Test: Tamron 28-75 2.8 @28mm vs. Super Takumar 28mm 3.5

Testing lenses isn't easy. I know from experience and some of you may remember a couple of lens reviews I have made and posted here. This thread is interesting from several aspects. For me colors have always been a big challenge (and no, I am not colorblind according to any test available). I have earlier tried to ask people here what they mean when they say things like "I prefer the colors from the "lens x". I still don't really grasp it.

A couple of weeks ago I was close to the final decision on what 50mm I wanted to use. To narrow down my choice I arranged a trial setup. It looked like this:



Above: One of the "overview" pictures from the failing test

I shot the same scene using five different 50mm lenses, I took great care of focusing and exposure and had to do only minor EV-adjustments when converting the raw files.

The gallery is here:
http://ImageEvent.com/jonas_b/fmforumsmonthly/5x50mmfailed
and contains two picture per lens, one overview and one 100% crop.

I would appreciate if any of the color aficionados reading this take some time and tell me what, if, color differences that can be seen.

If you are not a color aficionado but just interested in general... well, I would like to get suggestions for the next round. What can a good setup look like to give a reasonably good hint about bokeh. I have no problems with judging resolution and CA and can even run an IMA-test if I find the time and motivation. My earlier bokeh test have all been done outdoor in uncontrolled light.

Ideas? (Yes, I'm looking for the "perfect" 50mm lens. That may say something about me and your diagnosis are welcome together with your ideas.)

--
Jonas

All pictures are taken with a Canon 5D, the distance is 0.93m between sensor and the doll's face. The distance to the center part of the background is about 2.2m. It's an ordinary messy background, not very unlike what we get in real life when all the relatives are gathered in our living rooms.

There are 6 pair of pictures. The two first pairs are taken with the same lens with the second pair slightly OOF (the focus plane were probably around 0.92m)

All pictures are treatened with regards to white balance and exposure and they are sharpened exactly the same (just a little).

And the issue was: Ideas about a better setup? Opinions on colors?

--
Jonas

All pictures are taken at f/2. That is wide open for one of the lenses (easy to see). The other lenses are stopped down between 1 and 1.5 stops.

5 different 50mm lenses, 3 different brands.

But the main issue is not to guess lenses but to explain color differences, and, even more appreiated, suggest a better setup to see bokeh differences.

Hi Jonas!

Jonas B said:

But the main issue is not to guess lenses but to explain color
differences,

Click to expand...

Which differences?

Jonas B said:

and, even more appreiated, suggest a better setup to
see bokeh differences.

Click to expand...

I don't have one either. I think OOF highlights like small light bulbs can give good clues about the effective roundness of the aperture. Yet, one of the toughest tasks for bokeh smothering are lines and how they are rendered. Also, Bokeh changes between foreground and background and your test would need to reveal that, too. And finally (of what I can think of), Chris Gunn talks about foreground and background Bokeh-CA (or similar, sorry Chris), so you'd need a subject where you can see these CA lines properly.
Interesting task - for someone other than me

Cheers
Jens

--
'I only trust those photos I have faked myself.' (Me, 2007)
http://www.jensroesner.de/
--=! Condemning proprietary batteries since 1976 !=--

JensR said:

JensR said:

But the main issue is not to guess lenses but to explain color
differences,

Click to expand...

Which differences?

Click to expand...

Yes, that is one of the questions that puzzled me in this thread...

JensR said:

JensR said:

and, even more appreiated, suggest a better setup to
see bokeh differences.

Click to expand...

I don't have one either. I think OOF highlights like small light
bulbs can give good clues about the effective roundness of the
aperture. Yet, one of the toughest tasks for bokeh smothering are
lines and how they are rendered. Also, Bokeh changes between
foreground and background and your test would need to reveal that,
too. And finally (of what I can think of), Chris Gunn talks about
foreground and background Bokeh-CA (or similar, sorry Chris), so
you'd need a subject where you can see these CA lines properly.
Interesting task - for someone other than me

Click to expand...

I have an idea where measuring-tape and Christmas tree decoration is involved. I may also add high contrast details around the focus plane to show the axial CA and flare/ghosting. For CA I have used another setup but it can as well get included here.

Jens, to me this little test setup where only five lenses were involved a dissapointment: It showed no, or very small, differences between the lenses. How fun was that when I was chasing the perfect normal lens?

It's an interesting task to me: Do I have to get and convert a Summilux-R 50/1.4, or can I just smack on an old Takumar? And when does it matter? Is the bokeh from the Zeiss C/Y Planar 50/1.4 that bad? Or is it a DA* 55/1.4 that is the only sensible way?

Salut,

--
Jonas

Jonas B said:

Yes, that is one of the questions that puzzled me in this thread...

Click to expand...

Well, in all fairness, when I compare my superzoom to a good prime, I notice different colours. And if I compared various old primes of differing quality, I also noticed differences.
So, maybe your "mistake" was to compare all that fine glass!

Jonas B said:

Jens, to me this little test setup where only five lenses were
involved a dissapointment: It showed no, or very small, differences
between the lenses. How fun was that when I
was chasing the perfect normal lens?

Click to expand...

LOL. Terrible fate

Jonas B said:

It's an interesting task to me: Do I have to get and convert a
Summilux-R 50/1.4, or can I just smack on an old Takumar? And when
does it matter? Is the bokeh from the Zeiss C/Y Planar 50/1.4 that
bad? Or is it a DA* 55/1.4 that is the only sensible way?

Click to expand...

You should get a DS and an Industar 50-2, use it "wide" open and be done with it

Cheers
Jens

--
'I only trust those photos I have faked myself.' (Me, 2007)
http://www.jensroesner.de/
--=! Condemning proprietary batteries since 1976 !=--

JensR said:

-snip-
You should get a DS and an Industar 50-2, use it "wide" open and be
done with it

Click to expand...

Amazing. I have absolutely no reply to that.

Btw: Thank you for reading, looking, posting Jens. With 30 album hits and 63 downloaded pictures and 1 reply I think I have invented a way to measure the level of boredom and meaninglessness.



--
Jonas

I notice this in my tamron a lot, both my 28-75 and 90mm Macro have a noticable yellow cast.

I compared the 90mm to Sigma's 105mm Macro and the colour cast is very pronounced it almost made me regreting selling the Sigma.

Jonas B said:

JensR said:

-snip-
You should get a DS and an Industar 50-2, use it "wide" open and be
done with it

Click to expand...

Amazing. I have absolutely no reply to that.

Btw: Thank you for reading, looking, posting Jens. With 30 album
hits and 63 downloaded pictures and 1 reply I think I have invented
a way to measure the level of boredom and meaninglessness.

Click to expand...

Does it allow you to download pictures? I get a 'right-click disabled' message.

Jonas B said:

--
Jonas

Click to expand...

carpents said:

Does it allow you to download pictures? I get a 'right-click
disabled' message.

Click to expand...

My error. The gallery just got the default settings. Now you can right-click.

J

Picture 108: Canon EF50/1.2L @f2 (very little flare/ghosting, great bokeh)

Picture 109: The same lens but misfocused (FF approx 10-15mm), compare the background...

Picture 134: Canon EF50/1.4 USM at f/2 (standard thingy)

Picture 157: Zuiko 50/1.2 at f/2 (smalles f/1.2 lens I have seen, pleasent rendering)

Picture 185: Zuiko 50/1.4 at f/2 (late multicoated model, known to be "soft")

Picture 196: Leica Summicron-R 50/2 (good reference lens)

--
Jonas

Jonas B said:

Picture 108: Canon EF50/1.2L @f2 (very little flare/ghosting, great
bokeh)
Picture 109: The same lens but misfocused (FF approx 10-15mm),
compare the background...

Click to expand...

What's with the background? Looks as I would expect? No?

--
'I only trust those photos I have faked myself.' (Me, 2007)
http://www.jensroesner.de/
--=! Condemning proprietary batteries since 1976 !=--

As far as quality of the bokeh in these shots, I would rank them:
108
157 (although only because of 134's aperture blades)
134
185 (strange lop-sided bright rings around OOF highlights?)
196 (pretty noticeable green CA in the OOF highlights)

Nothing of note in the colors. The sharpness is fine for all of them.

As for your ultimate lens test subject, why does it have to be one? I thought your testing regime of a combination of tests is really the only way to do it. You can shoot USAF test charts for critical sharpness comparisons (under controlled light) and you can shoot a combination of 'real world' subjects like this one or bicycles or strange kissing-statues to help fully describe each lens. No one test scene is going to fully quantify a lens, nor is any real scene going to tax every single quantifiable behavior of it.

I know this doesn't help you compare a scene last year from a similar scene this year, but is that really so important? I guess I like to treat lenses like wine - I enjoy comparing two side-by-side to help me learn about each, but I don't really care if I can't say it was 'better' than one from last year!

Jonas B said:

Picture 108: Canon EF50/1.2L @f2 (very little flare/ghosting, great
bokeh)
Picture 109: The same lens but misfocused (FF approx 10-15mm),
compare the background...

Picture 134: Canon EF50/1.4 USM at f/2 (standard thingy)
Picture 157: Zuiko 50/1.2 at f/2 (smalles f/1.2 lens I have seen,
pleasent rendering)

Picture 185: Zuiko 50/1.4 at f/2 (late multicoated model, known to
be "soft")

Click to expand...

Based on only one picture, I would have guessed some sort of Russian lens.

Jonas B said:

Picture 196: Leica Summicron-R 50/2 (good reference lens)

Click to expand...

Another surprise. I would have guessed a Pentax.

Jonas B said:

--
Jonas

Click to expand...

Katsura said:

I notice this in my tamron a lot, both my 28-75 and 90mm Macro have
a noticable yellow cast.

I compared the 90mm to Sigma's 105mm Macro and the colour cast is
very pronounced it almost made me regreting selling the Sigma.

Click to expand...

Aren't you using a digital SLR? I could see how different lenses would render somewhat differently using different films, but a DSLR is correcting color balance based on what it senses coming through the lens and would deal with that. Same as trying to use color correcting filters on a digital camera...doesn't work.

JensR said:

JensR said:

Picture 108: Canon EF50/1.2L @f2 (very little flare/ghosting, great
bokeh)
Picture 109: The same lens but misfocused (FF approx 10-15mm),
compare the background...

Click to expand...

What's with the background? Looks as I would expect? No?

Click to expand...

Yes, it looks as one can expect. I included the misfocused picture 109 as I thought it was a nice illustration to how impossible it is to measure bokeh: The scene is the same, the distances are the same and even the lens is the same. The difference is that the focus plane is moved about 15mm closer to the camera. The result is that the OOF highlight in the background (a bit over 2m away from the camera) now are "bigger" (CoC is increased). A tiny minute detail changes the impact of the picture.

Ain't this pixel peeping taken to a nice level?

Now I'll just forget about this failed experiment, have a small glass of Caol Ila and go to bed.

Cheers,

--
Jonas

carpents said:

Jonas B said:

Picture 108: Canon EF50/1.2L @f2 (very little flare/ghosting, great
bokeh)
Picture 109: The same lens but misfocused (FF approx 10-15mm),
compare the background...

Picture 134: Canon EF50/1.4 USM at f/2 (standard thingy)
Picture 157: Zuiko 50/1.2 at f/2 (smalles f/1.2 lens I have seen,
pleasent rendering)

Picture 185: Zuiko 50/1.4 at f/2 (late multicoated model, known to
be "soft")

Click to expand...

Based on only one picture, I would have guessed some sort of
Russian lens.

Click to expand...

I would have guessed a Pentax lens thinking of how often the aperture blades closes down in an uneven way with some of them.
No, really, I could have guessed just about anything.

carpents said:

carpents said:

Picture 196: Leica Summicron-R 50/2 (good reference lens)

Click to expand...

Another surprise. I would have guessed a Pentax.

Click to expand...

Thinking of the green bright rings? Yes.

The Summicron-R 50/2 is a very nice lens. It does nearly everything just right. Thanks to the bright rings here (wide open, an unfair comparision) it at least has a bit of character...

--
Jonas

carpents said:

As far as quality of the bokeh in these shots, I would rank them:
108
157 (although only because of 134's aperture blades)
134
185 (strange lop-sided bright rings around OOF highlights?)
196 (pretty noticeable green CA in the OOF highlights)

Nothing of note in the colors. The sharpness is fine for all of them.

Click to expand...

To me there were surprisingly small differences. Yet, in real life when actually using the lenses for something worthwhile I can see differences. The setup was too simple.

carpents said:

As for your ultimate lens test subject, why does it have to be one?
I thought your testing regime of a combination of tests is really
the only way to do it. You can shoot USAF test charts for critical
sharpness comparisons (under controlled light) and you can shoot a
combination of 'real world' subjects like this one or bicycles or
strange kissing-statues to help fully describe each lens. No one
test scene is going to fully quantify a lens, nor is any real scene
going to tax every single quantifiable behavior of it.

I know this doesn't help you compare a scene last year from a
similar scene this year, but is that really so important? I guess
I like to treat lenses like wine - I enjoy comparing two
side-by-side to help me learn about each, but I don't really care
if I can't say it was 'better' than one from last year!

Click to expand...

Ftm I think it is all true, and put in a good way.

I was trying to find the "perfect" fast 50 (other contenders in this test where the Summilux-R 50/1.4, SMC Takumar 50/1.4, Zeiss C/Y 50/1.4 and an SMCP A50/1.7) as I really would like to have just one of them, forget all about the testing, learn to know it real well and sell of the rest of them. It would have became my main working lens, just as the FA35/2 has been for me. Now it doesn't matter.

Btw, focus shift or not, the EF50/1.2L was my most used lens with the 5D. Clear winner wide open, great bokeh, least flare/ghosting and sharp enough. The Summilux 50/1.4 is also great with a nice liquid rendering and at the same time bitingly sharp when stopped down to 2.8.

regards,

--
Jonas