Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
First solid measurements of lens variance
#11
Good grief...

 

For the record, I measured all of the lenses for this - in fact I've measured all 702 lenses tested so far for this project.  This is what I'm at Olaf for the summer to do, other than write software to work with the 'raw' data and present it nice. 

 

The otus lenses are manufactured by Cosina in Japan in the same facilities with the same testing equipment.  The best manufactures use testing similar to a dumbed down and sped up imatest setup, which is not remotely as sensitive as an MTF bench.  It is not surprising that they are hardly better than other ZE/ZF Zeiss lenses - they are built under the same capacity.  The mega-kilobucks lenses made in Germany by Zeiss do fare much better.  Despite being a zoom, the 28-80 is from what I can tell from the 3 I've measured, going to be as good as the Otus, despite being an enormously complex zoom.  Same for the 70-200. 


$4000 is a pittance for such a complicated instrument as the Otus.  The price range for a lens to be tested on an MTF bench per copy is typically $10,000+ in the US at least.  It is prohibitively expensive in terms of time, though I do think the software I've written for it provides the fastest method of anything available due to how hands-off it is. 

 

The 50 STM likely does as well as it does because it is drop-dead simple and it is lightweight, so there is no large inertia to decenter it.  It's also the newest Canon lens and is likely produced on an automated production line which is more accurate than humans doing drop assemblies.

 

Regarding the scores - the Otus vs 58mm Nikkor difference is insignificant.  If we took them out to 25 copies they would likely become the same.  Otus vs sigma art is a significant difference. 


Dave, failure copies are pre-screened. 


A true bad copy of the 24mm f/1.4G looks something like this: https://www.dropbox.com/s/c1u239jo5bjtiw...s.pdf?dl=0


While a decent one looks more like this: https://www.dropbox.com/s/xeqnnkx2oz31qh...s.pdf?dl=0

 

If my own standards were used, not a single copy of any rokinon model would pass inspection. 

 

In my honest opinion no lens should be scored lower than 6-6.5 which indicates about +/- 5-10% but alas, this is too tight a tolerance. 

 

Keep in mind that despite the Otus' similar score to the 58/1.8G, there is this massive difference present: https://www.dropbox.com/s/sfm1gkhripnsn6...O.pdf?dl=0


 vs https://www.dropbox.com/s/1rk45m9s4nv38g...O.pdf?dl=0

#12
I understood, even the worst Otus copy to be better than the best Canon 50/1.8 STM, no problem with that. But given most if not all manufacturers use the MTF method for testing, I expected the "best" lens not necessarily best manufactured, but tested the hardest way and with tightest tolerances.

 

Sure, quality control never produces quality - it's just making sure, that the variation doesn't go below a "more complaints than profit" point.

 

It's very easy, @AiryDiscus and @dave9to5: If Zeiss calls the highest price ever for a 55/1.4 it has to be the best, from copy to copy. Everywhere. Does a customer really needs to do a centering or MTF test by himself to be sure he/she got a good copy?

 

Plus, being the failure copies pre-screened: Doesn't this make the whole testing results questionable? As lensrentals doesn't stock a lot of expensive lenses but quite a number of cheaper ones. So, with pre-selection coming into play, we're looking here at the "avant-garde" of their production but not to the "out-in-the-wild" situation.

 

So, how much more possible would it be to get a lousy copy? 

 

Thanks for the samples, @AiryDiscus

#13
Quote:Good grief...

 

Dave, failure copies are pre-screened. 


A true bad copy of the 24mm f/1.4G looks something like this: https://www.dropbox.com/s/c1u239jo5bjtiw...s.pdf?dl=0


While a decent one looks more like this: https://www.dropbox.com/s/xeqnnkx2oz31qh...s.pdf?dl=0

 

If my own standards were used, not a single copy of any rokinon model would pass inspection. 

 

In my honest opinion no lens should be scored lower than 6-6.5 which indicates about +/- 5-10% but alas, this is too tight a tolerance. 

 

Keep in mind that despite the Otus' similar score to the 58/1.8G, there is this massive difference present: https://www.dropbox.com/s/sfm1gkhripnsn6...O.pdf?dl=0


 vs https://www.dropbox.com/s/1rk45m9s4nv38g...O.pdf?dl=0
  Thanks Airydiscus,

As I thought none of the lenses seemed to be excessively bad/decentered, the tests showing just variations in sharpness etc. 

#14
Quote: 

 

It's very easy, @AiryDiscus and @dave9to5: If Zeiss calls the highest price ever for a 55/1.4 it has to be the best, from copy to copy. Everywhere. Does a customer really needs to do a centering or MTF test by himself to be sure he/she got a good copy?

 

Plus, being the failure copies pre-screened: Doesn't this make the whole testing results questionable? As lensrentals doesn't stock a lot of expensive lenses but quite a number of cheaper ones. So, with pre-selection coming into play, we're looking here at the "avant-garde" of their production but not to the "out-in-the-wild" situation.

 

So, how much more possible would it be to get a lousy copy? 

 

Thanks for the samples, @AiryDiscus
 Yes, as I thought  Mr. Discus, this tests shows the favourable side of lens variation......... and it's dead easy to get a lousy copy, nothing easier!  

 The lens industry is the most unique industry in the world and if it wasn't for the retail laws of the right to return a product within fourteen days (depending on the country concerned) for refund/exchange, we would be at the "beck and call" of lens manufacturers business decisions!  

 

   Decentered lenses spend most of their life flying round in airplanes and postal delivery vans landing only  briefly on DSLRs to re-continue their circumnavigation of the globe.

 

 

https://www.flickr.com/photos/124690178@N08/

#15
Quote:Nope. I confirm my statement: it is indeed disappointing given the price point and the marketing message conveyed by Zeiss. Their very site states "The best standard lens in the world". Slightly pretentious isn't it?

Now, by reading this, you would expect their QC to be unmatched, especially given the ridiculous price they charge for it. However, it's not quite the case. The last 3 lenses below are easily ahead of it while being much much cheaper:
  • Zeiss 55mm f/1.4 Otus => 6.5
  • Nikon 58mm f/1.4 => 6.7
  • Zeiss 50mm f/2 Makro => 7.3
  • Sigma 50mm f/1.4 Art => 7.5
  • Canon 50mm f/1.8 STM => 9.3
Ironically, it's the cheapo Cannon 50 STM that leads the packs.

From a customer point of view, if I pay extra for "the best" quality, I'd also expect premium QC, better than any other lenses. Clearly, it's not the case here.
 

Actually I am not surprised that the 50 STM leads the packs. It is the slowest lens in this series.

The slower a lens is the more robust is it against decentering.

#16
Quote:Actually I am not surprised that the 50 STM leads the packs. It is the slowest lens in this series.

The slower a lens is the more robust is it against decentering.
The Zeiss 50mm f2 is slower, the Nikkor 50mm f1.8 just as slow.
#17
"We had previously published the Consistency number for 24mm lenses, so I'll include those in a table for comparison. Wide angle lenses tend to have more variation, so we hoped the 50mm lenses would have less variation than the 24mm lenses."

 

Would it not be more correct to state that the variation of the lenses (50mm and 24mm) is similar, but that at wide angle the same  variations show up in an "amplified way" with wider angle lenses in MTF tests?

 

If you take the focal length into account, the 24mm QC seems on par with the 50mm lenses?

#18
In a nutshell: Samyang lenses are a lottery.

[Image: VarianceByEFL.png]

#19
Quote:Actually I am not surprised that the 50 STM leads the packs. It is the slowest lens in this series.

The slower a lens is the more robust is it against decentering.
The slowest was the Zeiss 50/2 macro.
#20
Random thought: is there any correlation between the variation score and either number of groups or elements?
<a class="bbc_url" href="http://snowporing.deviantart.com/">dA</a> Canon 7D2, 7D, 5D2, 600D, 450D, 300D IR modified, 1D, EF-S 10-18, 15-85, EF 35/2, 85/1.8, 135/2, 70-300L, 100-400L, MP-E65, Zeiss 2/50, Sigma 150 macro, 120-300/2.8, Samyang 8mm fisheye, Olympus E-P1, Panasonic 20/1.7, Sony HX9V, Fuji X100.
  


Forum Jump:


Users browsing this thread:
1 Guest(s)