A week plus with the SD1 (long.....very long)

Started Jul 7, 2011 | Discussions thread
NarrBL Veteran Member • Posts: 6,135
Re: I believe this image is actually much better

Ok, Erik, and thanks for that. From your precision and reading through the whole subthread, plus some observation in a different viewing circumstance, I think all this can draw together happily.

  • first, taking your last, I see enough where you thought I was reacting harshly myself, and thus defended. Not quite the actual case; it was just that I had finally found a hard indication to explain what surprised me also about Rick's images as well as some number of others posted - the color profile issue (which has numerous twists and turns and also per-image-differing results).

And yes, i did presume everyone would be viewing that way; as much because there have been such high complaints across the forum as for any other reason. It seemed important to get that clear situation out, and in fact I picked Managarm's commentary as a nice place to do that. But in fact, he was using or could use Firefox, and his later comments that I am about to answer shows this wasn't probably the root of the things he was noting as less than optimal. Agree.

  • further, now that I have a little peace and an evening room where my laptop display works its best, I do see noise, very likely what both of you are observing, the green random pixel hash. It would literally hide in more intense background lighting; just something I am going to have to look out for.

This is very peculiar noise, I must say. It isn't often in truly dark areas at all. It does like to show some real strength in certain mid-scale areas, and also in certain transitional ones like the cloud-fog boundaries on the mountain. It seems to like certain colors to set it off, at a guess suggested in what's seen. To say the strangest part again, the noise happily appears in some mid-tone areas, and not in dark or light, which could also throw you off in looking for it. Once seen, though, it is there.

Furthermore, when you go to look, magnifying for better attention at 200% but with no saturation intensification, you also find something else: the small ten or 20 pixel dimension bounded areas just turned to monochrome grey. Leaves in the foliage framing the house at certain angles just turn grey, for example. I recall now that someone recently posted an outtake of another subject showing a similar effect; a few unlit light bulbs in a string just turned to grey masks in the same way. Very strange.

Here's yet one more strangeness. We're looking at an image indicating it was taken at ISO 100. Yet one taken at ISO 400 by the same camera doesn't seem to show this green pixel noise at all, even boosting saturation up to the color bloom point. That's http://www.pbase.com/rickdecker/image/136176846 .

Is noise reduction on in some shot processings and not in others? Is noise reduction responsible for the conversions of some small image features to grey masks? Is it rather something about blotch reduction? Too many good questions to ask.

I think to stop on this and see what comes out of Rick's reprocessing, setting proper colorspace and then converting, using the raw headroom to retrieve blown highlights, etc., setting sharpening to -2, etc.. This will improve all the viewing, and then maybe we'll see something to work with further on noise and-or masking issues.

In the end, then, I think we're arriving at a good investigation and discussion.


p.s. the reason for having made the mistake in assessing pbase images lower than full size: if you bring an image link over to another browser, it shows at whatever scale that browser might have used last at pbase -- in this case, wasn't Original size.

Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow