Due to a rather unfortunate discovery reported in a previous post I’ve had cause to drag out the GX20 and start playing with it again. (The Samsung GX20 that’d be.)
The mournfulness is occasioned by a number of things…
First there’s the whole “bugger, my fave cam’s fallen ill” sorta thing that’s left me annoyed and depressed in equal measure (though prob’ly a bit more depressed than annoyed).
And I haven’t yet reached the optimistic stage of assuming all can be made well again… I’m still in the doom & gloom (“aargh, I might as well just top meself now”) stage at the mo’.
Then there’s this huge “thing” I have about the GX20.
First, there’s those massive file sizes that cause my normal pic-processing machine to groan in agony and exhaustion. (Hmm… p’raps I should think about bunging in some more RAM. Damn. That’s another expense I don’t need!)
Despondency deepened by recollections of the colour issues I had when I first acquired the 20.
I think I’d just about got the issue resolved but its so long ago now I’m not entirely certain.
Of course, I could always press the Canon back into service. But the 400D vs the GX20? No contest far as I’m concerned.
So drag the beast out of storage and start having a play (not terribly enthusiastically, I have to admit).
“Might be an idea”, thinks I, “if I do a few sort of sky test shots” (cos, as I seem to recollect, ‘t’were sky colours that caused the greatest problem).
Of course, what I should have done was wait until full daylight (and hopefully a bright sunny day) but, once seized by the idea, well, it had to be then and there.
Which “then and there” was about half-six in the morning on what was shaping up to be a bit of an overcast day. In fact, it even began to spot with rain whilst I was out “doing my stuff”.
Anyway, the whole point of the session was a) to see what the colour rendition was like straight off-camera; and b) to process the same pics on different machines for comparison purposes, the thought being that maybe I could get away with using my more powerful machine (with a whopping 4Gb of RAM) to process the files rather than my customary machine with its pitiful 512Mb of RAM. And finally c), having reluctantly accepted that I’m likely gonna be stuck with using the GX20 for the immediately forseeable future, what’re the chances of being able to produce pics that are at least half-way acceptable? (That being the principal motivation of course.)
The real issue being the way different machines appear to render colours differently, seemingly no matter how much one tweaks the graphics properties.
I think I may have referred to this before (in fact, I know damn well I have) and finally come to the conclusion that all talk of colour consistency on the Web is a load of cobblers.
At the end of the day its all bloody subjective! Coloured by how the individual has their machine set up (yes, the pun’s intended).
Now my approach in processing these same pics on different machines was not to try to get them all looking identical but rather, focussing on one batch at a time, attempt to get them to reproduce the scene as closely as possible colour-wise as I remembered it.
I ran the test with three different machines, and what I actually ended up with was three completely different colour versions of the same scene. How totally bizarre. So which one’s right, if any? I’ll leave it to you to judge. Perhaps the only real question is, which one looks most “natural” (if any)?
Let’s start off then with a JPEG developed straight from the original RAW file. (I’d like to be able to say its completely unprocessed but I can’t cos it turned out that all the shots appeared somewhat underexposed (on “normal” camera settings so this has had one stop of exposure compensation applied. Oh, and… um… “straightened” a tad cos it seems I wasn’t holding the camera totally straight. But that’s the only “tinkering”).
So, this first one was output using the ACR (4.4 I think) profile…
and this identical one output using the embedded profile…
Now we come to the shots that I’ve messed around with (in each case principally tone adjustments, noise reduction and some sharpening) on the three different machines. One of these machines is supposedly “colour balanced”. Hmm.
More examples in the dedicated Flickr set.
(Tip: in the set they seem to display in a fairly random order, whereas in the photostream itself they actually display in a more organised fashion… mainly cos I edited the “upload times” to get them to do just that!)