Standard lens comparison

[blockend]

I've been looking for an objective test on 50mm standard lenses in the f1.8 to f2 range, the kind that came with SLR bodies from the 1960s to 1980s. There is a great deal of anecdote out there, and Googling mostly takes me to subjective comparisons between lens a and lens b, which are neither scientific nor comprehensive. Given that these are the most common lenses around, I'd be interested in any serious optical tests between manufacturer's standard lenses.

 

[Dan Fromm]

Why? I ask because when Modern Photography that used to be published mass tests of normal lenses for 35 mm SLRs differences between best and worst rarely amounted to much.

 

[blockend]

Why? I ask because when Modern Photography that used to be published mass tests of normal lenses for 35 mm SLRs differences between best and worst rarely amounted to much.

Do you mean there was not much difference between them in optical performance, or that the differences were irrelevant to practical photography? One of the reasons I wanted to pursue such an analysis is because there is a big difference in price between standard lenses, and once you take away mechanical quality, I wonder whether that cost difference is born out in optical performance. Is there, for example, a real and perceivable difference between a Leitz or a Contax standard lens, and a Chinon or Zenit lens at a fraction of the price?

 

[darkosaric]

It is hard to find 50mm prime lens that is manufactured after 1960 and that is not very very good optically. All 50mm lens after 1960 that I have tried (canon, nikon, Leitz, FSU...) were great or superb.

 

[darkosaric]

Is there, for example, a real and perceivable difference between a Leitz or a Contax standard lens, and a Chinon or Zenit lens at a fraction of the price?

Difference you will see if you shoot handheld in dark, wide open, 1/30s. Difference between lenses that you are talking about is more about the signature of the lens, and not so much about sharpness. And those things like signature, character... are not so easy to measure as sharpness - that is why manufacturers like sharpness - it is easy to measure, but it is not so important in real life (at least not to me).

 

[Nodda Duma]

Dan is correct, and is in line with what I'd expect as a lens designer. Fundamentally, all modern fast normal focus lenses are Double Gauss designs with very similar levels of correction. For the same 1st order optical parameters the design form optimizes out pretty much the same. The differences come down to meeting different size, cost, and weight requirements, or taking advantage of new technology such as computer aided design or aspheric manufacturing.

Sort the lenses by focal length and maximum f/#, then by decade in which the lens was designed (not manufactured). Lenses that fall in the same category will be very similar in performance.

From there, switch to reviewing subjective preferences from other users. Their feedback will tell you something about the coatings and the ergonomics.

You may find the bigger optical performance differences come down to how tightly the manufacture held fab and assembly tolerances. The more expensive the lens (due to tighter controls), the better chance your particular lens will have slightly better performance.

 

[blockend]

Thanks for the feedback. As the owner of a number of 50mm lenses from various manufacturers I have my favourites. This is sometimes because they appear very sharp, and in other cases because they render the image subjectively in a way that matches how I remember the scene. Quantifying that appearance is something that is rarely described in a way that is useful to photographers, for example people talk about "pop" as though it was an optical phenomenon with specific characteristics.

Having looked at all all kinds of lenses and the images they make over many years, I think the preference comes down to high contrast or low contrast varieties, which is associated (though not exclusively) with the evolution of optical coatings and computer aided design. Highly corrected, multi-element lenses do not always translate their advantages into a 2-dimensional viewing plane like a computer screen, and even less into a print, yet their rendering is prized by reviewers and attracts a high value, even though alternative renderings may be much more pleasing aesthetically.

 

[Nodda Duma]

the one thing I've never seen is spectral throughput measurements made on a spectral photometer. There could be a correlation there with "pop".

Honestly, I don't think there's a "smoking gun" measurement to correlate, but rather a combination of measurements and analysis. There will be missing puzzle pieces in metrics which require modeling with the original design file (stray light, narcissus, ghosting, etc)...but subjective opinion will hint is better or less controlled.

 

[Dan Fromm]

Do you mean there was not much difference between them in optical performance, or that the differences were irrelevant to practical photography?

Yes and yes.

Jason, I've seen people fight to the death over differences in the way lenses render color. I've always thought that most modern lenses' transmission by wavelength curves as measured with a spectrophotometer are nearly flat in the visible -- where they start rolling off badly at the edges of the visible seems to vary -- and that tests of color rendition with color reversal film are so vulnerable to variations in shutter speed (especially with leaf shutters), illumination and film processing as to be useless. If I'm mistaken please correct me.

 

[Nodda Duma]

- where they start rolling off badly at the edges of the visible seems to vary --

I don't think you're mistaken and this quote above is kind of what I was thinking too. I also think perhaps the absolute (not normalized) throughput would tell me something as well. As well there might be a percentage point variance across the spectral region of "full throughput". But honestly it's mere speculation on my end because I've never seen measurements made. Perhaps I'll bring a couple lenses (like one of my Nikkor 50 / 1.4 and Zuiko 50 / 1.4) into work and measure them on a lunch break. As you state that's only part of the story. Film's color rendition plays as big a part.

From the design standpoint the impact of all this on the final impression of quality gets pretty nebulous pretty fast, and I have to start relying on my experience and gut feeling to interpret measurements made if I got a sudden desire to characterize lens performance. Personally I'm more of a "proof is in the pudding" type of guy, so if the lens makes pictures I like then I'm going to use it.

 

[RalphLambrecht]

Difference you will see if you shoot handheld in dark, wide open, 1/30s. Difference between lenses that you are talking about is more about the signature of the lens, and not so much about sharpness. And those things like signature, character... are not so easy to measure as sharpness - that is why manufacturers like sharpness - it is easy to measure, but it is not so important in real life (at least not to me).

sharpness is not easy to measure at all;resolution is

 

[Gerald C Koch]

Difference you will see if you shoot handheld in dark, wide open, 1/30s. Difference between lenses that you are talking about is more about the signature of the lens, and not so much about sharpness. And those things like signature, character... are not so easy to measure as sharpness - that is why manufacturers like sharpness - it is easy to measure, but it is not so important in real life (at least not to me).

Most normal lenses are not optimized for use wide open and so a test of them wide open is really not a fair test. The Leica Nokton however is designed to be used wide open and is optimized for this purpose. Indeed its resolution decreases as it is stopped down. AFAIK all lens are optimized at only one aperture.

 

[darinwc]

Thanks for the feedback. As the owner of a number of 50mm lenses from various manufacturers I have my favourites. This is sometimes because they appear very sharp, and in other cases because they render the image subjectively in a way that matches how I remember the scene. Quantifying that appearance is something that is rarely described in a way that is useful to photographers, for example people talk about "pop" as though it was an optical phenomenon with specific characteristics.

Having looked at all all kinds of lenses and the images they make over many years, I think the preference comes down to high contrast or low contrast varieties, which is associated (though not exclusively) with the evolution of optical coatings and computer aided design. Highly corrected, multi-element lenses do not always translate their advantages into a 2-dimensional viewing plane like a computer screen, and even less into a print, yet their rendering is prized by reviewers and attracts a high value, even though alternative renderings may be much more pleasing aesthetically.

There are various informal lens tests posted around the internet. Most of them are not complete, so they may include a few lenses but not the entire range.

There are also lens tests from magazine s that are reposted around. I think modern photography and popular photography published test results. But those are suspect as well. The tests were done over a period of many years. Many things changed during that time, and they may have been performed by different people, different films, etc.

But I think you answered your own question here. I reposted it because I think it is a really good explanation. Too much value is placed on resolution.

 

[ic-racer]

I've been looking for an objective test on 50mm standard lenses in the f1.8 to f2 range, the kind that came with SLR bodies from the 1960s to 1980s. There is a great deal of anecdote out there, and Googling mostly takes me to subjective comparisons between lens a and lens b, which are neither scientific nor comprehensive. Given that these are the most common lenses around, I'd be interested in any serious optical tests between manufacturer's standard lenses.

Extensively tested by the photo magazines at the time. Results show most to be blurry at the corners and low contrast wide
open and all similarly sharp stopped to f8. That is good because most of the lenses are in propriatory mounts so you dont have much choice anyway.

 

[gone]

But subjective is all there really is! Data sheets are worse than useless, as they neglect IQ, which has nothing to do w/ measured crap. Find the lens that makes the best images according to your sensibilities. Or, just get an old non A. I. Nikkor 50 2, or a Leica R 50 2 Summicron (early version), and be done w/ it. I don't see how those can be improved upon, but that's only my highly subjective opinion.

 

[Xmas]

But subjective is all there really is! Data sheets are worse than useless, as they neglect IQ, which has nothing to do w/ measured crap. Find the lens that makes the best images according to your sensibilities. Or, just get an old non A. I. Nikkor 50 2, or a Leica R 50 2 Summicron (early version), and be done w/ it. I don't see how those can be improved upon, but that's only my highly subjective opinion.

If you were to take a single coated and multi coated lens but otherwise identical lens from Cosina Voightlanders range and shoot identical scenes with alternate lenses you should detect different signatures in typical scenes.

Lots of people bought a single coated lens new!

So two questions

Did they like their shadows flashed?
Did they like their colours pastelled?

The correction of aberrations would have been similar.

Most of the shots I took yesterday were with a single coated double gauss four group design you don't need to know the manufacturer or name plate.

 

[Nodda Duma]

Most normal lenses are not optimized for use wide open and so a test of them wide open is really not a fair test. The Leica Nokton however is designed to be used wide open and is optimized for this purpose. Indeed its resolution decreases as it is stopped down. AFAIK all lens are optimized at only one aperture.

All optics are designed and optimized at the widest aperture proscribed by the design requirements. Any additional optimization at smaller apertures relates to meeting focus shift or other specialized requirements.

Just to clarify, optimization of a lens specifically means designing for the best optical performance within the constraints of the requirements set forth for the design. It does NOT mean aberrations are completely eliminated or even diffraction limited performance.

 

[narsuitus]

Blockend,

Please keep in mind that the lens tests that were performed decades ago may no longer be valid for the old lenses that are still in circulation. Years of wear and tear may have had a negative effect on optical performance.

 

[blockend]

If you were to take a single coated and multi coated lens but otherwise identical lens from Cosina Voightlanders range and shoot identical scenes with alternate lenses you should detect different signatures in typical scenes.

Lots of people bought a single coated lens new!

So two questions

Did they like their shadows flashed?
Did they like their colours pastelled?

The correction of aberrations would have been similar.

Most of the shots I took yesterday were with a single coated double gauss four group design you don't need to know the manufacturer or name plate.

I have DSB and ML versions of various Yashica lenses. The first is single coated, the second multi-coated. The first is cheap and less undesirable, the second is more expensive (relatively) and desirable. The Contax version, more so again. In reality although they render slightly differently, neither is "better" than the other.

The underlying point is people are easily swayed by anecdote, more so in a democratic blogosphere, and subjective responses take on an authoritative, objective tone. My advice would be try to divorce financial value from aesthetic value and blind test your lenses. Don't be lead by popularity and price as an index of photographic quality.

 

[Xmas]

I have DSB and ML versions of various Yashica lenses. The first is single coated, the second multi-coated. The first is cheap and less undesirable, the second is more expensive (relatively) and desirable. The Contax version, more so again. In reality although they render slightly differently, neither is "better" than the other.

The underlying point is people are easily swayed by anecdote, more so in a democratic blogosphere, and subjective responses take on an authoritative, objective tone. My advice would be try to divorce financial value from aesthetic value and blind test your lenses. Don't be lead by popularity and price as an index of photographic quality.

The Cosina lenses I referenced were manufactured concurrently (batch after batch) and the same $.

A similar situation exists with some of the Olympus OM lenses where there are single and multi variants, where the users (some off) and collectors both favour the early lower volume single coated.

It is a YMMV effect.

 

[ColColt]

I just got through a 50mm side by side comparison between the Zeiss 50 f2 Planar and the much older Leica 50f2 DR Summicron. Hands down sharpness? They were close with a slight sharpness edge to the Summicron but the Zeiss was a tad more contrasty but not much. All in all, not a nickels worth of difference. That says a lot for a 1957 lens.

 

[ic-racer]

Here are a couple:

 

[Rook]

There's more to a lens than just optical performance. How the rings turn, overall weight distribution, readability and lay-out of the imprinted numbers, material durability, etc. Things you won't find on data sheets, but can have a significant impact on handling and ease of use. I'd rather have a less sharp but well-designed, user-friendly lens instead of an ultra sharp lens that is difficult to operate. Ultimately, if I need to struggle with my lens to operate it, my photography suffers.

 

[darinwc]

IC-Racer: Glad to see that the Zuiko had a good showing. What do the dots on the bars mean?

Regarding the slower 'normal' lenses. (for slr's) I find they fall into two catagories:
1. cheap, basic designed lenses. Sometimes OK but not the best quality control. Made as a 'starter'.
2. simple but high-quality, professional optics . Designed for top-quality resolution.

My favorites so far are the Nikon f1.8 and Olympus Zuiko f1.8.

Rangefinder lenses are a different breed altogether, I cannot speak about them.

 

[flavio81]

Here are a couple:

I have more of the same series archived on other laptop. Basically the ranking, generally speaking, based only on those optical metrics, was:
#1 Leitz, then #2 Canon closely followed by #3 Nikon.
Then the rest. Seems the test was about mid 1970s.

And no, Zuikos were not the best in any case, usually with lower transmittance and/or higher flare and more spherical aberration and chromatic aberration. This no doubt due to the obsession on making the lens smaller than it should.