Re: [darktable-dev] Re: Another real-life example

hi, On Mon, Oct 24, 2016 at 9:46 PM, Alex wrote: > He complains about magenta shadows. [..] > The issue is obvious in these images: > http://i.imgur.com/0sixStb.jpg > http://i.imgur.com/12eQb4y.jpg yes, those images do show the mean shift when clamping to 0 before

Re: [darktable-dev] Re: Another real-life example

Le lundi 24 octobre 2016 à 23:04 +0300, Alex a écrit : > I see this as chroma noise. Same issue as in the images I've linked, > just a lot less obvious. > > I'm talking about https://www.sendspace.com/file/uem3t6 here. Hey please! I have already acknowledged and said that in the original

Re: [darktable-dev] Re: Another real-life example

Le lundi 24 octobre 2016 à 22:46 +0300, Alex a écrit : > He complains about magenta shadows. Yes, but then: <<< The one you generated shows the surface of car body as covered with a little scratches (zoom to 1:1 and compare). >>> Those are luma noise. And I was responding to this (after

Re: [darktable-dev] Re: Another real-life example

Le lundi 24 octobre 2016 à 20:29 +0300, Alex a écrit : > > That's luminance noise and certainly not a problem to me. > > To me, it looks like chroma noise, but the defect is not really > obvious in this example. No please reread the exchange. At first there was chroma noise (I've explained how

Re: [darktable-dev] Re: Another real-life example

On Mon, Oct 24, 2016 at 08:29:38PM +0300, Alex wrote: > Zbigniew, > > Can you prepare a CR2 similar to this one? http://i.imgur.com/0sixStb.jpg Do you mean: to shoot something dark, on high ISO value - to get grainy result - and in similar lighting conditions? -- regards, Zbigniew

Re: [darktable-dev] Re: Another real-life example

Zbigniew, Can you prepare a CR2 similar to this one? http://i.imgur.com/0sixStb.jpg It can be the original CR2 from the same example, if you can find it or get it from OP. Or a similar scene that shows the same defect. I've got a theory about a possible fix, so I'd like to look into it.

Re: [darktable-dev] Re: Another real-life example

Zbigniew, > Anyway, I'm aware it's rather a matter of personal taste, but I'm > convinced > that after you examine both pictures again - and even more closely - > you'll > agree that this Silkypix-made is a little bit better. The one you > generated > shows the surface of car body as covered

Re: [darktable-dev] Re: Another real-life example

JLC> What I see: JLC> 1 - The exposure level is not the same on both the pictures - there is JLC> about 1 EV difference JLC> 2 - The silkypix.jpg is noise processed while the darktable.jpg is not. Most probably silkypix is denoising - I'm not opposed to that - what I was trying to explain is that

Re: [darktable-dev] Re: Another real-life example

J-L.C.> Is there something wrong as I don't SEE anything wrong P.O.> Same here. I don't see anything wrong. Encouraged by your answers just installed current version of Darktable - and what I can say, is... guys, I honestly can't believe you didn't see this:

[darktable-dev] Re: Another real-life example

OK, thanks - so it may be partial answer to my original question; since you don't see it using Darktable in "development" process, then probably its decoding procedures - at least the ones used to decode Canon's RAW files - are better than the ones in LR 5.7.1. You may want to examine my examples

Re: [darktable-dev] Re: Another real-life example

2016-10-24 10:44 GMT+02:00 Jean-Luc Coulon (f5ibh) : > I've downloaded your picture. > I've used the colour picker on the black car at the upper left. > I used the "zone" picker and drawn a square on the black car. > > Without pushing the exposure but with the base curve

[darktable-dev] Re: Another real-life example

I've downloaded your picture. I've used the colour picker on the black car at the upper left. I used the "zone" picker and drawn a square on the black car. Without pushing the exposure but with the base curve applied, I get: 80, 80, 78 With +1EV, I get: 118, 118, 115 Is there something wrong as