Yesterday I re-modified the Pentax *ist DL to get rid of some dust that was between the sensor and the low-pass filter, so today I set up the AF and did some real-world testing at Portchester Castle near Portsmouth in between heavy rain showers.

It must be said that this isn't the preferred weather for IR photography but the clouds were nice!
















All at 10mm on the Sigma 10-20, except the church at 18mm on the kit 18-55 and the distant shot of Portsmouth at 55mm.

The Giotto seems very portable to me - it is solid, light and the ball head is easy to cinch up to be movable without locking. Build quality is excellent, but it isn't cheap.

I've also used one leg of it as a monopod a few times, it takes moments to convert so that's a plus point compared to tripods that can't do this.

I've been a bit busy lately, but the recent fine weather has been good for IR so I just had to get out there.

Trees 1




Trees 2




Watching the Sea 1




Watching the Sea 2



All modified K20D, Sigma 10-20 @ 10mm.
First two are on top of St Catherines Hill near Winchester, 1/45th at f/6.7
Second two are at Mudeford on Sea, 1/60th at f/8.

A mixed blessing though - debug mode isn't yet available on the K-5 and K-r, and you'd never get an adjustment of +70 via the normal adjustment menu.

What I meant was whether the lens axis was truly perpendicular to the chart. The plane of sharp focus is only a few mm out across the chart, which could be accounted for by the lens being a few mm off the correct axis. The idea of adding a little bit of error is to see what the effect is - maybe it will look perfect with the lens moved 5mm to the right and re-aimed for example. Or not...

I'm just thinking that the difference between sides is only a few mm, so it might be instructive to deliberately skew the alignment to see what happens. If it changes a lot then it suggests the alignment is very sensitive.

It looks to me like the K10D one is showing the same general effect though, albeit a little bit back-focussed compared to the K5 one.

That would implicate the lens - or the alignment of the chart

Edited to add - try moving the tripod a centimetre or so to the right, centre the image and shoot it again to see whether it looks the same.

Some lenses can be very slightly off-centre as well, or it could even be a combination of that and the physical adjustment of the sensor inside the body. The sensor assembly is held by the anti-shake, and the whole enclosing assembly is on three adjusting screws. Out of the three K20D's I've converted to IR the position of those three adjusters has varied quite a bit, between bodies and between the three on a single body. This suggests that the sensor is normally adjusted to be parallel to the lens flange, but perhaps your test is showing some small residual misalignment.

I doubt you'd notice it in real pictures.

This thread on the other forum might be relevant: http://www.pentaxforums.com/forums/pentax-dslr-discussion/159611-k10d-gx10-ccd-p...

Um, the sensor isn't absolutely square to the chart. Very hard to achieve without lots of fiddling about.

It sounds interesting. I assume that's from a fairly severe drop. Let me know by PM what you want for it, I'll pay postage. I'm interested in how it differs from the K-m and K-r inside.