Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
08-11-2021, 11:42 PM - 4 Likes   #1
Loyal Site Supporter
Loyal Site Supporter
TDvN57's Avatar

Join Date: Jul 2011
Location: Berlin
Photos: Gallery
Posts: 1,137
645 Lens Tests - Identifying the top performers - Again

I don't own a lot of lenses, but too many for any reasonable person not afflicted with GAS . I decided to do another round of tests to become a little bit more consistent in my evaluation. I found a test chart referred to on Luminous Landscapes (New custom DNG profile for Pentax 645z) wherein the reviewer took a picture of a chart with the 645z DFA 55mm @ f2.8. He provided the DNG file in full resolution. http://213.112.34.223/ekr/Articles/P645D_vs_P645Z/

The chart goes to a maximum of about 12.5 lines per mm and the 55mm lens could resolve that with ease. I printed the test chart at the maximum resolution on my printer which did produce some fuzziness from 10 lines per mm onwards. Using the supplied image I could establish some calibration for the loss of resolution.

Overall I would say of the lenses I tested I did not find any lens that I would say it should not be in my bag at any given time. They are all excellent lenses but there are a few exceptional super performers in the bag. I list below the lenses I tested and what I have found in my own unscientific way.

DFA 25mm - Great lens - resolving 9 - 10 l/mm (lines per mm)

DA 28-45mm - Super performer - resolving 12+ l/mm

FA 45-85mm - Great lens - resolving 11 - 12 l/mm f8 to f11

DFA 55mm - Super performer - resolving 12+ l/mm

FA 80-160 - Super performer - 12+ l/mm f5.6 to f11

FA 120 - Super performer - resolving 12+ l/mm @ f 5.6 to f16

FA* 300mm - Unfortunately something went wrong with my tests shots and I will have to re-do. From previous tests I am inclined to list it as a super performer

FA 400mm - Almost super performer resolving 12 l/mm at f8 with comfortable 9 to 10 l/mm from f5.6 to f16. (Sweet spot at f8).

FA 400mm with TC 1.4 - Good combo with 9 to 10 l/mm at f8 & f11. Outside of the sweet spot it drops off dramatically. The f8 & f11 is the reading on the camera, not the actual aperture.

FA 400mm with TC 1.4 and TC 2.0 - Surprisingly good combo with 9 to 10 l/mm from f5.6 to f11. These are the aperture readings on the camera, not the actual aperture.
A* 600mm - Super performer resolving 12+ l/mm from f8 to f16.

A* 600mm with TC 1.4 - Super performer resolving 12+ l/mm at f5.6 & f8. This is the aperture readings on the camera, not the actual aperture.

FA 150-300mm - This lens is currently with my son, but as I recall and from pictures I took previously, this lens would be in the super performer group as well.


The surprise lens was the FA 400mm lens with excellent resolving capability. The only drawbacks are that it shows a lot of fringing and CA in high contrast settings.

So lenses I can absolutely recommend are the ones in bold above. The others are great lenses and will produce great results, they may just not be the top of the list I could test. One day I may be willing to spend USD500 or more on a pro resolution chart and then I can get even better results, but for that I'll rather buy another lens, or go somewhere to take pictures.


Notes for those interested:
1. My tests are the results of the lens and 645z camera combined.
2. The lighting was not calibrated and I had to set the white balance in the completed images
3. The resolution chart was printed at home at 4:3 aspect ratio and is 400mm high.
4. The image in Photoshop was sized accordingly to avoid distortions, yet at finer resolutions my printer could not adequately distinguish between the lines.
5. Sizes and lines per mm was measured in Photoshop once the size was calibrated.
6. The supplied image taken with the 645z and DFA55mm was much sharper than the equivalent image I took at the same settings of the reproduced chart. This drop in quality was taken into account in the assessment of the lenses.
7. I did not test the drop-off in resolution between center and edge of frame. This was done on purpose, because I was not interested in setting a standard for lenses, I was attempting to set them in order of highest performance, since they are already in my bag. My assumption was and is that all lenses in my bag will be sharper in the center than at the far edges.
8. The chart I used is not an ISO chart, yet is has many similar gradients at similar sizes.



Edit: changed FA 80-160mm and FA* 300mm lenses to bold as originally intended.

Attached Images
View Picture EXIF
PENTAX 645Z  Photo 

Last edited by TDvN57; 08-13-2021 at 12:22 AM.
08-12-2021, 12:33 AM   #2
Pentaxian




Join Date: Feb 2015
Photos: Gallery
Posts: 12,173
QuoteOriginally posted by TDvN57 Quote
Overall I would say of the lenses I tested I did not find any lens that I would say it should not be in my bag at any given time. They are all excellent lenses but there are a few exceptional super performers in the bag. I list below the lenses I tested and what I have found in my own unscientific way.
Nice report. IMO, old lenses are mostly limited by flare and CA, not as much by resolving detail , if we consider the existence of software tools such as Topaz AI sharpen being able to accentuate deep details. Given how cheap those used 645 lenses are, I'd seriously consider buying a lens trio (FA 35, FA75 and FA 120) and mounting them on a used Fuji MF camera with an adapter (used GFX cameras are a lot cheaper than Pentax 645z used, but I didn't go that route yet as I'm too deep into the Pentax FF system and not sure how much I'd practically gain from MF).
08-12-2021, 02:57 AM   #3
Pentaxian
angerdan's Avatar

Join Date: Feb 2015
Photos: Gallery | Albums
Posts: 2,631
QuoteOriginally posted by TDvN57 Quote
I decided to do another round of tests to become a little bit more consistent in my evaluation.
I'd put your useful informations into an table:

Lensl/mmapertureverdict
DFA 25mm9-10 Great lens
DA 28-45mm12+ Super performer
FA 45-85mm11-12f8 to f11Great lens
DFA 55mm12+ Super performer
FA 80-16012+5.6 to f11Super performer
FA 12012+f5.6 to f16Super performer
FA 150-300mm  this lens would be in the super performer group as well
FA* 300mm  From previous tests I am inclined to list it as a super performer
FA 400mm12f8, 9-10 from f5.6 to f16 (sweet spot at f8)Almost super performer
FA 400mm with TC 1.49-10f8 & f11 (camera aperture)Good combo. Outside of the sweet spot it drops off dramatically
FA 400mm with TC 1.4 and TC 2.09-10f5.6 to f11 (camera aperture)Surprisingly good combo
A* 600mm12+f8 to f16Super performer
A* 600mm with TC 1.412+f5.6 & f8 (camera aperture)Super performer

Last edited by angerdan; 08-12-2021 at 04:59 AM. Reason: FA 80-160mm also in bold
08-12-2021, 04:07 AM   #4
Loyal Site Supporter
Loyal Site Supporter
TDvN57's Avatar

Join Date: Jul 2011
Location: Berlin
Photos: Gallery
Posts: 1,137
Original Poster
QuoteOriginally posted by angerdan Quote
I'd put your useful informations into an table:

Lensl/mmapertureverdict
DFA 25mm9-10 Great lens
DA 28-45mm12+ Super performer
FA 45-85mm11-12f8 to f11Great lens
DFA 55mm12+ Super performer
FA 80-16012+5.6 to f11Super performer
FA 12012+f5.6 to f16Super performer
FA 150-300mm  this lens would be in the super performer group as well
FA* 300mm  From previous tests I am inclined to list it as a super performer
FA 400mm12f8, 9-10 from f5.6 to f16 (sweet spot at f8)Almost super performer
FA 400mm with TC 1.49-10f8 & f11 (camera aperture)Good combo. Outside of the sweet spot it drops off dramatically
FA 400mm with TC 1.4 and TC 2.09-10f5.6 to f11 (camera aperture)Surprisingly good combo
A* 600mm12+f8 to f16Super performer
A* 600mm with TC 1.412+f5.6 & f8 (camera aperture)Super performer
That is a great idea thanks. I think the 80-160 was also intended to be in bold. My mistake.

---------- Post added 2021-08-12 at 07:10 PM ----------

QuoteOriginally posted by biz-engineer Quote
Nice report. IMO, old lenses are mostly limited by flare and CA, not as much by resolving detail , if we consider the existence of software tools such as Topaz AI sharpen being able to accentuate deep details. Given how cheap those used 645 lenses are, I'd seriously consider buying a lens trio (FA 35, FA75 and FA 120) and mounting them on a used Fuji MF camera with an adapter (used GFX cameras are a lot cheaper than Pentax 645z used, but I didn't go that route yet as I'm too deep into the Pentax FF system and not sure how much I'd practically gain from MF).
I did find some purple fringing as most of it could be removed in ACR/Optics/Remove CA.

I agree Topaz Sharpening is lifting the raw files to another dimension, if applied lightly.

08-12-2021, 05:55 AM   #5
Loyal Site Supporter
Loyal Site Supporter




Join Date: Feb 2018
Posts: 1,892
Theuns, this is much appreciated. Do you have the 150? I'm trying to figure out if I have a bad copy - bad purple fringing on bright sources.
08-12-2021, 06:35 AM   #6
Pentaxian




Join Date: Oct 2013
Posts: 1,611
QuoteOriginally posted by TDvN57 Quote
645 Lens Tests - Identifying the top performers - Again
Thank you for putting this list together. It is a lot of work and much appreciated. I do not have any 645 bodies but bought the FA 45-85. FA 200 and the A120 in hopes of getting into the 645 system. Given how good my K1 is, I decided to forgo the 645. I sold the FA 45-85 but still have the FA 200 and the A 120. I use the A120 with an adapter on my K1 and it is fantastic for macro work.
08-12-2021, 07:40 AM   #7
Loyal Site Supporter
Loyal Site Supporter
TDvN57's Avatar

Join Date: Jul 2011
Location: Berlin
Photos: Gallery
Posts: 1,137
Original Poster
QuoteOriginally posted by ProfessorBuzz Quote
Theuns, this is much appreciated. Do you have the 150? I'm trying to figure out if I have a bad copy - bad purple fringing on bright sources.
I sold my 150 f2.8 some time ago but I still regret selling it. Fantastic lens. My reason for selling it was that at 2.8 the depth of field was too shallow to be practical and I couldn't discipline myself to stay stopped down a bit.

If I were to guess from memory, I would expect the 150 f2.8 to be similar to the FA120. I don't recall any bad fringing problems, perhaps someone may have recent experience with it.

08-12-2021, 09:33 AM - 1 Like   #8
Pentaxian
jslifoaw's Avatar

Join Date: Apr 2007
Location: Toronto/Victoria
Posts: 460
I've always found that the evenness out to the corners is what separates okay from great/excellent. Especially for times where the patient approach with a tripod and longer exposures isn't possible.

Most of the A and FA lenses have a bit of purple fringing that requires closing the aperture down a bit to eliminate but it's only an issue if you look very closely and if you have lots of overexposed elements like reflections of shiny surfaces with harsh lighting.

With the wide angle lenses, one testing risk is you might be using close distances to fit the testing chart in the frame; they might be better at infinity.

I recently took some cityscapes with the 150-300 and found it to be too uneven so buildings near the edges were quite blurry. I immediately switched to the 120 macro and had no such issues and the prints turned out great. Don't think it was camera shake.

Of the zooms I have, I would rank them (top to bottom) as 28-45, 45-85, 80-160, and 150-300. I'm not convinced the two telephoto zooms make 645(z) worth using over 35mm (K-1). The wider ones totally make it worthwhile; however, we have access to high quality modern telephotos for 35mm (e.g. DA60-250, DFA70-200, DFA70-210) and they also benefit from SR which helps a lot.

All the primes I have (25mm - 200mm) are excellent. Most are also fairly inexpensive with a couple of exceptions from the DFA series. Same with the 6x7 series.
08-13-2021, 12:44 AM   #9
New Member




Join Date: Dec 2018
Location: Perth, Western Australia
Posts: 12
QuoteOriginally posted by ProfessorBuzz Quote
Theuns, this is much appreciated. Do you have the 150? I'm trying to figure out if I have a bad copy - bad purple fringing on bright sources.
I see a lot of purple fringing with this lens too.
08-13-2021, 02:42 AM   #10
Pentaxian




Join Date: Feb 2015
Photos: Gallery
Posts: 12,173
How is the FA 75mm positioned in that list?
08-13-2021, 01:38 PM - 1 Like   #11
Pentaxian




Join Date: Jul 2009
Location: Pugetopolis
Posts: 11,008
For some lenes (eg can be used for portraits), rating them by resolving power only is not the absolute measure of a good lens.
08-14-2021, 08:28 AM   #12
Loyal Site Supporter
Loyal Site Supporter
TDvN57's Avatar

Join Date: Jul 2011
Location: Berlin
Photos: Gallery
Posts: 1,137
Original Poster
QuoteOriginally posted by tuco Quote
For some lenes (eg can be used for portraits), rating them by resolving power only is not the absolute measure of a good lens.
I fully agree with you. As I also mention in my assessment, all the lenses on the list are very good lenses, I use them all in different situations.

For portraits I use the 45-85 and the 80-160. Landscape I sometimes use the 28-45 but find I mostly use the 80-160.

Wildlife (not birds) I either go with the 150-300, or lately the fa* 300. I did a lot of wildlife with the 600mm on a gimbal and monopod. For birds I use the 600 or the 600+tc1.4.

The only time I found the 645z system inadequate was when a friend asked me to take pictures of his martial arts practice moves. In that situation I really needed a fast frame rate, and IQ and pixels were way down the list of must haves.
08-16-2021, 12:34 AM - 1 Like   #13
Loyal Site Supporter
Loyal Site Supporter
TDvN57's Avatar

Join Date: Jul 2011
Location: Berlin
Photos: Gallery
Posts: 1,137
Original Poster
Update

I tweaked my printer settings a bit and printed a 1200mm x 900mm version of the test chart with finer resolutions of 24 l/mm. I then did some spot checks to see if my original results would change. There is a slight improvement because of an improved quality chart, but the relative results between lenses would remain the same. All these were measured with the whole test chart filling the full frame of the camera.

I also did a couple of test shots where I moved as close as possible to the test chart and took as close a picture of the high resolution areas, which I now have printed to 24 lines/mm. The original results can be doubled. Thus a lens tested at 10 l/mm for the full chart in view, when composed to show the smallest possible portion of the high density parts of the chart, it resolved 20+ l/mm.

And as I mentioned before these are not lab conditions and not ideal by any chance for "testing" conditions. Although I felt it does represent daily use conditions.



---------- Post added 2021-08-16 at 04:10 PM ----------


.

PIXELS AND LENSES

I would like to highlight a few items to put these results in perspective and give a "handle" to some of the issues.

The 645z has a pixel pitch of 5.31 microns, which translates to 188 pixels per linear mm. Compare that to lenses resolving around 20 to 30 lines per mm, the discrepancy seems to indicate that we don't really need the ultra small pixels.

Thus the notion that pixel size translates into resolution is a misunderstanding.

The physical size of the sensor actually contributes more to the sharpness of the image rather than the pixel size.

Thus a 645 FF sensor will produce a sharper image of the same composition than a 645 crop sensor. And a 645 crop sensor will produce a sharper image than a 35mm FF sensor, etc..

During the de-mosaicking of the pixel values it seems likely that the surrounding eight pixels will contribute to the colour and luminance value of the ninth pixel in the center, stepped one pixel at a time.

This interpretation process can increase or decrease edge contrast to a much finer resolution than the line resolving power of the lens (approximately eq to 60 lines per mm vs 20 to 30 lines per mm). Hence edge contrast can enhance the appearance of sharpness.

The higher density of pixels compared with lens line resolution also results in higher dynamic range, purely because luminance and colour are not bound to line resolution. Which may explain why Pentax always refers to the quality of the final picture instead of concentrating on line resolution of their lenses.

Although the smaller pixel pitch should theoretically improve the interpretation of dynamic range there is a point of diminishing returns when the pixel pitch is too small to capture adequate light.

That would suggest that a 100mpx 645 FF sensor (4.6 pixel pitch) would be a better performer than the 150mpx 645 FF sensor (3.76 pixel pitch).

In both instances the images will be about 20% sharper than the current 645 cropped sensor, just because the FF sensors are about 20% larger.

Apologies if the above is a bit cryptic but I am trying to keep it short and to the point.

As always comments and opinions are welcomed.

Last edited by TDvN57; 08-16-2021 at 01:12 AM.
08-16-2021, 03:52 AM - 2 Likes   #14
Pentaxian




Join Date: Feb 2015
Photos: Gallery
Posts: 12,173
QuoteOriginally posted by TDvN57 Quote
Thus the notion that pixel size translates into resolution is a misunderstanding
Yes and no. It's very simple: the lens is like an analog filter, there is no hard stop to what information it transmits, just at some point the variation of image micro-contrast (high frequencies) melts into noise (random uncertainty) of the observing light sensitive surface. The digital sensor samples what the lens projects onto it, and it cuts all periodic details right at nyquist (half number of pixels/mm). Therefore, with 14bits pixel depth (that's capable of detecting tiny variations of contrast, much much lower than the traditional 50% MTF contrast) there is no such thing as "Sensor out-resolve the lens", until optical details get soaked into sensor noise.... which can be as much as resolving power as 400lpp/mm but only at low ISOs. At higher ISOs, most extra optical detail would fall below pixel noise, and therefore the higher pixel count would be wasted. The choice of pixel pitch is more a matter of pixel performance, data efficiency (the more pixels the less efficient) and how versatile is the camera with regards to ISO range.

---------- Post added 16-08-21 at 13:09 ----------

QuoteOriginally posted by TDvN57 Quote
In both instances the images will be about 20% sharper than the current 645 cropped sensor, just because the FF sensors are about 20% larger.
Not necessarily because ALL lenses (included the Zeiss Otus) have their lp/mm drop from center to edges. My experience with printing taught me that a square print (from the center of image crop), can be printed as large as the full 3:2 image un-cropped, due to lens optical properties being circular (closer to square), not rectangular. Now, to judge a camera resolution, I don't count the total number of pixels, I don't look at the horizontal resolution (long side), but the vertical resolution (short side) tells me the largest size of prints I can make for a given print quality objective. I would prefer cameras with square sensors.

Last edited by biz-engineer; 08-16-2021 at 04:14 AM.
08-16-2021, 09:29 AM   #15
Loyal Site Supporter
Loyal Site Supporter




Join Date: Dec 2017
Photos: Gallery | Albums
Posts: 1,137
@biz-engineer. Hear hear.

@TDvN57. A way to think of the pixel resolution is to divide the pixels per mm by the ratio of object distance to focal length. Hence, 188 pix/mm at the focal plane using a 100mm lens is equivalent to 18.8 pix/mm a meter away. When using a resolution chart, particular care is needed to know just what scale the USAF test chart is actually presenting. (I'm not suggesting that you didn't do that.)

The Nyquist frequency for 188 pix/mm is 94 cy/mm. This frequency yields zero output, although the MTF has some bumps after that point. By color band, the Nyquist frequencies are at lower frequencies at half the color sampling pixel density. I have doubts that the Beyer filter function actually raises the true spatial frequency response in the color bands, but may join in creating a higher intensity MTF depending on how the nature of the estimation algorithm optimized for a particular nature fits the nature of the actual nature under observation.

For comparison, Portra 160 has at least 30% response (green, blue) at 100 cy/mm, and red is over 10% response. As noted, these respective MTFs will be modulated by the lens MTF, which varies over the field and may also vary by color.

To put it another way, with digital a field of pure red roses viewed from a particular distance might have missing blooms with the not-missing blooms having high contrast, while with film no blooms are missing, but all have lower contrast.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
645d, 645z, aperture, camera, chart, combo, f8, fa, ff, lens, lenses, line, medium format, mf, mm, pentax, performer, pixels, resolution, results, sensor, tc, tests, topaz
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Abstract Performers gilberteid Post Your Photos! 1 04-24-2020 02:10 PM
Hello again, and again.. and again..... ghigoblin Welcomes and Introductions 9 01-11-2017 02:22 PM
Streets Oxford street performers Tekmongoose Post Your Photos! 2 08-03-2016 11:33 AM
People Chinese New Year performers bobbotron Photo Critique 5 02-21-2016 03:52 PM
For Sale - Sold: SMC D FA MACRO 1:2.8 100 mm WR -----PRICE DROPPED again- again-again-again watchman323 Sold Items 12 12-09-2013 11:18 AM



All times are GMT -7. The time now is 08:12 AM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top