Zeiss ZM Biogon-C 21mm f/4.5 (current)

[Guest Walt]

Advertisement (gone after registration)

I suppose this is as good as any thread to ask this question.

 

I hear people talking mostly about contrast and sharpness in lenses, but I don't hear a term for the quality of the grayscale, which is most important to me. Though I recognize the higher contrast of the Zeiss lenses, they also have a long, smooth, subtle, elegant midtone range, that is distinctive and usually, in my opinion, better than that of many modern Leica lenses. The 28/2.0, which is the lens I use most on the M8, has much of this Zeiss grayscale quality. The 28/2.8 ASPH has very little of it, and I don't like to use this lens. I find the grayscale of this lens ugly and "lumpy." The Leica 35/2.0 ASPH, which is a contrasty lens, is somewhere between the two 28's on the quality of grayscale issue. The CV 35/2.5, which is a lower contrast lens, has a beautiful grayscale. So, I don't find that the quality of the grayscale is much related to the contrast of the lens.

 

What is it I am seeing here? Is this grayscale issue not discussed or is it being discussed in terms I am not recognizing?

 

Walt

[lct]

My only problems with grayscale are when the lens is very contrasty. Otherwise shooting raw and using iCorrect give me the feeling that i can get the results i need rather easily.

[hankg]

Though I recognize the higher contrast of the Zeiss lenses, they also have a long, smooth, subtle, elegant midtone range, that is distinctive and usually, in my opinion, better than that of many modern Leica lenses. The 28/2.0, which is the lens I use most on the M8, has much of this Zeiss grayscale quality.

 

The transitions on the Zeiss lenses are very smooth. The sharp to out of focus transitions are reminiscent of some old classic lenses. This way of rendering must impact the step to step transitions in the grey scale. I have no idea how optical design affects this behavior. I'd love to see an illustration of 2 lenses that exhibit the 2 sides of this behavior like the 28/2 and 28/2.8 have any samples that would show the difference on the web Walt?

 

The combination of moderate (or high but not to high) contrast and smooth tonal transitions makes for a RAW file with tremendous flexibility. You can locally hike middle tone contrast, plug up some shadows and open others, exaggerate detail or lose it selectively and the image holds together with out developing abrupt ugly transitions and artifacts. To much contrast or abrupt transitions in the grayscale up front is like to much sharpening or smoothing in camera it really hobbles your ability to shape the image to your personal taste and vision.

[Guest Walt]

To elaborate a small bit, I am talking about a smooth, long grayscale, but one that is not "dull," heavy, or leaden. Perhaps it is smooth gradation with "local contrast." With a flat lens I can get some of this with large-radius USM, but some lenses are distinctively good at this. Zeiss lenses are typically good at this. Modern Leica lenses, to generalize, typically look lumpier or stepped rather than smooth. I am talking about the appearance of the midtones.

 

Hank, since I never shoot the same scene with two different lenses, I'm not sure I can illustrate this. I'll think about it. But over several thousand images, the differences between lenses becomes obvious and I can usually recognize the 28/2.0 in a thumbnail. I can usually also recognize the 28/2.8 also, unfortunately.

 

Walt

[carstenw]

Walt, although you have not used the 28/2A and 28/2.8A side-by-side, can you post images which are characteristic of each?

 

Have you tried the pre-Asph 28/2.8?

[lct]

28/2 & 28/2.8 asph, 100% crops at f/2.8 and f/5.6 (R-D1, 200 iso).

The 28/2.8 asph is a bit more contrasty with a slightly harsher bokeh.

Same in BW but i did not compare side by side so far sorry.

 

28/2 & 28/2.8 asph, 100% crops at f/2.8:

http://tinyurl.com/rv7w/EPSN2739_crop02web.jpg

http://tinyurl.com/rv7w/EPSN2742_crop02web.jpg

 

28/2 & 28/2.8 asph, 100% crops at f/2.8:

http://tinyurl.com/rv7w/EPSN2747_crop01web.jpg

http://tinyurl.com/rv7w/EPSN2745_crop01web.jpg

 

28/2 & 28/2.8 asph, 100% crops at f/2.8:

http://tinyurl.com/rv7w/EPSN2742_crop01web.jpg

http://tinyurl.com/rv7w/EPSN2743_crop01web.jpg

 

28/2 & 28/2.8 asph, 100% crops at f/5.6:

http://tinyurl.com/rv7w/EPSN2740_crop02web.jpg

http://tinyurl.com/rv7w/EPSN2742_crop02web.jpg

 

28/2 & 28/2.8 asph, 100% crops at f/5.6:

http://tinyurl.com/rv7w/EPSN2748_crop01web.jpg

http://tinyurl.com/rv7w/EPSN2746_crop01web.jpg

 

28/2 & 28/2.8 asph, 100% crops at f/5.6:

http://tinyurl.com/rv7w/EPSN2740_crop01web.jpg

http://tinyurl.com/rv7w/EPSN2743_crop01web.jpg

[Guest Walt]

Advertisement (gone after registration)

Walt, although you have not used the 28/2A and 28/2.8A side-by-side, can you post images which are characteristic of each?

 

Have you tried the pre-Asph 28/2.8?

Carsten and LCT-

 

I'll do this when I have time. I can't tell much from LCT's post because of the color, the 100% crops, the subject matter, the web image, etc.

 

Yes, I've used the pre-aspherical 28/2.8 (the 46mm filter thread, the last before the current lenses) and I think it is excellent, perhaps the very best of the 28s. I am not using it only because for the same weight and size I can get another stop. I couldn't recommend this lens too highly.

 

Walt

[Dan States]

To elaborate a small bit, I am talking about a smooth, long grayscale, but one that is not "dull," heavy, or leaden. Perhaps it is smooth gradation with "local contrast." With a flat lens I can get some of this with large-radius USM, but some lenses are distinctively good at this. Zeiss lenses are typically good at this. Modern Leica lenses, to generalize, typically look lumpier or stepped rather than smooth. I am talking about the appearance of the midtones.

 

Hank, since I never shoot the same scene with two different lenses, I'm not sure I can illustrate this. I'll think about it. But over several thousand images, the differences between lenses becomes obvious and I can usually recognize the 28/2.0 in a thumbnail. I can usually also recognize the 28/2.8 also, unfortunately.

 

Walt

 

Walt, I always felt that this quality was because of their superior control of flare. The Zeiss lenses render highlights with perfect clarity and shadows without the ever so slight haze of lesser lenses. It's not about "sharpness" because these same qualitys are evident in their vario lenses as well.

 

You can get really subtle gradations in the highlights that made images easy to print from black and white film. The same translates to digital in my experience.

 

As for recognizing the 28 2.0...that's EASY....look for perfect sharpness, high color saturation and strong vignetting (on film). Minus the light falloff it's just about the perfect lens (too bad about the loony price).

 

Best wishes

Dan

[scott kirkpatrick]

Sean Reid's recent reviews of the M-mount CV21/4 P suggested that the new model, which is supposed to be optically identical to the older LTM mounted CV21/4s, gives better contrast and resolution out to the corners of the field. This provoked some testing to see if the LTM CV21's had a lot of sample variation, perhaps leading to what sounds like de-centering. A good sample of the CV21/4 LTM was subsequently found, but the question remains if the new model is also a risky proposition.

 

I just received a CV21/4 P from Cameraquest, and it is sharp across the field according to a quick acceptance test, conducted on tripod, with focus-bracketing, shooting my front door in fading sunlight. I compare it with a good recent copy of the Elmarit 24/2.8 asph. Both are shot at f/4.0, and if I was not aligned perfectly, the effects would show up for both lenses, as the camera was not moved between sets.

 

24 Elmarit asph at center, 100% crop:

and the same area shot with the CV214P:

 

Now the worst corner for both lenses, the lower left -- first the 24 Elmarit

and then the CV21P:

 

The 24 is a little crisper both on center and at the corner, but neither has anything to be ashamed of, I think. Look at the RR articles on the 21mm rangefinder lenses to see what poor performance looks like. So the statistics for the CV21/4 P are now two good out of two samples. Maybe the M mount is a little more stable than the LTM.

 

scott

[J_Brittenson]

If the exposure is the same, wouldn't a lower contrast lens only cast a veil over the shadows? How would it actually draw out MORE shadow detail?

It's like preflashing film to sensitize it. It overcomes the toe. Digital is the same, it has a noise floor and preflashing (or adding a little extra fill light in the shadows) will raise the dark end of the tonal scale above the noise floor. So when the camera clips the dark end (by subtracting an offset in the ADC) to eliminate noise this portion of the scale that would otherwise not register is left above the clip point.

 

Veiling flare is similar, but tends not to be evenly distributed and tends to be excessive meaning you don't get clean blacks at all. And that's clearly not a good thing.

 

Here's a shot from the 35/2 (II), clear bright sun at about 3pm.

 

 

Contrast is very nicely controlled here. Even the hot part on the nose isn't totally offensive.

 

Low contrast doesn't mean low resolution or lack of clarity, here's a crop processed in PS CS2 with no sharpening other than the DNG processing:

 

 

Curiously, this lens doesn't need the 1/3 to 1/2 EV exposure adjustment I use normally since it seems to produce a slightly darker image overall.

 

If I had shot this with a ZM lens it would have been line art. Yech!

[Dan States]

I'm sorry but preflashing tones down highlights by foging what would otherwise be pure white photo paper. It does not create detail in the shadows and neither can a lower contrast lens on a camera unless you EXPOSE more. Taking this theory to an extreme would mean that the Summar is the ideal 50mm lens! It certainly does condense the response curve!

[hankg]

I'm sorry but preflashing tones down highlights by foging what would otherwise be pure white photo paper. It does not create detail in the shadows and neither can a lower contrast lens on a camera unless you EXPOSE more. Taking this theory to an extreme would mean that the Summar is the ideal 50mm lens! It certainly does condense the response curve!

 

You want a balance. If you are shooting test charts the more contrast the better the result. But if you had to shoot a real 3D scene say where the model was wearing a dress with black brocade and white lace and the client wanted very dramatic lighting but wanted to see all the fine detail in the black and white fabrics? In the studio you can match the lighting ratios to the dynamic range of your sensor. You will get the full greyscale from highlight to shadow with no clipping if you set up the lighting contrast to not more then say 6 stops max white to black. You could then selectively punch up the contrast as much as you wanted in post.

 

But what if you had to shoot outside in full sun -lots of deep shadows and bright sunlit highlights with no flash, diffusers or reflectors (which is how most documentary and street photographers shoot) you must find a way to compress those 10+ stops down to a range that can be recorded by the sensor. You can't control the lighting and you can't record more range then the sensor is capable of - so that leaves the lens -you want enough contrast but not to much. The ideal solution is a sensor with more dynamic range but until that happens you work with what you have. In the conditions I describe the Zeiss will deliver results that are unusable and even the Leica 28/2 will be at the edge. You might have to resort to an old Canon 28 for the best result trading some detail in the midtones to hold detail at the extremes.

 

This is why I think the balance in the 28/2 is about perfect -it is by no means a 'low contrast' lens. It is in many respects most like a Zeiss ZM with not quite so much contrast. My ideal lens would be a lens with the price and image signature of the Zeiss ZM's (resolution across the frame, OOF drawing, tonal transitions) but with the contrast of a preASPH Leica.

[sandymc]

It's like preflashing film to sensitize it. It overcomes the toe. Digital is the same, it has a noise floor

 

No, not so. Preflashing works, in some situations anyway, for film because of the non-linear toe to the response curve. That doesn't exist for digital. Noise will be the same - same number of stray electrons. Preflash might hide that, but only at cost of destroying detail.

 

Sandy

[cme4brain]

Sean - Your excellent review of the 21s definitely helped me to take the decision, for which thank you.

 

Stephen

 

My CV 21mm skopar seems to be a good copy, with excellent corner sharpness.

[J_Brittenson]

BTW, contrast affects the perception of "sharpness", not resolution per se. The detail is still there in a lower contrast file, its just less obvious until and unless local contrast is increased. Sometimes, its preferable not to.

The great thing about digital here, especially cameras and backs without an AA filter, is the ability to capture low-contrast detail. Compared to film, the M8 cares vastly less about local contrast! With film there's a very real tradeoff in fine texture, but with digital at the current 10MP/6.8u pitch it's a non-issue IMO.

 

I do think the higher contrast modern lenses are preferable when using controlled lighting. You get much closer to what you see (to my eye at least!) and it's easy to dial back contrast, throw in a reflector (or flag, if more is needed) or tweak a specular to perfection. Or when shooting in extremely flat light outdoors, such as twilight.

 

But like you say, it's a matter of taste. I mainly posted to object to the notion that a low-contrast lens is the same as a less well corrected high-contrast lens. It's clearly not. It may still be as well or even better corrected for some aberrations (like chromatic) while less good in other areas (like oblique-ray spherical).

[sbelyaev]

It is a nice lens.

Small, light and optically, probably, as good as elmarit-m 21/2.8 asph. I compared both lenses. For practical purposes, I found no differences in terms of picture quality.

[sbelyaev]

I was looking at some pictures and noticed that Biogon is much worse that elmarit in terms of light fall-off (on M8). Elmarit gives a very even illumination while biogon gives darker not just corners, but noticable light fall-off starts around the middle of the image. I haven't noticed it before because I didn't have pictures that include the sky. It was impossible to correct that light fall-off in Lightroom.

Well, I guess if one wants an ultimate quality he/she should pay accordingly. There is no way around it.

[lct]

My CV 21mm skopar seems to be a good copy, with excellent corner sharpness.

Thanks to your M8 firmware and/or microlenses i guess.

On the R-D1, the CV 21/4P vignettes a lot more than the Elmarit 21/2.8 asph.

[sean_reid]

Microlenses

 

Cheers,

 

Sean

[Advertisement]

Hallo,

Look here: Leica M

[sean_reid]

I was looking at some pictures and noticed that Biogon is much worse that elmarit in terms of light fall-off (on M8). Elmarit gives a very even illumination while biogon gives darker not just corners, but noticable light fall-off starts around the middle of the image. I haven't noticed it before because I didn't have pictures that include the sky. It was impossible to correct that light fall-off in Lightroom.

Well, I guess if one wants an ultimate quality he/she should pay accordingly. There is no way around it.

 

I imagine your Zeiss is uncoded? If so, you are comparing apples to oranges. Compare them both coded or both with the lens detection turned off.

 

"Well, I guess if one wants an ultimate quality he/she should pay accordingly. There is no way around it."

 

That myth is still making some companies lots of money. Sometimes quality and cost are related and sometime they are not. And "quality" can be defined a lot of different ways.

 

Cheers,

 

Sean