For more information on stock photos by Peter Spiro available for downloading from istockphoto.com, please click here.

To view some of my recent digital photographs, available for stock photography purchase, please click on this icon to view my portfolio at istockphoto.com

One of the great things about the Pentax system is that every bayonet mount lens ever made by Pentax (going back to 1975) can still be used with a high degree of functionality on the most modern Pentax SLR. These lenses can be focussed manually, and used in aperture priority automatic exposure. (Nikon is the only other brand which has allowed its older lenses to be used on modern autofocus cameras, but generally with less metering capability then Pentax.) Hence, it is worthwhile keeping track of the tests of the older lenses. Lenses last much longer than cameras, and many of the older lenses are available in the used market.

Pentax autofocus cameras even provide focus confirmation when used with older manual focus lenses: a green diamond lights up in the viewfinder to indicate the point of correct focus. In the first generation autofocus bodies, the SF series, there was even an arrow that showed what direction the focus ring should be turned in manually to attain focus.

(One important point that should be noted: some non-Pentax brand K-mount lenses have protrusions that will cause them to get stuck when put on an autofocus camera. Ricoh program lenses are especially dangerous. No non-Pentax lens should be put on a Pentax autofocus camera until it has been carefully examined.)

As of December 2004, Pentax has two digital SLR models on the market. They should be given credit for maintaining a greater degree of lens compatibility than any other manufacturer. Any K-mount lens made since 1975 can be used to give aperture priority auto-exposure on a Pentax DSLR with an extra push of a button.

Why is this important? For some kinds of photography (still lifes, scenics, architecture), manual focussing is still ideal, and if you are going to focus manually, it is preferable to have a lens that its designed for it. Nothing beats the smooth precision of the focussing ring on a Pentax M or A lens. Autofocus lenses are made with a much looser mechanism, in order to allow the motors to turn them easily. For this reason, the lens elements are often less well aligned, and optical quality is lower as a result. You won't find any manual focus lenses you can put on a Canon SLR. You can put some older Nikon lenses on a Nikon DSLR, but the exposure meter won't function. This backwards compatibility is a unique and important feature of the Pentax DSLRs.

There are many well-known shortcomings to lens tests, but nevertheless they do give a useful indication of the approximate ranking of a lens along the quality scale. Modern Photography magazine used to report resolution tests on several different lenses in almost every issue, and I found a local university library that has a nearly complete collection of these. I have presented below the results for all the Pentax lenses I could find. Most of these lenses are readily available in the used photo equipment market.

The now-defunct magazine Modern Photography was merged into Popular Photography around 1989. Jason Schneider and Herbert Keppler now occupy roughly the same positions at Popular Photography that they did previously at Modern Photography. (Long may they flourish. Mr. Keppler's column is always the best part of the magazine.)

The largest lens test archive is maintained by Klaus, who discusses some of the problems of lens testing in more detail (look for the "Easy Lens Guide" in his Photo Zone). The results from the Modern Photography tests appear to correlate reasonably well with those from the other magazines for the few lenses that appear in both.

For many types of photography (though by no means all), the single most important lens characteristic is lens sharpness. High resolution is particularly desirable in nature, scenic and architectural photography. It should be noted that there is no perfect measurement of resolution either. It is not an absolute, and tests give different results depending on the contrast of the subject. The amount of resolution that is needed for a subjectively sharp shot depends on the size of the enlargement made. That is why Popular Photography's tests nowadays do not present the raw resolution data. Instead, they give a "subjective quality factor" that depends on the size of the enlargement. They suggest that the pictures taken by good versus bad lenses are pretty much indistinguishable in 4x6 inch prints, but become noticeable at 8x10.I’m sceptical about this, and I think they are being too kind to the manufacturers of inferior lenses in taking this approach.Their point about print size is valid only if you view the print from a “normal viewing distance.”If you happen to like looking at your 4x6 prints up close to see the detail in the picture, you will pretty quickly notice differences in quality among lenses.

Contrast is also an important lens characteristic. However, weak contrast in the lens can be partly overcome by selecting film and paper (for prints) that yield higher contrast. There is no remedy for low resolution.

The final sharpness of a picture is a function of the lens, the film, the accuracy of focussing (which is enhanced by wide aperture lenses), and the steadiness of the camera.Camera steadiness is maximized by putting it on a tripod and locking up the mirror, but this is not always feasible or desirable.Other things equal, a faster shutter speed will enhance sharpness.While the rule of thumb often used is that the shutter speed should be “at least” as high as the lens focal length being used, for most photographers sharpness is maximized at a shutter speed considerably higher than this.A high quality lens, which reaches close to its maximum sharpness at f/5.6 or even f/4 is desirable because it allows the use of faster shutter speeds with relatively slow, fine grained films.This is one important difference between the quality of zooms and single focal length (“prime”) lenses.Many zooms produce excellent results, nearly indistinguishable from that of prime lenses, at f/8 or f/11.However, this smaller aperture requires either slower shutter speeds, or restricts their use to fast films or bright sunlight.In the table below, the aperture at which the lens reaches its maximum resolution is noted.

Another caution to bear in mind with resolution tests is that, for practical purposes, they are always conducted with the target being photographed at about 50 times the lens’s focal length.E.g., for a 50mm lens, the target is about 2.5 meters from the camera.This distance may be typical for certain types of pictures, such as portraits and interior shots,but is not relevant for scenics and architecture.Wide angle lenses, in particular, tend to exhibit somewhat more curvature in their field of focus, which means that the centre and edges of the picture are in focus at different distance settings.This is more of a problem at close distances, where depth of field is shallow, than at long distances.To cite on example, tests by Yoshihiko Takanami find that the Pentax-M 35mm f/2 set to f/5.6 is considerably less sharp at the edges of the picture than the Pentax-A 50mm f/1.7.Mr. Takanami conducted these tests at 50 times focal length, and I have no reason to doubt that his results are correct for that distance.However, I have found that for pictures taken at infinity, the 35mm f/2 is in fact noticeably sharper at the edges than with the 50mm, and in fact it is probably my sharpest lens for infinity focus.MTF tests, such as Photodo’s, are conducted at infinity focus, and so may be more relevant for wide angle lenses.

The same point also applies to some extent to the Pentax "pancake" lens, the 40mm f/2.8. While it is not quite as sharp as the 50mm lenses in general use, when focussed at infinity there is no discernible difference.

Some actual photographs I have taken with some of these lenses at infinity focus can be seen in a separate article.

Some recent advertisements from Carl Zeiss claim that the Zeiss Planar T 60mm macro lens can resolve 150 lines per mm. Perhaps one should expect no less from a lens that sells at a discount mail order dealer for about US$800. If true, this would give it a higher resolution than any lens ever tested in Modern Photography. The peak I found there was 110 for a 90mm macro by Tamron. Of course, this is the best resolution at the best f/stop for the lens.

Specialized macro lenses are made for maximum resolution, which is why they cost so much. However, ordinary lenses by even the two top brands, Leitz and Zeiss, do not come close to this. The resolution of a lens depends on both its optical design and the quality of its production. Of course, you can get to the point of diminishing returns in lens fineness. To get an 11 x 14 inch print, a 35 mm negative has to be enlarged quite a bit more than a 6cm negative, so you can achieve a higher effective resolution with an average quality lens on a medium format camera than you could with the ultimate lens on a 35 mm camera.

One of the criticisms frequently made about lens tests is that the magazine tests a single lens, and there is deviation in manufacturing quality. I'm pretty sure that the frequency distribution of this is such that a test based on an outlier lens is more likely to show worse quality than the average for this model. This is because there are more ways to make a lens bad than to make it unusually good. The maximum quality of any particular model has an upper limit set by the lens design and the manufacturer's standard for fineness of lens grinding, which depends in turn on the types of abrasives used. There isn't going to be a case where a worker "forgets" and accidently goes way past the fineness expected. However, there might be a case where he gets lazy and doesn't finish grinding one of the surfaces properly, and this will get into circulation if the company has poor quality control.

Another concern is how carefully the magazine did the test in the first place. In the Modern Photography tests, there were a couple of cases of what had to be either bad test results or serious typos. One of these was for the Pentax-F 50mm f/1.4. The test showed a string of 98's for resolution, while the written comments said that it was "slightly above average." If it had really been 98, it would be a lot more than "slightly" above. Therefore, I have not included it in the test table.

The numbers in my table represent the average resolution and contrast for the four stops from f/4 to f/11. For almost all lenses, performance drops off significantly after f/11. Except when you need extreme depth of field, there is not much excuse for going past that, since you can just boost your shutter speed to reduce the exposure in bright light. Leitz does not even have an f/stop greater than f/16 on most of their lenses. I have also included the resolution at full aperture for each lens, so you can gauge how useful that aperture is. The resolution wide open for some of the old fast lenses appears to have been quite good.

For the vast majority of lenses, the peak of centre resolution is at f/8. For a few it is f/5.6, and even fewer f/11. However, corner resolution usually peaks one stop smaller. This is because corner resolution benefits from increased depth of field. Part of the reason why corner resolution is lower is because the field of focus of a lens is actually a curve, whereas the test pattern is a flat surface. Hence, if the tester sets the focus for the centre, the corners will be slightly out of focus.

The reasons why lens resolution peaks at around f/8 are very clearly explained in an article by Karl Volk. Stopping down a lens tends to minimize the aberrations that are most prevalent on the periphery of the lens. However, as the aperture is reduced, the laws of physics intrude, via the phenomenon of the diffraction of light through a small opening. The two factors tend to balance out at f/8. In theory, if designers developed a perfect aspheric lens with no aberrations, its best performance would occur wide open.

It is interesting to note that the central resolution on the classic Pentax 50mm lenses was as high or higher than competing lenses from Leitz, Zeiss and Nikon. However, Pentax has a distinct pattern of being inferior at the corners, indicating that in their lens design they did not go to as much trouble to try to achieve a flat field of focus. The importance of this obviously depends on what you are trying to photograph. If it is a central object (such as a person), with other things around it at other distances, then corner sharpness will be relatively less important.

All four pairs for which tests of both M and A versions exist show the A lenses as being somewhat better than their M series predecessors, even though in most cases the basic optical design was unchanged. In particular, corner resolution and contrast are better. It is possible that Pentax improved its manufacturing quality at the time that it introduced the A lenses. One difference that is apparent to the naked eye is the coating. If you look straight at an A lens, the front lens element is nearly invisible, which suggests a superior coating technology.

Another interesting question is the quality of zoom lenses compared to fixed focal length lenses. The results for the few lenses shown here indicate that zooms, in some part of their range, could be as good or better than some fixed focal length lenses.

However, the same zoom lens can be fairly good in one part of its range (such as the 28-80mm @50mm) and relatively poor at another (the same lens @28mm).

It should also be remembered that there can be considerable sample to sample variation. This is especially the case with zoom lenses, where small mechanical misalignments can seriously impair performance. Ideally, anybody considering buying a lens should do so from a source where a refund is available in case the particular one they get is a dud.

Two of the lenses included in the tables are the Pentax-M 40mm "pancake" and a corresponding 45mm lens from Zeiss. I have the Pentax version, and when it's on my ME it makes the camera so compact that it fits comfortably into a coat pocket. It is a very nice lens, and when focussed at infinity it produces excellent results. For both the Pentax and Zeiss lenses, the performance of the pancake lens, while quite good, is below that of the 50mm f/1.7. This accords with my own experience of pictures taken with my 40mm versus 50mm lenses. The former are sharp, but the latter are even sharper.

The sharpest lens of all in these tables appears to have been the SMC Pentax 50mm f/1.4 lens that came out with the original K cameras in 1975. However, it should be cautioned that testing standards may have varied over the years. Resolution tests are particularly sensitive to the film used. To demonstrate this principle, one article in Modern Photography featured resolution tests conducted using an ultra-fine grain photomicrographic film. With this test, lenses that were normally rated at about 70 lines/mm delivered about 100 lines/mm. Given that they were aware of this fact, one would hope that they used the same film over the years for their tests (assuming that whatever film they started out using was still in production.) In fact, an article in Modern Photography in 1982 claimed that their tests had been unchanged for about 10 years. As of 1982, they assigned slightly different verbal interpretations to the resolution numbers, but the tests for generating the numbers stayed the same.

SMC Pentax Manual Focus Lenses
	Central Resolution	Corner Res.	W.O. Cent. Res.	Central Contrast
18mm f/3.5	56	32	50	70
20mm f/4 M	63	36	60	73
24mm f/2.8	66	39	49	47
24mm f/2.8 A	70	46	49	57
28 mm f/2.8 M	60	50	58	54
28 mm f/2.8 A	66	47	57	63
40mm f/2.8 M	64	46	56	60
50 mm f/1.2	64	49	42	60
50 mm f/1.4	73	58	54	68
50 mm f/1.4 M	63	56	63	62
50 mm f/1.4 A	68	49	46	68
50 mm f/1.7 M	71	51	69	61
50mm f/1.7 A	70	62	51	69
50mm f/2 M	61	46	46	40
85 mm f/2 M	52	57	40	74
135 mm f/3.5 M	50	43	45	51
135 mm f/2.8 A	47	44	44	60
200 mm f/4 M	44	39	40	57
300 mm f/4 A	45	39	41	59

The 200mm and 300mm lenses have considerably lower resolution than the shorter focal length lenses. Apparently, this is typical, and lenses of other brands such as Nikon and Canon had similar low numbers. In this instance, most smaller aperture zoom lenses that go to 200mm appear to perform better than the 200mm prime lenses.

Are autofocus lenses inferior to the same lens design in a manual focus mount? It is sometimes claimed that this would be the case, since autofocus ones (in order to turn rapidly with minimum friction) have very loose internal mechanisms that let the optical part shake around relative to the film plane. The tests are not very conclusive in this regard, since I did not find many that were tested in both manual focus and autofocus. In the table immediately below, the autofocus Minolta and Nikkor are not obviously worse than the manual focus Leitz and Zeiss lenses. By contrast, the Pentax 50mm f/1.7 is somewhat weaker in autofocus than in manual. The contrast is also worse on the autofocus version, and this can hardly be due to the autofocus mechanism, but instead suggests that Modern Photography happened to get a bad sample. Other tests of the autofocus Pentax 50mm f/1.7 find it to be near the top of its class, such as the WWW lens test archive.

Non-Pentax Lenses for Comparison
	Central Resolution	Corner Res.	W.O. Cent. Res.	Central Contrast
Minolta 50mm f/1.7 AF	66	58	49	65
Zeiss Tessar 45mm f/2.8	63	61	50	61
Zeiss Planar T 50mm f/1.7	67	60	50	63
Leitz 50mm f/2 R	66	57	55	63
Nikkor 50mm f/1.4 AF	63	56	63	62
Tokina 70-210mm f/4-5.6 @210	59	53	56	60

In the table below, I have included test results from the 35 -70mm f/3.5-6.7 zoom lens on the Pentax IQ Zoom point and shoot camera, to give a sense of how its quality compares to the lenses made for SLRs. Central resolution at 35mm is decent, but the corners are poor, while at 70mm even the central resolution is quite poor.

SMC Pentax-F Autofocus Lenses
	Central Resolution	Corner Res.	W.O. Cent. Res.	Central Contrast
50mm f/1.7	60	50	45	54
35-70mm f/3.5-4.5 @35	59	48	51	43
@50	56	49	50	45
@70	56	50	50	44
35-70mm f/3.5-6.7 @35	64	38	57	43
@70	42	39	40	51
28-80mm f/3.5-4.5 @28	48	31	44	54
@50	61	50	56	51
@80	63	47	51	44
70-210mm f/4-5.6 @70	56	46	40	46
@135	47	43	43	50
@210	41	35	47	49

The FA 28-70mm f/4 aspheric lens has tested out as among the best lenses in its class, and I can attest to its high quality from personal experience. It is better than any of the 28-80 lenses that Pentax has made. Unfortunately, it has now been discontinued.

A large number of tests of modern Pentax lenses can be found on the Internet at www.photodo.com. These are tests done using an MTF (modulation transfer function) measuring apparatus without taking any actual pictures. These tests calculate a weighted average of the sharpness and contrast combination for the lens, estimating the total amount of useful information that comes through the lens at different apertures. They may not be a perfect guide to how good the lens is for the particular type of picture taking you do, but it is as close as one can come to finding an absolute objective standard.

The following is a sampling of Photodo test results for some of the most common lenses. The score is out of a maximum of 5. The results confirm that single focal length lenses remain much sharper than even the best zooms overall. However, at their best apertures (generally f/8) and focal lengths (generally the short to medium range), zoom lenses can be very good. More details, including performance at different apertures and focal lengths, can be found at www.photodo.com. There seems to be an emerging consensus in the Pentax on-line discussion group that the Photodo results are meaningful. At first, there was a lot of skepticism about the fact that the 28-200mm Tamron tests out as much better than the corresponding Pentax-label lens, even though both are made by Tamron. However, the result seems to be confirmed in practice, and it is suggested that the Pentax version uses an older optical formula without the low-dispersion glass that is in the latest Tamron version.

	Photodo Overall Score
SMC Pentax-FA 28mm f/2.8	3.8
SMC Pentax-FA 43mm f/1.9	3.6
SMC Pentax-F 50mm f/1.4	4.6
SMC Pentax-FA 50mm f/1.4	4.2
SMC Pentax-F 50mm f/1.7	4.4
SMC Pentax-A 50mm f/2.0	4.0
SMC Pentax-F 35-80mm f/4.5.6	2.2
SMC Pentax-FA 28-70mm f/4	3.3
SMC Pentax-FA 28-80mm f/3.5-5.6	2.0
SMC Pentax-F 70-210mm f/4-5.6	3.4
SMC Pentax-FA 70-200mm f/4-5.6	2.9
SMC Pentax-FA 28-200mm f/3.8-5.6	1.7
Tamron LD 28-200mm f/3.8-5.6	2.7

There are also many fine lenses made by third party manufacturers that fit Pentax cameras. Indeed, as the market for SLR lenses shrinks, Pentax has occasionally re-badged lenses made by companies such as Tamron and Cosina as Pentax lenses. A very useful compilation of information about third party lenses is maintained by Robert Monaghan. One feature that many third party lenses lack is Pentax's super-multi-coating. I have found with two Sigma lenses I have tried that flare in backlit scenes is much worse than with SMC Pentax lenses. SMC coating is not just a marketing gimmick -- it is something that really makes a difference.

I have also written up some views on focussing accuracy with Pentax cameras.

Pentax was a bit too slow in coming out with a DSLR when I was making the move to that format, so I switched to Canon (although seeing the new K10D I'm feeling a bit remorseful about having switched). However, as an afficionado of nice old metal-bodied manual lenses, I was pleased to discover that adapters are available to use them on Canon digital SLRs. To see how classic manual focus lenses such as Pentax (Takumar) M42 can be used on Canon DSLRs such as the Digital Rebel XT (EOS 350D), click here.

To read a review of Quicktax versus Ufile income tax software for Canadians, please click here.

Return to my photo home page.

COPYRIGHT NOTICE Copyright (c) 1998, 2000 by Peter S. Spiro. This document may be freely distributed. It may be reformatted for purposes of compatibility. It may be freely used for personal and educational purposes, but it may not be used for commercial purposes without prior written consent of the author. It may be included in toto or in part as components of other documents with proper attribution.