Zoom...today vs yesterday

[Dali]

Cooltouch, interested to hear more about the Vivitar S1 28-90. From what I read, vignetting is a plague with this lens. Did you have a similar experience? I always found this focal range to be the perfect fit with my habits but mixed reviews cooled me down...

 

[whojammyflip]

I have been doing some testing of zoom lenses, using MTF slanted edge tests. I like to test and quantify performance myself and have been comparing 1980's manual zooms, and used a modern Pentax DA 18-55 for reference. Previously I posted to APUG about comparing Pentax to Olympus, in particular the fabled MIJ 50 vs the Pentax M 1.7 50, where I found the MIJ was not all that its claimed to be. Similarly, in summary here, the 1983 Vivitar S1 28-90 is sharper than the modern Pentax 18-55, particularly at the wide end where the difference is substantial.

====
Lots of whats written on the internet about lenses is nonsense. Potentially even this post. I like Photrio/APUG as it seems full of photographers with hands on experience, with little tolerance of stuff that doesn't make sense or isn't true.

Post 98: "Your "door stop" remark then wasn't made from any personal experience or deep knowledge of using zoom lenses, but purely out of prejudice and your ignorance of their capabilitys"

Similarly in post 106: "Sometimes guys just repeat things they Hear/Read so they seem like they agree with all the other experts"

Post 100: "Why don't you just shoot with a pinhole camera?"

It strikes me as happening a lot on the internet, people feel free to post stuff which is just shot from the hip. Then others start regurgitating it, and it somehow gains credibility as it remains unchallenged as if the ultimate truth.

====
In post 20 there was written:

"The difference between most 40 year old zoom lenses and most modern zoom lenses come down to computers - computer aided design and computer aided manufacture. The computer made it possible to design and manufacture at much higher quality, for reasonable cost...."

I don't dispute the computer aided design improved design, but when did this start to materialise in a commercial consumer zoom lens? I would pinpoint the year as being 1983, for normal zooms. By 1983, there were zooms being produced which are sharper than zooms today. For example, the Vivitar Series 1 28-90 is substantially sharper than my current Pentax DA 18-55 kit zoom. Admittedly, I am comparing a top end zoom from 1983 with a current kit zoom.

I think this is a reflection of the 'minimum viable product' philosophy; that the current DA 18-55 is designed to maintain an affordable entry point for a new comer to photography, it is designed to only meet the lower quality requirement. I note that in testing over on other sites, the DA 18-55 is regarded as being one of the best kit zooms.

This consideration of what is good enough was captured in post 88 "Note in this example the design is "good enough" to meet the requirement without providing unnecessary performance: Over-designing can be as disastrous as under-designing".

====
In post 49, there is the claim that version 2 of the Vivitar Series 1 70-210 zooms is the 'best'. This is an ambiguous statement, as the definition of best is not clear. By what criteria is that pronounced? A lot is written about the Vivitar Series 1 70-210 zooms. Once you start designing some MTF testing, it becomes clear that testing will requires many photos to be taken. You have various important variables such as f-stop, focal length and test distance. You could potentially have to evaluate a lens by processing 100 different images. Also, the testing needs to be across the frame. Using MTF Mapper, something like 1300 slanted edges are evaluated with each capture. This is where computers are better than humans. Obviously, the 1300 edges are then aggregated, getting summary stats like means and standard deviations of contrast for a given resolution.

In my testing, I looked at the long zooms including the first 3 Vivitar Series 1 70-210 zooms, the two Tamron SP 70-210 zooms, Kiron's 70-210 f4 zoom, Kiron 80-200 f4, Kiron 70-150, Pentax M 75-150, Pentax M 80-200. I also included a Tamron 278D 80-210 and Pentax F 80-200. Having measured them, I think all the Vivitar Series 1 70-210 zooms are great. You cannot really go wrong with any of them. The Kiron made 70-210 has prime like performance at 100mm. The Kiron 70-150 was the best performer at 70mm. The Tamron 278D was amazing at f8 135mm, but awful when opened wider than f8. I wonder whether the resolution I saw from the Tamron 278D was due to it having a massive lens hood.

====
On zooms performance relative to primes, it really varies according to focal length and f stop. I was very surprised that the performance of the zooms can be on a par with the primes once stopped down to f8/f11, or even exceed a prime. This was contrary to what I had read previously and common internet lore. The difference was most noticeable at the long end, ie 135 or 200mm some zooms are sharper than my primes, notably the Tamron SP 19AH. If you are looking at the ultimate performance, resolution limit of the lenses, then the primes probably have it (on average across f stops or across the frame), but its a very close run thing, and in usage something I dont think you would notice. I was seeing all the good lenses reaching up to 60-70 lpmm for MTF25. At this level of resolution, it really matters whether its mounted on a tripod. Anthing hand held and the motion from the camera will mean the difference in the lens performance cannot be seen.

 

[AgX]

You made some viable points, especially concerning image qualities. Some here speak of resolution, some of contrast, some meanwhile refer to MTF. Whereas the MTF concept is something already established in the early 50s and still unknown to many. And computers entered lens design already in the 50s too.

Intersting I found the turning point of 1983 you gave, I shall think about this.

But then there are many other aspects too, from speed over bokeh to handling. I somewhere listed that all.

 

[AgX]

In this context seemingly fits my question, why a Canon 100-200mm 5.6 made it all the years from the FL- over the FD- to the New-FD-mount?

 

[Deleted member 88956]

...
Ferrania backer? hahahah KS is a ripoff. kiss your money goodbye.

And what is your evidence to back that up? It helps thinking some comments through before posting. You made a blatant attempt to discredit an operation that is complex, has had some success and as of right now it isn't over until it is.

 

[Nodda Duma]

The 1983 turning point makes complete sense. That was 8 years after the industry-standard Code V optical design software developed by Optical Research Associates was released as a commercial product that could be installed at an optical company’s facility. Code V was a big leap forward in computer-aided lens design capability, in no small part due to its accurate modeling, superior optimization routines, and extensive database of starting-point designs. 8 years seems like a big delta, but one can imagine ~2-4 years for the management at a company like Nikon to approve the purchase of such a capital expense, a couple years for the lens designers to train and confirm the results, and then 2 more years for a Code V - designed lens and put it into production.

 

[Diapositivo]

IMHO Today lenses are much lighter, more compact and cheaper. Industry attention focused on those aspects rather than extreme sharpness or correction of other aberrations.
In today's digital world, even Hasselblad relies on software to correct certain distortions. Sharpness seems to be the desire of very dedicated photographers, some of them actually, while the majority of photographer just count the camera's megapixels.
Quality as we know it is not understood any more, as can be easily seen by the amount of digital images with burned (clipped) highlights.
The average time people passes in front of an image is also probably reduced to a fraction of what it was at the time of film.
Images are thought to be "consumed" in one second and a half, two seconds.

The hiatus, in expectation of quality and dedication to good results, between professional and "casual" photography is probably bigger than ever.

 

[faberryman]

IMHO Today lenses are much lighter, more compact and cheaper.

None of these attributes is true.

 

[Mainecoonmaniac]

None of these attributes is true.

I beg to differ. Some newer lenses have plastic barrels which makes them cheaper to manufacture and reduces weight. I was surprised when I bought my first Canon DSLR and the kit lens was plastic. Versus my old Canon FD lenses which were all metal. I thought the lenses felt cheap, but the pictures I made with them are every bit as sharp as my old lenses.

 

[faberryman]

Maybe I go back too far as prices go, but all the new lenses seem much more expensive. And they are all larger due to autofocus motors. I have been put off with each new announcement of a lens when I see the picture of it. Today's FF cameras and lenses are no where near as compact as their film counterparts. It is why I shoot Fuji. It approaches the size of my OM bodies and lenses. All the FF cameras and lenses are bloated.

 

[abruzzi]

I think the market is broad enough to encompass both trends. A kit 50 from 40 years ago was heavier and in many cases better built than a kit 18-55 today. But there is also an obsession with high resolution MTF designed lenses, spare no expense or size. My old pentax 40mm ƒ2.8:

The new Sigma 40mm ƒ1.4:

Ok, it's two stops faster, but that is another trend in lenses at the moment. I also wouldn't assume that plastic is worse--there are very well made lenses in plastic housings, and cheap metal lenses.

 

[AgX]

In the 80s there were different design approaches to zoom lenses:
-) higher speed
-) higher ratio
-) more compact built

 

[MattKing]

It is difficult to compare old vs. new, because the markets are different.
For a long time, a zoom lens was a specialized accessory that people were willing to pay extra for. The "standard" lens, that most people used when they compared cameras, benefited from economies of scale.
Then, for a long time, there were "kit" zoom lenses, which benefited from economies of scale, but needed to be small, light and economical, in order to keep camera prices reasonable. A fixed focal length "standard" lens became a somewhat more expensive accessory. There were also both lower priced zooms and higher priced zooms competing for the photographer's extra money.
Now, people who buy cameras are becoming more and more rare. So some of the cameras are directed to people who spend a lot of money, and the zoom lenses that serve as a standard lens are very good and quite expensive.
For other cameras, the standard lenses are quite economical, so not as good.
And fixed focal length lenses have become relatively more expensive.

 

[Nodda Duma]

In the 80s there were different design approaches to zoom lenses:
-) higher speed
-) higher ratio
-) more compact built

Different *requirements*. Similar design approaches.

 

[Paul Verizzo]

Several people have mentioned 1983 as an important year in zoom lens technology.

It so happens that I bought a Tokina 35-200mm in anticipation of a trip to the Grand Canyon with my daughters who lived in another state that year. Almost 7:1, very cutting edge for the year.

It served me well.

Now I have an 18-200 on a digicam. Due to sensor size, it works out to be about 28-300. Holy moly, over 10:1. And compact.

Progress.......

 

[AgX]

A 10x zoom one already got in 1970.
For Super-8 ...

 

[Diapositivo]

Modern lenses make extensive use of plastic lenses, not just plastic barrels. That allows producers to produce aspheric lenses at a much lower cost.
Since the '80, we had a great reduction in weight (keeping lens complexity equal) because of plastic lenses. But also the complexity went much further.
It's not just the internal motor, but also the higher zoom ratio or the increased luminosity.
In the '80 AF zoom existed: they were heavy and expensive. Internal-focus lenses were expensive and rare. Lenses making use of aspheric elements were very expensive. Apochromatic lenses were rare and expensive.

Inevitably there has been in the last decades a huge progress, both in weight and cost and performance. It's true for any field of human activity.

But this progress was not focused on higher resolution or lower distortion for standard fixed-focal lenses, but on weight, zoom ratio, luminosity, cost.

 

[Paul Verizzo]

A 10x zoom one already got in 1970.
For Super-8 ...

That's cheating. I was referring to 35mm. Harder to do.

 

[Ben Hutcherson]

In the '80 AF zoom existed: they were heavy and expensive. Internal-focus lenses were expensive and rare. Lenses making use of aspheric elements were very expensive. Apochromatic lenses were rare and expensive.

It amazes me that I can go out and buy a $100 DSLR kit lens(typically 18-55mm) and get a complex design that usually includes both aspheric elements and some type of exotic low dispersion glass. The former, of course, use to be pretty much the exclusive domain of ultra fast lenses, and the latter super-teles.

My Nikon 14-24mm f/2.8 is the sharpest UW lens I've ever used at all focal lengths. I don't mean the sharpest wide zoom-I mean the sharpest lens covering these focal lengths period. Of course, it also weighs a couple of pounds and has a front element roughly the size and shape of half a grapefruit(and can't use filters easily) but it's sharp. This is a lens that's under $3K new, although of course I didn't buy mine new.

I have both an older 80-200mm f/2.8 AF-D(push-pull, no tripod collar), and a newer 70-200mm f/2.8 AF-S VR(first gen). The xx-200mm f/2.8 is a staple lens for both Nikon and Canon, and they seem to try and leapfrog each other every few years in design. My 80-200mm is decent enough wide open, while the 70-200 is outstanding at all apertures as long as you don't look too closely at the corners wide open. The newest version, the 70-200mm f/2.8E FL, is supposed to be the sharpest zoom of this class that Nikon has made.

Yes, it's heavy, and it's also 1-1.5 stops slower than the AF primes I would use in its place, but the combined cost and weight of the lens is lower than buying those lenses individually. Nikon has neglected a lot of their sub-300mm primes long enough, too, that I don't know if they are any better than the zoom(with the possible exception of the macro lenses, although they have benefits the zoom lacks).

BTW, I'm applying my comments to both film and digital use. Yes, I use both-both have their benefits for different applications for me, but at the end of the day what matters to me is that the image looks like what I envisioned. Yes, I have shot sports on film with manual focus lenses, and been able to get good shots. Back in the mid-2000s when I was in high school, film was still cheaper than a good DSLR for me, and I suffered through the grain of films like Superia 1600 and Tri-X pushed to EI 3200 when I needed to(I couldn't afford to shoot TMAX P3200 regularly then, esp. not in the quantities I was shooting Tri-X, and it didn't really make a difference in the school news paper printed in half tone). That was even using manual advance cameras on occasion, although my senior year of high school I did "upgrade" to a Canon T90 and though having 5ps on tap was a wonderful thing. Nothing beats a good wet print on fiber based paper for B&W, nor can I match the colors I get from Velvia if I use it on the right subject. Still, though, when I'm at one of my nephew's baseball games or chasing kids around photographing them or whatever it's easy to get use to fast AF(with tracking) and useable ISO 6400 so that I can stop action. I rarely shoot bursts-I'd rather rely on timing-but if I did it's hard to stomach that I can shoot an entire roll of film, rewind it, and have another loaded in well under 30 seconds in the F5.

 

[Luckless]

Modern lenses make extensive use of plastic lenses, ...

Which lenses are you referring to with this?

I can't think of any model lens I've seen, beyond exceedingly cheap ones seen in highly questionable webcams, basically toy cameras, or some bottom of the barrel gear for telescopes/microscopes that made use of plastic optical elements. And now you have me curious if anything from a middle or upper quality tier has ever made serious use of a plastic element in the actual light path, aside from coating films on glass/crystal elements - Even most smart phones cameras seem to rely on non-plastic lens elements.

 

[miha]

Nikon AF 28-70mm f/3.5-4.5 from 1991 sports a plastic aspherical element. It must have been one of the first of this kind.

 

[Diapositivo]

It's thirty years that aspheric elements are built in optical plastic. Glass aspheric elements are very expensive. Nowadays aspheric elements are much more widespread because they are made out of optical plastic. I would be surprised if there is a lens with glass aspheric elements in production nowadays.

Extremely-high quality lenses do use plastic elements. It's twenty years I don't follow closely industry novelties, but the trend begun decades ago.

Now I am looking at the Nikon Nikkor lens range. There's a AF-S NIKKOR 400mm f/2.8E FL ED VR, gee, 400 f/2.8, weighting 3800g, 16 elements in 12 groups. The front lens is 14.2 centimetres in diameter. When I was young such lenses did not exist in the photographic domain, and simpler lenses needed a two-tripod-arrangement to support them properly.

"Plastic" in this context means very sophisticated, high-technology material, not the Kodak Instamatic lens. I think they are all derived from metachrylate ("plexiglass"), which is more transparent than glass itself, has a superior resistance to shock than glass, and is lighter. Lens producers will not refer to this as "plastic" of course. They will not tell you, or will not stress, that there are "plastic elements" in your lens. The plastic elements are probably the more expensive, but they would be much more expensive if made in glass proper.

Many lenses now - and since many years - will not any more stress characteristics such as internal focusing, extra-low dispersion, apo-chromatic behaviour, aspheric surface, because they became much more common than they used to be. Part of the reason is in the use of "plastic" materials.

 

[Nodda Duma]

Glass aspheric elements are not as expensive as you think with modern production techniques (molded glass, deterministic polishing, MRF). I think you’ll find that most aspheric elements in modern optics actually *are* glass....once you get above the disposable camera level of quality.

It’s not the 1990s anymore.

Btw the optics in the most prevalent consumer cameras — smart phone cameras — consist of molded aspheric glass elements. Equivalent plastic elements would be too large.

 

[wiltw]

Computers are used in lens design since the mid-50s.
In the meantime not only computer aided manufacture but new lens materials emerged.

Could you cite known lens design using computers in the 1950's ?
It is known that the UNIVAC 1 was not developed until the early 1950's, and the US military wanted to get one in the mid-1950's.

A bit of history on Wikipedia says,

"In 1974, the Ponder & Best (Opcon/Kino) Vivitar Series 1 70-210mm f/3.5 Macro Focusing Zoom (USA/Japan) was widely hailed as the first professional-level quality very close focusing "macro" zoom lens for 35mm SLRs. Ellis Betensky's (USA) Opcon Associates perfected the Series 1's fifteen element/ten group/four cell formula by calculations on the latest digital computers.[179] Freed from the drudgery of hand computation in the 1960s, designs of such variety and quality only dreamt of by earlier generations of optical engineers became possible.[180][181] Modern computer created zoom designs may be so complex that they have no resemblance to any of the classical human created designs.

"The optical zooming action of the Series 1 was different from most earlier zooms such as the Zoomar. The Zoomar was an "optically compensated" zoom. Its zooming cell and focal plane compensating cell were fixed together and moved together with a stationary cell in between.[182] The Series 1 was a "mechanically compensated" zoom. Its zooming cell was mechanically cammed with a focal plane compensating cell and moved at different rates.[183] The tradeoff for greater optical design freedom was this increase in mechanical complexity."
​

A paper on the evolution of optical design mentions

" The complex conjugated of the Fourier transform of the reference is placed in the Fourier plane and therefore multiplied by the Fourier transform of the input scene. Lens 2 performs a second Fourier transform that gives in the output plane the correlation between the input scene and the reference. Implementing a complex filter with the Fourier transform of the reference was the main challenge of this set-up, and Vander Lugt proposed in 1964 to use a Fourier hologram of the reference as a filter [12]. Figures 2(b) and 2(c) show respectively, the output correlation peak for an autocorrelation when the correlation filter is a matched filter and when it is a phase only filter [13]."
​

...and goes on to state

"At the beginning of the seventies, the digital computers were able to compete successfully with the coherent optical computers for SAR applications and finally they won and this was sadly the end of radar signal optical processing."
​

Then in 1978 paper
Lens Design Without A Computer - SPIE Digital Library
© (1978) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE)

" Prior to about 1930 all lenses were designed by logarithms, and from 1930 to about 1960 they were designed by hand with the aid of a mechanical desk calculator and a book of sine tables.
The use of an optimization program on a large computer for the design of lenses is a very recent phenomenon. Prior to about 1930 all lenses were designed by logarithms, and from 1930 to about 1960 they were designed by hand with the aid of a mechanical desk calculator and a book of sine tables. It is the purpose of this paper to indicate the kind of procedure that used to be followed when lenses were designed in this way. A numerical example is included to illustrate the process.
© (1978) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE)."​

 

[AgX]

You are right, the acquisition date of of a computer does not say when what lens was released designed by means of that computer. And I cannot name you the latter.

In any case, computers for optical design were first acquired in both parts of Germany between 1950 and 1958. Germany was ahead of the Univac as such..