[clackers]

I'm much too polite - it's a failing. I can never ask for money back after I've loaned it to a friend and they forget to return it. Isn't that tricky? The most I can do whenever I'm over at their home is to break something of that approximate value.

Thinking about where I was going in life, I went to a book store and asked the saleswoman where the Self Help section was.

She said if she told me it would defeat the purpose.

So, this week's tip is about buying a lens. New always carries the risk that one or more of the elements inside left the production line not exactly aligned, and the effect will be softness on one side of the image and focusing problems.

Secondhand lenses may have had an owner who never recognised it was what we call a 'poor copy' which is now out of warranty, and the bonus possibility that they may have dropped it at some stage and introduced some decentreing. Be suspicious of dents.

We pay a lot of money for lenses and we're entitled by law to keep sending 'em back until we get one that's up to spec. There was a PF member who on his third copy got a Sigma 100-300mm f4 that was as sharp as it was designed to be.

Obviously, the companies that sell the cheapest lenses have the worst manufacturing tolerances and do the least amount of quality control checking, but as Lens Rentals will tell you, there are defects and decentreing in any lens of any brand - even Canon L glass.

Since it seems to be a deliberate strategy to make the customer a defacto QC tester, we need to check as soon as we receive our purchase, and keep all the packaging in case we need to box it back up again and mail it off. I will not buy from a vendor that does not accept returns.

We probably don't have a laser collimater lying around the house to check our lens. You can print sharpness tests and set up the camera not far away from them, but there is enormous pressure to make the setup perfectly perpendicular or the results will be misleading.

There is a quick and dirty test we can all do before doing anything more precise, as our esteemed member @Beholder3 writes. The gist is to take four shots of a relatively distant object, with it in each of the corners, and compare them on your computer afterwards. A lens with perfectly aligned elements will have them all with the same sharpness.

You can confirm a fault by taking a picture of a flat detailed scene rotating the camera 90 degrees between shots, and you should see the soft side 'follow' your rotation in the resulting pictures.

How to Check Your Lens for Decentering - Articles and Tips | PentaxForums.com

To finish with there's the story of how after months of plucking up courage, a guy decides to take a parachute jump. But after leaping out of the tiny Cessna, he pulls the ripcord - and nothing happens.

Alarmed, he pulls his reserve chute cord - and again, nothing happens.

As he's plummeting towards the earth, he spots another man shooting upwards at tremendous speed.

'Do you know anything about parachutes?' cries our guy, as the man passes him.

'No,' is the yelled reply, 'Do you know anything about gas cookers?'

The rest of the series here: Clackers' Beginners Tips (Collected) - PentaxForums.com


Image Attrib: Tamasflex, CC BY-SA 3.0 <https://creativecommons.org/licenses/by-sa/3.0>, via Wikimedia Commons

[MossyRocks]

Another great way to see if a lens is decentered is to shoot a starfield with it. Those point sources of light will show the problem rapidly.

[gatorguy]

Keep 'em comin' Ian.

[clackers]

Keep 'em comin' Ian.

Thanks, mate. PJV and you guys have been good with suggestions about forum format. I might get to thirty and get bored, but until then ...

I post them originally in the Pentaxians Facebook weekly (I hope we can pick up some new members from there by me sometimes linking to here such as in the above article) but of course being social media they really are gone and forgotten within hours.

---

Last edited by clackers; 04-22-2022 at 05:29 PM.

[clackers]

Another great way to see if a lens is decentered is to shoot a starfield with it. Those point sources of light will show the problem rapidly.

Yeah, it'd be a good test bed, they're all effectively the same distance away so field of curvature should be insignificant. But there's uneveness due to rotation if not pointed to one of the poles, right, MR?

[Broadside]

I don't know what the forum's feelings on it are, but I'd suggest that a little variation in the corners (ie. a small amount of decentering) is to be expected in consumer lenses. Ones that have no variation (ie. are perfectly centred) are close to being unicorns. It's only a problem when the difference is significant and noticeable to you.

At the risk of starting a fight, I'd say I prefer a lens of sharpness 9 in the centre and corners 9-9-8-7 over a lens of sharpness 8 in the centre and corners 8-8-8-8.

[clackers]

I don't know what the forum's feelings on it are, but I'd suggest that a little variation in the corners (ie. a small amount of decentering) is to be expected in consumer lenses. Ones that have no variation (ie. are perfectly centred) are close to being unicorns. It's only a problem when the difference is significant and noticeable to you.

At the risk of starting a fight, I'd say I prefer a lens of sharpness 9 in the centre and corners 9-9-8-7 over a lens of sharpness 8 in the centre and corners 8-8-8-8.

Yes, I'm sure that's the big companies' justification for less QR, Broadside.

They will say, let the particular customer decide whether the copy's good enough for them. They may not even notice that it's 8-8-6-4 by your system, that the microadjustment has to be +8 on the body, and accept that focus at short range is hit and miss.

The company may even be able to send an imperfect returned product back on to the sales shelf, hoping the next customer's standards aren't as high.

Of course the strategy is financial. The higher the QC, the cost is passed onto all buyers. Refunding or exchanging 6% of lenses sold is cheaper than carefully testing 100% at the factory. *Collectively*, we buyers benefit in dollar terms, it's just a shame about unlucky individuals. It was inconvenient for me to return a Samyang 14mm to place of purchase, but that's the deal.

Even Zeiss do this. AFAIK, only their cine lenses are manufactured in Germany. Everything else, including those $5000 Otuses, is outsourced to the plants and workforces of Cosina in Japan.

---

Last edited by clackers; 04-23-2022 at 05:35 PM.

[MossyRocks]

Yeah, it'd be a good test bed, they're all effectively the same distance away so field of curvature should be insignificant. But there's uneveness due to rotation if not pointed to one of the poles, right, MR?

For wides and ultrawides that shouldn't be a problem since one can use 200/focal length for your shutter speed and get good enough results for seeing if it is decentered. With longer lenses decentering is less of a problem but even a 50mm lens you can do 4s exposures and easily tell if there is a decentering problem. Also decentering with astro shots looks more or less like coma but only happens on one side or corner and not the others so even with trails one should still be able to tell since things won't look uniform.