So

[David Kilpatrick]

I now have confirmation that I can attend the Sept 9 Euro press gig locally to me (but not the 'experience' they have laid on for the visitors - doesn't bother me, I have done these things before and it's no fun on home ground).

I have been doing some tests on the Sigma 24mm f1.8 as a potential lens for full frame. Bad news - it already gets very soft at the edges on APS-C even at f5.6 let alone wide open, I imagine it would be unusable on full frame. In fact lenses are my main concern with the full frame format.

David

[David Kilpatrick]

Well, here's an f5.6 carefully autofocused with the Alpha 350:


And here is a section from near the middle at 100 per cent:


Here is the extreme left at the same level:


and here is the extreme right:


Yes, it has amazing close-up abilities - to within the lens hood almost touching the glass:


Flower of Himalayan balsam on the river bank - at this macro distance, the loss of edge sharpness is not so important.

David

[PhotoTraveler]

Congrats on the invite. "Experience" eh, for some reason I just pictured that being everybody goes in a room, and a giant vat of orange (cinnabar) liquid dumps on you.

On the corners, just crop to 16x9, no more bad corners . I crop most shots to 16:9, so it definitely helps, unless your doing poor mans PC lens, in which 2 corners could still be an issue. Now if Sony was to make a 4/16-24PC lens, no problem, of course that isn't going to happen in this reality , But a 20mm PC lens would be ok, and then a 24mm TS. I expect the A900 will have 16x9 like the A700, that chops off 2mm top and bottom, which gets rid of a good amount of corner.

Now a question is what rumors will pan out on those 2 wides. Is the one a 1.4/24, or is it something more like a 18 or 20mm. If it's a 1.4/24, that might make the sigma irrelevant, unless you have cost constraints that is.

Do you expect the Minolta 2.8/24 not to do well? Or are you looking at the sigma because you need some speed in that range? I've never really understood the purpose of lenses like the canon 1.4/24L I suppose on FF/Film it can be nice for close up shots in dark nightclubs/concerts, but it just sounds like a lens of limited use that cost a lot.

I now have confirmation that I can attend the Sept 9 Euro press gig locally to me (but not the 'experience' they have laid on for the visitors - doesn't bother me, I have done these things before and it's no fun on home ground).

I have been doing some tests on the Sigma 24mm f1.8 as a potential lens for full frame. Bad news - it already gets very soft at the edges on APS-C even at f5.6 let alone wide open, I imagine it would be unusable on full frame. In fact lenses are my main concern with the full frame format.

David

[David Kilpatrick]

Yes, it's a unique lens, but I was hoping it might prove ideal for the Alpha 900. There are Alpha 900s out there! Sony staff are using them. I do not know what lenses they have. Keep your eyes open, the cameras are on the loose and I now know of three people (at the very least) from Sony who are using them for their own leisure shooting, taking them on holiday etc. But- with what lenses? Is there a 24-105mm new design out there yet? This I don't know.

I'll be taking the Sigma 24 to the launch on Sept 9th anyway.

David

[PhotoTraveler]

Take a 1.7/50 with you. Would be fun, if you get your hands on said camera to put on the littlest little guy in the whole system. Folks see the current uber-camera with the 80 dollar lens and maybe get some of the right folks to think about how they should make such a lens.

[Greg Beetham]

Well, based on Don's excellent results with the Sigma 24 there is now a question mark hanging over that one of yours David...or the camera, I wonder if it is possible for you to take a couple of shots with another body too see which it is, your particular lens, or the camera...sheesh mate the clouds in the background were as sharp as the scenery
Greg

[David Kilpatrick]

I would guess the Sigma has a fairly flat field at the distance Don shot, but a strongly curved field at near infinity (it focuses so close that there is bound to be closer than normal optimal point for field flatness). Also, the A350 is a bit more critical of lenses than the A700, and possible slightly less accurate for focusing. And distant trees are about the most difficult subject for any digital SLR.

David

[tbroadley]

I was thinking about lens that I have or do nor have that would work with the a900. What I came up with was the following:
Minolta 28/f2
Minolta 50/f1.7
Minolta/CZ 85mm/f1.4
Minolta 100mm/f2
Minolta/Sony 100/2.8 macro
CZ 135mm/f1.8
Minolta 200mm/2.8
Minolta 80-200/2.8
CZ 24-70/f2.8 zoom (don't have)

I'm not sure my Minolta 300/2.8 would hack it nor am I sure my beloved 28-135?f4-4.5 would also. In response to David comment on the 70-200/2.8 SSM, does the 80-200 have the same field curvature issue? For most of my style, the issue of distortion at the long end of the 70-300 will probably not be a problem. In any case, IF i'm close on the above, that will be a useful range for me. I particularly look forward to the 100/f2 and CZ 135/1.8 with the a900 especially for doing trains

cheers
tim

[Greg Beetham]

I would guess the Sigma has a fairly flat field at the distance Don shot, but a strongly curved field at near infinity (it focuses so close that there is bound to be closer than normal optimal point for field flatness). Also, the A350 is a bit more critical of lenses than the A700, and possible slightly less accurate for focusing. And distant trees are about the most difficult subject for any digital SLR.

David

I guess it's possible for a lens too have a shifting focus plane (fore and aft) depending on distance of the subject focus, I'll take your word for it, although you'd think lens designers would be aware of this possibility when designing the lens and minimising that to as close as nil as possible, looking at the photos it does look like the sensor is either slightly too close to the rear element (or slightly too far) and thus slightly defocussed, if the lens was producing a curved focussed field one would think there would be some parts of the image that was sharpish and some soft, say the center area better resolved and the corners soft (or vice versa), this looks universally soft all the way across, anyway it would be interesting too see a couple of similar shots with an A100 or your A700 David, too see if there is any difference.
Greg

[David Kilpatrick]

In response to David comment on the 70-200/2.8 SSM, does the 80-200 have the same field curvature issue?
tim

No, it's a better lens for close ups, although you can't get physically as close. It does not the same type of internal focusing. In several ways the old 80-200mm APO G/HS is a better lens than the 70-200mm SSM.

David

[Greg Beetham]

David,
I've been trying too take some decent shots of some trees, with not a lot of success (I have three lenses that can "do" 24mm) and I was going too see which, if any, could produce a half decent job of some foliage in the near distance, problem is, this is our windy part of the year Aug, Sep, so it's been difficult too get shots that one could say are representative of the lenses...so far.
Greg

[David Kilpatrick]

Foliage at a distance is the most demanding subject for digital, because the contrast in green foliage is determined by the red filtered pixels (if you want to check out why this is so, try shooting a landscape using a deep green filter on b/w film, then try it with an 8X red). Many people think that it is the green pixels (half the total in a Bayer pattern sensor) which decide the sharpness of green objects, but that is the wrong way round. Green pixels actually aid the sharpness and contrast of red or orange coloured objects most of all, which is why if you want to show off the resolution of a digital camera, nothing is better than a studio image of a bowl of fruit including oranges!

The red pixels in a Bayer array usually record the least sharp image because the AF system is set to work best with green/blue wavelengths (this is also why green/blue subjects focus badly - by allowing maximum sensitivity in the green band, the AF phase contrast detection system lowers discrimination between greens). Red is then normally focused behind the true focal plane, unless your lens is totally apochromatic - which even those lenses labelled APO are not. This means that the contrast element in green foliage (the shades of green emphasised in difference by the action of the red filters on the R pixels) is rendered slightly less sharply focused. The actual degree is TINY, but it's enough at pixel level to have some effect on the appearance of trees and leaves. It also gets worse the closer you get to infinity; most lenses can behave apochromatically at macro distances because of the great latitude in depth of focus and depth of field at 1:1.

A secondary factor is the fractal nature and random edge orientation of foliage shapes. Combined with the edge sharpening applied to any de-Bayer process, leaves will always appear softer than something geometric with 0, 90, 45 degree straight edges (etc). A third factor - as you point out - is that leaves are often moving while brick walls are not. And because the contrast of foliage is mainly determined by the R component, it's only one in four pixels, so foliage is really being rendered as a 3 megapixel image when you shoot 12 megapixels.

Technical? Yes. Intensely technical, very easy to understand if you have worked with lenses and filters, really hard to understand if you have not. The important thing to remember is that for any given colour of subject, it is complementary (opposite) filter colour which gives texture and contrast. Skin tones are successful in digital shots because it's mainly the green filters (half the total of the RGGB pattern) which give skin texture, depth, and variation in the fleshtone hue.

David