ATRAX PHOTO by Clifton Beard | Lens Sample Variation For Real

Lens Sample Variation For Real

February 11, 2015  •  1 Comment

I have been looking to source a very wide angle lens for My Fuji X-system cameras. While the 14mm f2.8 is a fantastic lens and is the ideal  21mm equivalent for most landscape work, I wanted an ultrawide of 10mm or wider for an upcoming trip to Hong Kong and for astro photography.


Samyang do a very well-rated 8mm fisheye lens, which is hugely wide and I know from my brother's own copy of this lens that it can be a stunning performer.  It provides a uniquely wide angle view and will be my widest lens, the current honour going to Nikon's 14-24mm on full frame.  Not only that, but its image projection is far less distorted than many fisheyes, making it more suitable for general wide angle work. 


I decided to source a Hong Kong lens due to the relatively low risk at this price point and the considerable saving over official UK stockists, so I ordered from OneStop Digital, based in Kowloon.  Ordering and delivery was very efficient and the lens arrived within a few days.  The Samyang 8mm f2.8 fisheye I ordered was the newer version ll, which has an improved optical performance and better focus ring over the version l.  On opening the box impressions were good and you have to hand it to Samyang that their lenses do offer very good value at the price point they are pitched at.  The lens seemed well-finished and reassuringly dense.  It was well-packaged in a foam shell and came with its own little pouch.  Its not premium build but it would not be fair to expect that.  In terms of handling the only real negative was the rather stiff focus ring which clearly lacked the refinement of premium manual focus lenses such as the Nikkor AIS and Zeiss, but hey, its no big deal on a fisheye and look at that price again!  


However, as soon as I popped it on my X-T1 and took a few very quick test shots in the fading light, I immediately saw what I thought was a significantly asymmetric performance across the frame.  I carried out more tests using more careful methods over the next couple of days, and yes indeed, the lens was pretty gruesome on the left side.  For astro work, which means wide open performance, this was going to be a paperweight!  In all honesty the softness/blur could easily be identified on the camera LCD and at normal viewing sizes on my 27 inch iMac.  At 100% pixel view it was absolute mush in places.  The entire left side of the image field was far weaker than the right, which showed remarkably good sharpness and contrast almost right out to the very corners.  For approximately one third of the image field from the left side the image was weak.  Not only was it soft but there was clear indication of that ugly smearing and double image blur you often get with optical defects.  It may be easy to argue than a general softness is "within specs" but such asymmetry can never be right!  Anyway, I tested it further using a tripod, careful enough alignment, electronic shutter, remote release, 2 camera bodies etc and got the same results.


​Upon stopping down the performance did improve (obviously), but generally it did not get to an acceptable level on the left side until f8.  It was never as good as the right side which was very decent at f2.8, good at f4 and really excellent at f5.6.


Here are a few images that illustrate what I am trying to say, although it may be harder to see at blog sizes.  Remember that a fisheye lens has huge depth of field so plane of focus, field curvature and defocus are not the issues...its the uneven performance that is.  According to Photopills (great photographers App) at a focus distance of 2m DOF should extend from 0.72 metres to infinity at f2.8, to give but one example.


In the below image of the bridge, look at the left side and in particular the left corner where the railings are weak compared to those on the right.  Higher up the left side has poor definition compared to the right, and this extends out to infinity.


​In the below image and near 100 percent crops, look at the left side and lower left corner.  The definition in the railings is totally different each side, with those on the left having an ugly double blur.  Look through the railings to infinity on the road and the differences are massive in rendering of detail and contrast on the road surface and foliage.  This is especially clear when looking at the central crash barrier, white line paint and foliage of the central reservation, which is at lower left/right on the crops.  On the right detail of the foliage and the road surface texture can even be seen, but on the left this detail is non-existent.






In this example taken about 2 metres from a nice, symmetrical tower, look at the definition in the bricks and the contrast differences across the frame.  Softness is very evident on at least the left third of the frame, worsening the further left you go. The crops are not from the extreme corner but just a bit up from there on each side to show the bricks clearly.  The ground is mush on the left side, whereas it shows detail at right and the crops ruthlessly show the difference in performance on the bricks.






I clearly did not want to keep the lens as I do not want to shoot stopped down all the time, so contacted Onestop Digital for a return.  They urged me to test the lens properly and in fairness they perhaps do get some misguided returns.  They did annoy me a bit by suggesting that Samyang's optics were not as good as other brands because they are cheap...almost "well, what do you expect!"  Many test sources on line show this model of lens is not rubbish but should be a high performer.  OneStop's return policies were not reassuring, stating that if they disagreed with my findings then they could charge me all the postage costs and an inspection fee.  The cynic in me thought it is bound to be considered as "within specs" as there is no advantage to them in considering otherwise.


Anyway, I insisted on a return and this was efficiently handled by using their courier account.  I contacted them again and was not surprised to hear that they could not find an issue!  I quote....


"Thank you very much for your email.


We tried the lens and we could not find any problems with it. We also sent it to Samyang and they also didn’t see any problems with the lens.


We also compare your returned lens with another lens of the same model, and they more or less gave the same results.


Would you like a refund on this?"


I remain doubtful that it ever went near a Samyang technical or service department and was reported on with 3 days of arriving back in HK, but I cannot prove or disprove this (you can make up your own minds).  Its a free country (here anyway!), so they have the right to disagree with me about the lens,  but I did tell them I found it hard that they were unable to see such a clear issue.


The offer of a refund was welcome, but it showed that they did not want a customer who was perhaps more discerning than average and was prepared to question.  Of all the dozens of lenses for Nikon and Fuji I have bought over the years, this is only the 3rd one I have returned due to lower than expected performance and let's face it, duff lenses do exist, so I do not think I am being awkward.  These Hong Kong dealers have a business model to sell high volumes internationally and make a small profit.  They want to sell, ship and be done.  If anything messes up, their profit margin goes, so I can see this was a pain in the arse for them as well as me.  I guess its nice to have a choice as to the price/service equation and we have to make our choices in that regard.  If you ordered a lens or camera and got a fine sample then you would be very happy, but service is not going to be like your local shop if you have to return an item.  I was not enamoured by some of the comments in their emails, but in the end they did treat me fairly and within a day of asking for a return my Paypal account was credited with the full amount.  I would have created merry hell had they tried to deduct postage or inspection fees but they did the right thing.


So, off to find a UK supplier, here I go!


Update as of 12/02/2015....I have just taken delivery of another sample of the Samyang 8mm f2.8 and initial quick tests show it to be vastly better than the first copy.  The image quality seems pretty remarkable across the frame for a fisheye lens.  I feel sorry for whoever gets my first sample as presumably it will be resold, seeing as the suppliers could not identify the optical issues I out and check your lenses as soon as you take delivery!


I’ve recently bought a pre owned example of the same lens, albeit in its previous incarnation ( by the way, I understand from the comments on other blogs, that the “ improvements” of the new ones concern mostly the people who use this on a full frame camera rather than smaller formats such as the aps-c).

I had read of a sharpness discrepancy, such as you describe, before in an article written, I believe, by Ian Norman, someone who used this lens especially for astronomic photography.

His findings and comments are on

"The first lens that I received was a bad copy. It showed a distinct decentering effect which caused the whole left side of the frame to be noticeably blurry compared to the rest of the image, particularly when used wide open at f/2.8. ..."

I have to say that I haven’t tested my lens under such very strict and rigid conditions but my copy doesn’t seem to show the problems that you both have found. I suppose this is a hit and miss and that the assembling of this lens must be particularly critical.

I think that the microscopic misalignment of some of the many lenses which make up this lens, could, and indeed apparently do, result into a certain and uneven sharpness.

Possibly this escapes whatever quality control tests are carried out on the lens. But, I wonder if all of this is only the lens responsibility.

I also own a 10-24. That lens is spectacular although, under certain circumstances, it too might show some results which are, shall we say less than pleasing.

I did some testing with this lens while I had a X-M1 on loan from the shop because my X-E1 was being repaired. Now I have a X-T1, it’s a very long story.

I haven’t kept those pictures so I cannot show you them but I shot, at 10mm, a light house. The vertical frame at the top showed the top of the lighthouse at the very edge of the frame. There there were some wires and aerials and the definition of those fine detail was not very good. The rest of the picture was fine. I was pretty shocked. After all, at the time this was one of the most expensive lenses made by Fuji and the most expensive lens in my kit!

However this never showed again in any other picture that I took with it . So, I wonder if a loss of detail could have been determined by the sensor and only under certain light conditions. This could also explain why Samyang couldn’t find anything wrong with their testing equipment. Those are probably working with lasers and the don’t go out and shoot pictures with the lens to test it, as you and I indeed do.

Anyway, as yet, my 8mm hasn’t seen all that much use and time will tell. Whatever little I’ve shot with it already makes me feel that my limited investment was money well spent.

Disclaimer: I am not trying to excuse Samyang, and I am not affiliated to either Samyang or OneStop Digital.
No comments posted.

January February March April May June July August (5) September (1) October (1) November December
January February (1) March April (2) May June July (1) August (1) September October November December
January February March April May (4) June (1) July (1) August (2) September October November (2) December (1)
January February March April May (2) June July August September (1) October (1) November December
January February March April May June July August September October November December