 YRUNVS's gear list:YRUNVS's gear list
Olympus XZ-1 Nikon D600 Nikon D7100 Nikon D610 Nikon AF-S Nikkor 70-300mm f/4.5-5.6G VR +12 more
syberman7 Contributing Member • Posts: 814
Re: The Naked Truth About Bokeh
3

Marianne Oelund wrote:

An inspired post once again. Thank you.

The 58 G never seems to loose sharpness behind the focal plane - it just gets murkier and murkier! Very strange behaviour. I guess that's what makes the bokeh so pleasing, and what irks the 'sharpness' freaks the most.

(In case you don't understand my reasoning - I'm talkiing about the central green dot that seems to stay sharp, while a halo gradually builds around it as it falls behind the plane of focus.)

 syberman7's gear list:syberman7's gear list
Nikon D810 Olympus E-M1 Olympus E-M1 II Nikon AF-S Nikkor 24-70mm f/2.8G ED Nikon AF-S Nikkor 58mm f/1.4G +10 more
Andre Yew Regular Member • Posts: 360
Re: The Naked Truth About Bokeh

Nice work Marianne! For focusing, if you have time, I look for the part which is not green or magenta. The focal point should be buried in that in-between mush somewhere.

Nikonmaniac0620 Junior Member • Posts: 47
Re: The Naked Truth About Bokeh
6

I've never seen such a scientific measurement towards bokeh characteristics of different lenses!

The two concepts: "scientific measurement" and "bokeh" have so far been like polar opposites. Yet you managed to combine them into one single coherent work.

Keep up the good works!

OP Marianne Oelund Veteran Member • Posts: 7,062
Questions to Ponder

Refer to the composite images in the original post, and see if you find these questions thought-provoking.  This is not a test and you do not need to post answers (unless you discover something that you consider important):

1a.  Compare the overall sizes of the foreground and background blur circles at -60mm and +60mm.  Where are they similar?  Where are they different?  When they are different, what does this suggest?

1b.  Blur circle size should reduce as the lens is stopped down, for example the f/2 blur circle should be only 70% of the diameter of the f/1.4 blur circle.  Look at the foreground blur circles for the 58G at f/1.4 and f/2:  Why are they almost the same size?

2a.  We can expect complex multi-element lens designs to potentially have multiple blur-circle features.  Where do you see evidence of two different-size blur circles being superimposed (in particular, study the f/1.4 series)?

2b.  In the f/1.4 cases that you find from question 2a, what would happen if you subtracted away the f/2 blur circle that is below it?  Do you get a simpler structure?  What would the result of that subtraction represent?

3.  Look at the background circles for the 85/1.4G at f/1.4 and f/2.  Which ones have a softer perimeter?  Would this suggest that the 85/1.4G could have smoother bokeh at f/2 than at f/1.4?

4.  For each lens:  How much shift is there in best focus, from f/1.4 to f/2?  How much shift is there from f/2 to f/4?

5.  In the background blur circles, which portion of the image is from light rays passing through the lens perimeter, and which portion is from central rays?  Does this differ for the foreground blur circles?  [Check your answers against the final answer for question 2a.]

6.  There are a few lens designs (not Nikon) which incorporate an ND filter to help fade the perimeter of the blur circle when the lens is used wide open.  Would the 58G benefit much from such a filter for its background bokeh?  For its foreground bokeh?  How about the 85/1.4G?

7.  Look at the second image in the opening post (macro rails), for the 58G f/1.4 case.  If you view this at the size it appears in that post, where does best focus appear to fall on the scale?  Open the image and display at 100%:  Where on the scale do you find the sharpest detail (regardless of contrast)?  Are the line edges sharper than those in the 85/1.4G f/1.4 case?

-- hide signature --

Source credit: Prov 2:6
- Marianne

snellius Contributing Member • Posts: 834
Spherical aberration

Spherical aberration is the cause of what we observe here.

The first stop will change much more in the background blur than to the foreground blur

paulski66
paulski66 Veteran Member • Posts: 3,171
Re: The Naked Truth About Bokeh
1

I can't lie...I was hoping for some tasteful nudes with shallow Dof.

 paulski66's gear list:paulski66's gear list
Panasonic Lumix DMC-LX5 Nikon D750 Nikon AF-S Nikkor 50mm f/1.8G Nikon AF-S Nikkor 85mm f/1.8G Nikon AF-S Nikkor 70-200mm f/2.8G ED VR II +2 more
OP Marianne Oelund Veteran Member • Posts: 7,062
Complex Lenses
2

snellius wrote:

Spherical aberration is the cause of what we observe here.

Good basic illustration, thank you.

But what I am most interested in exploring here, goes beyond basic SA, to the particulars of complex lens designs.  The f/1.4 blur circles for the 58G and 85/1.4G exhibit many detail differences beyond what one expects from single-element SA.  This is what question 2 alludes to.

There are many more lenses to look at.  Hope you will enjoy the journey.

-- hide signature --

Source credit: Prov 2:6
- Marianne

OP Marianne Oelund Veteran Member • Posts: 7,062
Enticement?
3

paulski66 wrote:

I can't lie...I was hoping for some tasteful nudes with shallow Dof.

I appreciate your honesty.

And if I am to be as honest, I must admit to realizing that the thread title could generate that type of interest - but I would hope that you are not disappointed with what you are finding here instead.

-- hide signature --

Source credit: Prov 2:6
- Marianne

Paul B Jones
Paul B Jones Senior Member • Posts: 2,460
Re: The Naked Truth About Bokeh

Wow. Accessible and erudite at the same time. Great work. Thanks.

-- hide signature --
 Paul B Jones's gear list:Paul B Jones's gear list
Canon EOS-1D Mark IV Canon EOS-1D X Canon EOS 7D Mark II Canon EOS-1D X Mark II Canon EOS M6 +14 more
snellius Contributing Member • Posts: 834
Re: The Naked Truth About Bokeh

Marianne Oelund wrote:

Unique 58G wide-aperture background blur circle with diffuse perimeter and bright central peak

1b. Blur circle size should reduce as the lens is stopped down, for example the f/2 blur circle should be only 70% of the diameter of the f/1.4 blur circle. Look at the foreground blur circles for the 58G at f/1.4 and f/2: Why are they almost the same size?

This is because there is not focused properly on the smaller apertures. It is better to use AF fine tuning or live view AF with closed aperture. The result would look like this.

It is certainly a good initiative to point out the users of such lenses. I'm doing this for ten years , but most do not believe the problems with focus shift. As they pay a lot for their stuff and the name Nikon is on it, they only want to hear it is pervect. It is often difficult to understand that more and more compromises resulting from the high-end equipment.

Vladi Stoimenov
Vladi Stoimenov Regular Member • Posts: 446
Re: The Naked Truth About Bokeh

Awesome work!

Thank you Marianne.
Vladi

 Vladi Stoimenov's gear list:Vladi Stoimenov's gear list
Panasonic Lumix DMC-LX2 Nikon D610 Nikon D90 Nikon AF-S Nikkor 70-200mm f/4G ED VR Nikon AF-S Micro-Nikkor 105mm f/2.8G IF-ED VR +6 more
MisterHairy Senior Member • Posts: 2,202
Re: The Naked Truth About Bokeh

I believe that Marianne mostly works with cameras which do not correct for focus shift (this is true for most Nikons to date) so her results are correct for her (and many others') usage.

Yours are representative of only the most very recent generation of Nikon cameras.

-- hide signature --

Really beautiful photograph!

 MisterHairy's gear list:MisterHairy's gear list
Nikon D700 Nikon D810 Nikon Df Nikon PC-E Nikkor 24mm f/3.5D ED Nikon AF-S Nikkor 70-200mm f/2.8G ED VR II +19 more
Rudolf Morf Forum Member • Posts: 67
Re: The Naked Truth About Bokeh
2

Marianne Oelund wrote:

There is quite a bit of information in these figures, and much more could be said, but I will leave the discovery to you. Hopefully you will find some interesting surprises.

There is indeed very much information in those figures. Thank you very much for this excellent work and for sharing it with us.

I have always wondered what it is that makes the 58g so special. I recommend looking at the old post https://www.dpreview.com/forums/post/53524950 in which results from lenstip for the 58g, the 50mm Otus und the 50mm Sigma Art have been compared. What struck me was the fact that the numbers is the Lensalign pattern remain quite readable in both for- and background even at f/1.4 fot the 58g, but very quickly become completely distorted or disfigured for the two other lenses which have been so highly praised.

Your tests provide new insight !

-- hide signature --

Rudolf

just Tony Senior Member • Posts: 1,787
Re: Complex Lenses

Marianne Oelund wrote:

But what I am most interested in exploring here, goes beyond basic SA, to the particulars of complex lens designs. The f/1.4 blur circles for the 58G and 85/1.4G exhibit many detail differences beyond what one expects from single-element SA. This is what question 2 alludes to.

The three-layered point spread function of the 58/1.4 really caught my eye. That outer shell arises from the annulus between f/2 and f/1.4.

There is a very slightly similar behavior in the 105/2.5 Ai/AiS and 85/1.8K manual lenses. At the wide open position there is a broad haze around focused points, which disappears with the first clicks of their aperture rings. Let's call that the "old school" approach.

I would guess that the two aspheric elements in the construction of the 58/1.4G were used to precisely tailor this non-traditional SA characteristic. This is a more deeply engineered bokeh.

By the way I'm guessing that both of your example lenses are decentered.

noirdesir Forum Pro • Posts: 13,381
Re: The Naked Truth About Bokeh

MisterHairy wrote:

I believe that Marianne mostly works with cameras which do not correct for focus shift (this is true for most Nikons to date) so her results are correct for her (and many others') usage.

Yours are representative of only the most very recent generation of Nikon cameras.

Do you know which cameras this includes?

OP Marianne Oelund Veteran Member • Posts: 7,062
Re: Complex Lenses

just Tony wrote:

The three-layered point spread function of the 58/1.4 really caught my eye. That outer shell arises from the annulus between f/2 and f/1.4.

There is a very slightly similar behavior in the 105/2.5 Ai/AiS and 85/1.8K manual lenses. At the wide open position there is a broad haze around focused points, which disappears with the first clicks of their aperture rings. Let's call that the "old school" approach.

I would guess that the two aspheric elements in the construction of the 58/1.4G were used to precisely tailor this non-traditional SA characteristic. This is a more deeply engineered bokeh.

By the way I'm guessing that both of your example lenses are decentered.

Thank you for your commentary, which is a nice addition to this thread.

Yes, there's slight evidence of asymmetry with these lenses, but it's not enough to make me concerned.  Wait 'til you see what's coming in the 50mm group . . . 

-- hide signature --

Source credit: Prov 2:6
- Marianne

OP Marianne Oelund Veteran Member • Posts: 7,062
Declaration of Neutrality: Macro Example
9

Macro lenses are unique in a number of ways, besides their close-focusing capability.  Typically, they have neutral SA correction, low distortion and flat fields.

Here is a set of blur circles for the Micro 100mm f/2.8D.  Rather than the under-corrected SA common in most lenses, we see that it actually has slightly over-corrected SA, producing very soft circles for the foreground, and just a little brightness at the edge of the background circles which gradually fades with increasing distance:

Very nice foreground bokeh softness, with reasonably neutral background bokeh.

In-focus sharpness is very good, DOF is shallow with fairly fast transition to out-of-focus.  There is a little evidence of asymmetry (decentering), but I'm finding that this amount is not uncommon in moderately-expensive lenses and the effect in photographs is not usually noticeable.

-- hide signature --

Source credit: Prov 2:6
- Marianne

Keyboard shortcuts:
FForum MMy threads