Sigma's new HSM 12-24mm DG II

[edrice]

agorabasta,

Thanks for raising my awareness of this issue. I realize I've seen these distractions many times before but never really properly identified them for what they were. I guess because it really didn't matter so much. I had become vaguely aware of Cornerfix from other readings but never pursued it because most of these little programs seemed like they add extra steps for what could already be accomplished in Photoshop itself, so over the years my collection of little uilities has dwindled. I read over the Cornerfix documentation and just having to first convert to a DNG is not without extra effort and then additionally havng to build profiles for each lens seems even more so. I'd rather be spending my time with creative aspects.

I weigh all this against just adding a selective color layer and in less than two minutes having colors tuned to what I want and this is part of my standard workflow anyway. No muss, no fuss. So I prefer the manual method and can probably fine tune it to what I want better than a Cornerfix profile could anyway. Besides I like the more creative aspect of manual control. Probably no lens is going to give me the exact colors I want all the time and I try not to let any lens get in the way of the result - which is why for me, the 16mm is a great little lens.

I'll take a further look at Cornerfix but most likely will continue the current path.

Ed

[agorabasta]

While there can't be any problem with a subjective approach, there's still some basic sanity in having a 'standard' calibrated response from the nearly-perfect image-capturing systems now so easily available to us.

Sure we can swing the colours to our likes, e.g. turn evening into morning (see below); but there is a much better chance doing it the best possible way if the original capture is first brought to some standard calibrated response beforehand.

So if there is a simple automatic proc doing exactly the normalising procedure to get us that standard starting point, it must be a good thing. The bad thing is that the correction doesn't come free of some losses.

[agorabasta]

Some good news. Just checked 16mm with the UWA converter on Nex7 - no purple shift to talk about, merely a few percent lower green wrt red/blue in the corners, essentially unnoticeable.
Corner smear is there as expected, need to crop by 1.25 to cut the smear off. So altogether it makes a good enough 15mm f/3.5 APSC at 15Mp res, but noise is also 0.7EV stronger due to crop.

[edrice]

Some good news. Just checked 16mm with the UWA converter on Nex7 - no purple shift to talk about, merely a few percent lower green wrt red/blue in the corners, essentially unnoticeable.
Corner smear is there as expected, need to crop by 1.25 to cut the smear off. So altogether it makes a good enough 15mm f/3.5 APSC at 15Mp res, but noise is also 0.7EV stronger due to crop.

Good work! I haven't had a chance to get back out with mine to see the effects, but that's good to hear.

Ed

[agorabasta]

Here's a sample 16mm+UWA image taken at f/2.8 with Nex7; first the full image downsized, then centre/focus and the right lower corner 100% crops.
All sky colours are real - no shifts as indicated by the colour of the car in the corner. All blur appears fairly isotropic; no radial blur directionality observed here, quite unlike the bare 16mm SEL results.

Processing 'to taste' in Lr4.1 RC2, no drastic measures applied at all.