The thought in the title of this post occurred while messing around with the Sigma SD10, trying to see what kinds of color casts occur at high ISO settings (not that I ever use high ISO, I'll hasten to add). It was cold outside in the Real World and it's so hard to assess the color of tree-trunks and twigs and other natural stuff. So, no prizes for guessing what I did. ISO 1600, 1/250 sec, well-lit, don't leave home without one:
It's a Sigma X3F (raw) file opened in ACR 5.4, NR'd to the max and sharpened a bit.
Quite frankly, that is an astoundingly good picture, garnered from the much-maligned early Foveon sensor which is not noted for it's high-ISO, long-exposure imaging quality. Now, if the shot had been taken in low light with a correspondingly long exposure, I'm sure that pure mush would have been the non-recoverable result! But we digress . . .
So, why is the image so good? It's pure software that does it. It makes one wonder, when reading debates about the "ISO performance" of this camera versus that, whether the contribution made by software is really appreciated by everybody. Or does it just go without saying?
I have another converter 'X3FWorkshop', written by Arvo Jagel, that gets right back to basics and applies no noise-reduction at all (even less than dcraw!). Opening the very same X3F file with X3FWorkshop is a bit of an eye-opener (same image, remember):
To paraphrase Crocodile Dundee: "Now that is noise . ."
A little off-topic, but the SD10 in-camera medium-size preview JPEG comes out quite well:
Not quite as "good" as ACR, though . . .