Help judging performance of my FE55z, FE35f2.8z horizon/infinity

Started Jul 5, 2016 | Questions thread
l_d_allan Veteran Member • Posts: 5,093
Re: Different testing rigor for "good copy vs bad" compared to getting valid MtfMapper numbers?

JimKasson wrote:

l_d_allan wrote:

I'm wondering if you would make a distinction between a "level of attention to detail" that is adequate for "good copy vs bad copy testing", vs a higher level for getting valid, repeatable MtfMapper numbers within +/- 10 lp/ph

Plus or minus ten? With lenses like the Otus 85, I barely get +/- 50 even with a focusing rail and a razor blade! Lesser lenses aren't such a problem.

My bad. I considered using a [fill-in-the-blank] instead of a specific number, but was too lazy to re-read several of your blog articles on what was the typical granularity you were getting. My recollection was that some TLW (or LR?) articles have numbers with 3 significant digits, such as 1230 or 2340, implying +/- 5, but those may have been LR averages.

Am I incorrect to think that:

If I were going to do brick wall testing, I'd print up five ISO 12233 charts, and attach them to the wall so that one is in the center and the other four are at the corners. Then, if I didn't want to do slanted edge testing, I'd read the extinction resolution off the rosettes.

Would that be like what TDP provides in their "Image Quality" section, such as:

Again to my shame, I've always found such extinctions to be difficult for me to interpret by subjective eyeball. Am I the only one this applies to?

My speculation has been that the number in the middle might be close to the "extinction resolution" you mention ("32" in the review above) . Or not?

My experience is that is doesn't take much wind at all to make it difficult to tape targets to a brick wall.

If I was "taping targets", I'd also be inclined to use 9 targets rather than "just" 5 for coverage of the APS-C and rule-of-thirds parts of the sensor.

However, my observation is that at proper distance, there is invariably so much detail in the brick walls that the hassle of taping targets isn't justified for the level of rigor I think is satisfactory.

Mirrors are also difficult to place on brick walls, in my experience.

I've reluctantly given up on being able to get valid MtfMapper numbers, so that reduces my motivation to use printed targets. However, I'll agree that for purposes of comparisons between lenses, it might be worth trying again for my "baseline prime" FE55 or adapted Canon EF 35mm IS  (both of which I consider to be mini-Otii).

I presented a version of this test that doesn't require precise alignment on my blog, but I now believe the OOF LED is a superior test and much less likely to produce false positives.

Is that the "silhouetted razor-blade based protocol", which seemed oriented to getting valid MtfMapper numbers?

If not, a link, por favor ? Sorry for again being lazy, but I'm not finding that specific blog article with searches including "extinction" and "rosette".

  • ...
  • that a typical DPR member can get valid captures
  • from which to evaluate whether they have a "good copy" or not.

I dunno. From what I see posted in the forum, I despair of achieving all that.

Or, is a simpler de-centering test mostly or fully adequate for "good copy vs bad copy"?

My belief is that it is, but I don't' know that for a fact, as I don't have any lenses that measure as materially decentered, so I don't know how sensitive the test is.

Prof. Hank has such lenses, and has tried the OOF LED test with them, and says that it works. I tend to take Prof. Hank's word on things like this. I respect him a lot.

IIRC, Roger C. at LR has also written about how to test for de-centering. However, I also get a sense from his articles that some or much of the Sony q/c problems may come from de-centering and hard-to-adjust elements, but he also didn't "know for a fact" how much it represents.

My speculation is that perhaps valid "best practice" for "good copy vs bad copy" might be done in multiple stages:

  • relatively simple de-centering test (OOF LED or LR approach)
  • if passes and concerns still exist about the LUT (Lens under Test), more involved far-horizon testing at a tilt (if adequate scene available)
  • if passes and concerns still exist about the LUT, more involved with something like a brick-wall
  • if passes and concerns still exist about the LUT, more involved with silhouetted razor blade, perhaps with motorized, computer-controlled focus rail

Opinions differ, but I think that Sony has earned a reputation for so-so q/c. Thus, the unfortunate reality seems to be that purchase of a Sony lens implies a willingness to tackle the learning curve of valid "good copy vs bad copy" testing. Big bummer.

I also recall RC mentioning lens elements being tilted within the barrel can happen.. That may or may not be similar to de-centering in effect and prevalence. I think of them as separate issues, but that may be ignorance on my part.

Would the relatively simple test for de-centering also detect tilt'ed elements, if applicable?

 l_d_allan's gear list:l_d_allan's gear list
Panasonic Lumix DMC-TZ3 Panasonic Lumix DMC-FZ47 Canon PowerShot S110 Canon EOS 5D Mark II Canon EOS 600D +27 more
Post (hide subjects) Posted by
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow