Beware tilt adaptors

[David Kilpatrick]

I'm now on my third tilt-only adaptor for NEX. The first, a Kipon, was very expensive and has no reference markings or definite axis, making it rather hard to be sure of the setting although the lens can be locked in position once set. The second, a Russian one provided with a 50mm lens, has markings from 0-8 to cover 8 degrees of tilt so a setting can be learned and repeated (for example, it's pretty easy to run tests to find out what setting you need for a landscape when standing with the horizon in a certain position, and just dial this in).

My third, a Minolta MD fit Roxsen Chinese adaptor, has the scale engraved from 1 to 8 (WRONG) and also engraved backwards so that 8 is set when the lens is at zero - and looking on eBay, it seems that every single variation of this type of adaptor has the same stupid marking error as they have all copied each other. So 8 degrees of tilt means setting 1 and no tilt means setting 8. There are only seven divisions unlike the Russian which has one division for each degree and is the right way round.

But both these have no zero lock, or any kind of position lock; also any tilt adaptor needs to be made perhaps 1mm thinner than a normal adaptor, and I'm finding even the normal adaptors are being made generally too thick by perhaps 0.2mm meaning too many lenses will never focus on infinity. Tilt adaptors need to focus through infinity as they tilt the lens from a point behind its optical centre and therefore tend to move it forward, and very few lenses have enough focusing range beyong infinity to cope with this.

All very frustrating as there's no reason why these products should not be designed properly.

David

[alphaomega]

Reference statement

Tilt adaptors need to focus through infinity as they tilt the lens from a point behind its optical centre and therefore tend to move it forward, and very few lenses have enough focusing range beyong infinity to cope with this.

Q1 Is it not possible to compensate for this by using an appropriate F stop?
Q2 Despite these faults, can one actually get some benefit out of using one of these adapters i.e. over and above manipulation in LR or PS?

[David Kilpatrick]

Depending on the lens design, they work well enough and the lens I want to use (17mm f/3.5 Tokina) seems to be able to hit infinity/landscape settings OK on the Roxsen/Pixco adaptor.

The resolution of the A7R is such that even f/16 or f/22 won't pull in adequate depth of field for a foreground/background composition (think of the typical rocks/sea shoreline shot or Cuillins with waterfall). Not at pixel level view. If the same file is knocked down to 12 megapixels or 9 megapixels, the problem disappears - what can look soft becomes sharp enough.

I'm looking at this from the Alamy sngle. 9 megapixels is the minimum size (2400 x 3600) and at the moment I regularly reduce 16, 20 or 24 megapixel images to this size whenever noise levels, movement, or depth of field problems could be visible. Others I leave as large as I can. There's not a lot of point in shooting with the A7R unless you can keep the file maximum size, which is four times the area of the minimum, and twice the linear scale. Depth of field has always been calculated on the basis of a 10x8 print viewed at one foot at 50 cycles per inch (1/100th" as the smallest detail). In practice we tend to view prints at a little more than one foot, and often at larger than 10 x 8 - A4 or A3 - but this still holds approximately true. The A7R produces an image four times this size, and needs twice the resolution, the assumption being that you are viewing a 20 x 16 print at one foot - very close! Alamy's inspection is equal to viewing a 40 x 60 inch print at one foot.

Bsically f/16 is giving you the same depth of field as f/8 would be for a 12 megapixel full frame sensor or f5.6 for an APS-C, or for example an OM-D at 16 megapixels MFT. It's a bit of shock to the system! Having worked with medium format digital, I know that it's the pixel count which really makes demands on your technique, not the format. A Hasselblad 16 megapixel back is less demanding than the A7R - they were actually 37mm square imagers, so when you use the A7R it's like making a 35mm shape crop from the Hasselblad, but at three times the resolution density.

The solution is to use a tilt lens and also to stop the lens well down - both.

David

[classiccameras]

Probably the best period of my short history with APS-C DSLR's and the most satisfying IQ and DOF was with my old 12MP Canon 450D and a 12MP Nikon D5000. For me 12MP was the sweet spot.

[bakubo]

I have never done focus stacking, but I saw an article in Outdoor Photographer magazine about it while my wife was buying some groceries this morning. I just skimmed though it. For many situations it seems like it may be the way to go. Only works if the subject isn't moving, of course. It has pretty much the same issues as multiple exposure HDRs. But, if the subject isn't moving and you are on a tripod then it seems to solve several problems. You then have to combine the exposures using stacking software. It seems that many of the exposure time techniques used in the film days (graduated ND filters, tilt/shift lenses, etc.) while still usable are being supplanted by post exposure techniques on the computer. I am not making a recommendation here since I have not tried these things, just a suggestion of something to possibly investigate.