Surprised at the IQ of this lens...

[rpcrowe]

I was originally hoping to use my collection of Canon lenses with my Sony APSC cameras but, the AF did not work out to my satisfaction... I purchased a used Sony 28-70mm f/3.5-5.6 which is the kit lens for a full frame Sony e-mount camera. I wanted to see what a native E-mount auto focus was like (I found out that it is great) and I figured that his would be a decent lens to shoot family parties and similar occasions. I really didn't expect any kit lens be of an great quality. However, I am quite content with this focal length on an APSC format camera for shooting people!

Yesterday, I did a shoot in some rather cramped quarters where my 50mm was a bit long and my 30mm prime was a bit short. So I thought today, "What the heck, I'll bring the kit lens also." I ended up shooting virtually all day, at the themed shoot, using that lens indoors with studio strobe lighting. Like the thread title states, "I am surprised at the IQ of this lens..." Of course, I was shooting around f/8 to f/11 and most lenses with an f/5.6 maximum aperture should produce their most decent quality at those apertures... But. I only paid about $125 USD on eBay for this little light weight lens...

[DanK]

I don't doubt that it is fine, but an image at such low resolution doesn't show us much. In the light box, it is only 1279 x 862, and it will be lower resolution yet on many screens if viewers don't look at it in the light box. Even images from low-quality lenses can look fine at this low resolution.

This is why I find it odd that there are so many "sample images" articles on the web about new lenses. Unless they post links to large files, the images always look good.

[Manfred M]

+1 to Dan's comment. The only way that I generally look at IQ in an image is when viewed at 100%. This is not something we will likely ever need when viewing on a screen, but it becomes quite important to us printers working on large format prints.

That means posting in a lossless image format too, something that no website I have ever used handles.

Anything looks good on a small screen. I've had many images that look great on the screen on the back of my camera (even when I magnify a portion of it), but when I get home an look at it on my large screen, the defects come out quite clearly.

[rpcrowe]

I have it in full resolution JPEG on my smugmug site but and it looks pretty darn good. On the other hand, it is quite unhandy today in working with very high resolution images.

One must realize that, despite my love of large prints, the standard display media these days is the Internet. I am making a totally wild-ass guess that for every image that is printed (disregarding the Walmart and drugstore 3x5 inch tiny prints) there are there are thousands of upon thousands of very good to excellent images on photo websites everywhere you look. Perhaps, except for personal pride, the majority of photographers may no longer really need the optimum resolution that many of today's top-line cameras/lenses are capable of producing...

This is not the case with wedding photography which most often has prints as the ultimate product nor with some other commercial uses. This also is certainly the not the case with photographers like Manfred who are involved with producing the ultimate quality prints...

However, I am beginning to think that today's photographers may be in a state of jousting with windmills in the search of the ultimate photo quality at any price of money and gear weight. IMO, the camera and lens that one needs today is one that will produce acceptably good quality FOR THE USES THAT THE ORIGINATOR MAKES OF THEM...

However, in comparing quality... Even in low resolution, I can see that my Tokina 28-70mm f/2.8 ATX paired with the Canon 7D Mark-2 runs rings around the Sony combination. I am mentioning this combination since both the Canon 7D2 and the Sony APSC are crop sensor cameras while the two lenses are equal in focal range plus both camaera/lens combinations are in the same area of used pricing (with the Sony perhaps just a bit lower)...

[Manfred M]

Richard - I went to your SmugMug site and downloaded your first image, imported it to Photoshop and reviewed it at 100%.

The lens does give reasonably decent results but is not as sharp as higher end glass.

Here is a crop at 100% of a recent studio shoot using the same aperture range as you have.


1. Maeve eyes taken with a "pro" Nikkor f/2.8 70-200mm lens at f/8







2. Celyne Eyes




If you go into the Lightbox view and toggle the image back and forth, the difference between the two images is clearly noticeable and the Nikkor is noticeable sharper. Of course, we are looking at a difference in cost that is an order of magnitude higher than the kit lens, so one would expect to see some difference.

PS - I think the Sony's eye focus algorithm is part of the problem too. The image looks like the camera front focused as the knife blade and the fingers around the hilt of the knife appear to be more sharply focused than the eyes.

[rpcrowe]

I never expected the kit lens to be as sharp as a better (more expensive lens) what has happened in the Sony line of lenses, except for the G-Master (inordinately expensive and heavy) the Sony zooms, even the G lenses and including the famous Sony-Zeiss lenses are not (by DXOmark standards - if you believe their testing methods) are certainly not a great deal better than the little 28-70mm kit lens...

As an example: The little inexpensive 28-70mm lens gets an overall score of 16 with a sharpness score of 8 for a average used price range of $125-150 USD while the Sony Zeiss Tessar 16-70mm f/4 gets an overall score of 17 with a sharpness of 8 and costs between $600-700 USD.

We know that as you go up the scale of lens quality, you get less and less improvement for the extra money that you spend. However, that seems like a great price to pay for really little additional quality Especially as the Sony 85mm f/1.8 lens (for which I paid $450 used) gets an overall score of 46 )placing it in the excellent category) and a sharpness of 40.

Of course, sharpness is not the only criteria in judging a lens and the DXOmark tests might be a bi flawed but the difference between 16 or 17 and 40 is tremendous. That is the reason that I generally stick to primes with the little Sony cameras and aso the reason that I have a pair to shoot in tandem.

It appears that at the present moment, the Tamron 28-75mm f/2.8 lens with a Sony mount may be the leading contender in the Sony "Bang for Your Buck Club"... However, even that lens has me wondering It seems like it is basically the same lens as the previous one with Canikon mounts only adapted for Sony E-mount cameras and costing $879 while the Canikon mount lenses cost only $499...

I do find that a zoom lens can be pretty important in certain shooting circumstances and I guess that I have been spoiled with the excellent Canon zooms available such as the 24-70mm f/2.8 in both the i and ii versions as well as one of my favorites, the EFS 17-55mm f/2.8 IS lens.

[Manfred M]

However, even that lens has me wondering It seems like it is basically the same lens as the previous one with Canikon mounts only adapted for Sony E-mount cameras and costing $879 while the Canikon mount lenses cost only $499...

I suspect that you are 100% right. Companies like Sigma, Tokina and Tamron design a generic lens and then build in an adaptor that takes into account the camera's flange distance and electronics.

The reason for the higher price? Possibly just to recover their development / tooling costs for the Sony mount and after it has been on the market for a year or so, the price should drop. If they have to pay Sony a licence fee to use their mount, then this could be reflected in the price as well.

[xpatUSA]

+1 to Dan's comment. The only way that I generally look at IQ in an image is when viewed at 100%. This is not something we will likely ever need when viewing on a screen, but it becomes quite important to us printers working on large format prints.

That means posting in a lossless image format too, something that no website I have ever used handles.

Anything looks good on a small screen. I've had many images that look great on the screen on the back of my camera (even when I magnify a portion of it), but when I get home an look at it on my large screen, the defects come out quite clearly.

I agree with the sentiment. To examine IQ, I normally view at 100% (not smoothed) in FastStone Viewer and often more with integer multiples of 2X, 3X, 3X ... etc.

Currently embroiled in a raging battle about that over here where it is disparagingly call pixel-peeping:

https://www.dpreview.com/forums/post/62952488

As to a lossless image format, DPR handles PNG and GIF ... though I've never tried them here. Also JPEG can be loss-less or almost, for example FastStone's RGB photometric instead of Y'CbCr:





Did JPEG2000 ever take off ... ?

[DanK]

I think there are two separate issues in this thread. One is the relative quality of this lens, about which I know nada. the other is whether it is informative to post a low-res image to document quality. The only way it is useful is if the low-res image looks bad. If it looks good, it isn't necessarily informative.

[Manfred M]

Interesting to see the memorial to the famed lens designer Ernst Abbe. Otto Schott received his doctorate in glass chemistry from this university and Carl Zeiss studied there too. The three of them ended up founding the Carl Zeiss company and with it the modern optical industry. Abbe did the lens design, Schott crafted the specialized optical glasses and Zeiss supplied the mechanical and machining know how to put everything into a working package.

[rpcrowe]

I can definitely see the difference in sharpness between this kit lens and my prime lenses on the A6400 such as the Sony 50mm f/1.8 OSS and the Sigma 30mm f/1.4. I shot with the primes on Saturday and used the zoom on Sunday.

I like using two camera bodies when I am shooting with primes however, there was only one flash trigger available for me at the photo shoot. When I am shooting with my own flash equipment, I own two triggers for my Canon and Sony bodies. (ie: two R2 Pro S TTL Transmitters).

IMO, there is really no mid-range Sony zoom lens, except for the ultra expensive and equally heavy G-Master models, that has very good to excellent IQ. You would think that the thousand dollar Sony Zeiss 16-70mm f/4 OSS lens would shine on an APSC camera but, it doesn't really.

Although some folks don't think much of the DXOmark lens rating system, I think that you can compare two lenses within that system to see which has the better sharpness. The above Sony-Zeiss lens has just minimally increased sharpness as compared with the 28-70 Kit lens. However my prime lenses for these cameras, such as the Sigma 30mm f/1.4, and both the Sony 50mm f/1.8 and 85mm f/1.8 really excel in the area of sharpness. However, the 70-200mm f.4 G lens is no slouch either...

Presently, it seems that the 28-75mm f/2.8 Tamron may be the best bang for the buck in mid-range zoom lenses for the Sony e-mount camera but... it is really a bit heavy for an APSC Sony body.

I think that I will just continue using my primes for photo shoots (despite the fact that a zoom is a bit easier to use) and save the kit lens for stuff like family get togethers...

[Manfred M]

A couple of thoughts on my part Richard.

Lens resolution (i.e. lines per mm) is only one measure of sharpness and frankly most modern OEM lenses are quite sharp. What people perceive as sharpness is often micro-contrast; where light and dark areas meet. In most circumstances this is far more important than the resolving power of a lens. Resolution is only meaningful when tested in lab conditions; a camera that is solidly mounted (no camera movement) and is generally not achievable in most real world photography, especially when one hand-hold the camera.

Most of the "Zeiss" lenses one sees for Sony cameras are in fact Sony lenses wearing the Zeiss badge. This is little more than a marketing exercise on the part of Sony and Zeiss. If you want a real Zeiss lens, then you had better be willing to shell out some really serious money for the modern Batis (Sony only, I believe) or Otus (made for different camera brands). These are Zeiss designs and specs, but are built in Japan for Zeiss by Cosina.

With respect to the whining about DxOmark lens reviews; most of the complaints are just plain silly. One can poke holes into any test protocol, but as long as we are comparing "apples to apples", then we can at least see how different lenses compare to each other. There never has been a perfect test approach in the digital age as camera variables (sensor and processing engine) are every bit as important as the lens to the final results. In the film days, at least that aspect of the test could be fixed, so long as the film manufacturer did not update the film itself or the processing chemistry.

Most of the images I have taken this year are with my Nikkor 28 - 300mm lens; definitely not high end glass; it's a touch soft, has strange aberrations and is somewhat prone to lens flare. I've taken award winning images with my Panasonic Lumix 14 - 140mm lens too. Good technique, composition and impact are the most important part of image making.

The only place where I do find a difference is when I make large prints (A2 / A3 sizes or 13" x 19" and 17" x 22" or larger). Here one can see difference between different sensor sizes and the softening that comes from upsampling. If I weren't a printer and exclusively posted to the web, I would think differently, but when it comes to printing, size matters and I honestly wish I could afford a medium format digital camera. If I could, there is no question I would be shooting that format.

[xpatUSA]

A couple of thoughts on my part Richard.

Lens resolution (i.e. lines per mm) is only one measure of sharpness and frankly most modern OEM lenses are quite sharp. What people perceive as sharpness is often micro-contrast; where light and dark areas meet. In most circumstances this is far more important than the resolving power of a lens.

I think that much of the literature does include contrast as well as spatial frequency when talking about "resolution".

For example, the popular "MTF50" (a measure of contrast) is stated at a spatial frequency, the higher the better. The lens plots of MTF (again, contrast) are given for specific spatial frequencies, e.g. 10 lp/mm, 30 lp/mm, etc.

And there is the so-called "extinction" metric, the spatial frequency at which MTF falls below some value (sometimes 9% per Rayleigh).

In other words, spatial frequency and contrast ratio are inevitably linked together, and micro-contrast is simply a particular part of that domain, rather than being able to be discussed separately.

Bob Atkins discusses it all rather well here:

http://www.bobatkins.com/photography.../mtf/mtf2.html
.

[Manfred M]

I think that much of the literature does include contrast as well as spatial frequency when talking about "resolution".

For example, the popular "MTF50" (a measure of contrast) is stated at a spatial frequency, the higher the better. The lens plots of MTF (again, contrast) are given for specific spatial frequencies, e.g. 10 lp/mm, 30 lp/mm, etc.

And there is the so-called "extinction" metric, the spatial frequency at which MTF falls below some value (sometimes 9% per Rayleigh).

In other words, spatial frequency and contrast ratio are inevitably linked together, and micro-contrast is simply a particular part of that domain, rather than being able to be discussed separately.

Bob Atkins discusses it all rather well here:

http://www.bobatkins.com/photography.../mtf/mtf2.html
.

Ted - I suspect you, I and very few others would bother reading at this level of analysis.

This is why sites like DxOmark are so popular, they provide a single number ranking and a bit more data for people who are willing to drill down a bit deeper. Nice, simple and open to criticism, because of the simplification.

As I've said before, this is all well and good, but unless the photographer is shooting in close to ideal conditions, the theoretical performance of the best lens will be masked by the reality of the shooting situation.

[rpcrowe]

When I open a RAW image from my primes vs. the 28-70mm kit lens, there is absolutely no question that the prime lenses are superior in all respects to the little zoom. It is absolutely evident that the primes are sharper and have more pleasing contrast than the zoom. It is "almost" like the earlier contests between zooms and primes; before the zoom lens technology had advanced to the point at which it is now.

I remember the first zoom lens for a 35mm still camera that I ever used. It was a 43-86mm Nikkor somewhere between 1966 and 1968. Shooting with this lens, I thought this is wonderful! I am always going to use this lens However after viewing my slides when they returned from processing, I changed my thoughts, these are terrible... I never shot with that lens again!