A Hilltop Ruin

[Dave Humphries]

Hi Donald,

Do we know if the concentric circles are placed dead centre of the captured image (before cropping)?

I ask because I am trying to think how or why DxO would apparently use something like that and the most logical idea that comes to mind would be the CA removal, since the amount of that applied would increase axial'ly from the lens (uncropped image) centre.

If the effect isn't image centred, then I'm at a loss too.

Either way, it doesn't really help beyond 'understanding' the precise cause - it doesn't solve the problem and short of using DPP instead of DxO, won't help.

Hope we get to the bottom of it soon,

[Dave Humphries]

I was just looking at your website, and wonder what would happen if you tried SEP2 on "A wall and 2 trees"

Wendy

Good choice Wendy

[ScoutR]

Donald, have you figured this out yet?

Wendy

[Donald]

Donald, have you figured this out yet?

Wendy and Dave.

Given all that's been going on in the last 72 hours, including the fact that I was working all day Sunday, I haven't got back to this since problem since Saturday.

But, Dave - you raise another fascinating point. The effect is image centred and seems not to be in any way linked to where I place SEP2 control points. The fact is that I crop in DxO. So what I'm exporting to to then be used is the cropped image on which the artefact is centred.

This is a bit like a medical diagnosis. You go through and eliminate all the options and, hopefully, you end up with what the problem must be. I'm away to try further re-processing on this and the image suggested by Wendy, to see what does (or doesn't) happen dependent on settings. Watch this space.

[Donald]

Still struggling.

Compare the following, done at Wendy's suggestion.

The first is with the RAW processed in DxO. But please note - every auto correction option was switched off - CA, Noise, Distortion, etc etc etc. So waht you've got, so far as I understand the science, is the most basic RAW processing. I have then cranked up the sliders in SEP2 in order to reveal the problem.

The second one is processed using Canon's own DPP and, again the sliders have been similarly cranked up to see 'inside' the image.

I'm now going down the DxO support team route. But am not hopeful given the lack of response from them that others report.

So, to save my sanity, would you agree that this has got to be a problem at the RAW processing with DxO stage of the process? But why have I, or you, never seen it before?

(Click on them to view in the lightbox)

ps - This is the full frame.

[Colin Southern]

Hi Donald,

I've only 1/2 followed this, but I'm assuming that we've still talking about the concentric circles?

If so - are we sure that's it's DxO doing it? Have we checked the RAW file straight from the camera? Have you tried converting to DNG via something like Adobe's DNG converter? Have you examined the metadata to see what DxO is adding?

PS: Looks like your sensor is a bit overdue for a clean!

[Donald]

Hi Donald,

I've only 1/2 followed this, but I'm assuming that we've still talking about the concentric circles?

If so - are we sure that's it's DxO doing it? Have we checked the RAW file straight from the camera? Have you tried converting to DNG via something like Adobe's DNG converter? Have you examined the metadata to see what DxO is adding?

PS: Looks like your sensor is a bit overdue for a clean!

Colin

Yes, we're still talking circles. Can you see them in the first of the two images above?

I'm pretty sure it's a DxO problem.

Dave checked the RAW file of another image (The Forth Bridge) and, like me, couldn't see anything in the RAW file. The RAW for the above winter scene is here. I'd be grateful. if you were able to look at that to see if you can see anything.

After my recent problems with the computer, my copy of Elements 7 decided to stop working. I'm goign to try and re-load it, so that I can then convert from DxO as a DNG instead of a TIFF, to see if that makes any difference, because I certainly wasn't seeing it then. (Note - When using SEP2 as a stand alone, you have to give it a TIFF. Can't give it a DNG).

I haven't checked the metadata to see what DxO is doing, but as I wrote above, I stripped out all the auto corrections on the above.

I'll report further, but any other thoughts would be welcomed.

ps - That one was taken before my recent sensor clean, which also serves to tell us that, thankfully, Dave's question about whether some damage might mave been caused to the anti-aliasing filter could have been the problem.

[Colin Southern]

Hi Donald,

Yes I could see them in the 1st of the two image, and no, I couldn't see them in the RAW file. Sounds like DxO might be approaching it's use by date

Hope you don't mind, but I couldn't resist having a quick play with the file ...

[Donald]

Colin

I know you are a passionate supporter to Adobe products and wonder why some of the rest of us don't use them all the time.But the fact is I bought DxO and for a long time now it's been a super tool that I have been very, very pleased with.

Because of that, and for financial reasons that mean I can't afford to just ditch one thing and buy another, I need to continue to work to try and find a solution to this. That's why I'm going at this one so much and am asking for help from anyone able to give it.

[arith]

Of course, it's beautiful, Donald!

Erm, there's something curious. There seems to be a target in the sky - a round circle with radiating concentric circles.... It's making me have a hard time seeing the picture. Is it because of my silly, falling off monitor?

Nope can't see it; I did see something odd when I stuck a blue filter on one of my church photo's, some devilish faces: turned out to be just patterns in the ironwork.

A bit cornish; need some fog and .... memory has gone. Heathcliff

[Colin Southern]

Colin

I know you are a passionate supporter to Adobe products and wonder why some of the rest of us don't use them all the time.But the fact is I bought DxO and for a long time now it's been a super tool that I have been very, very pleased with.

Because of that, and for financial reasons that mean I can't afford to just ditch one thing and buy another, I need to continue to work to try and find a solution to this. That's why I'm going at this one so much and am asking for help from anyone able to give it.

And good on you for doing it! (I was just kidding about approaching it's use by date by the way)

[arith]

Is it possible that there is nothing really wrong with SEP 2 even if you use some control points on your image? My wild guess is that the adjustments available in GIMP might probably be amplifying the effect of that control points you used while inside SEP 2 and so making those concentric circle anomaly slightly visible. Normally, what I do is to use SEP then apply a mask in photoshop rather than use the control point of SEP. Not a good process but works for my taste. Apologies if I have offended you by asking about GIMP.

This is a screen capture of a control points' mask layer showing the concentric circles. This is not the actual image, this is only the masked layer showing the extent as to how far the control points' adjustment will affect the image.

I know what that is cos I use cheap software, Colin knows what it is cos he told me what it was ages ago.

[ScoutR]

Well, I can't help much, but I will report that I see the circles clearly in the first image and not in the second. However, in the second image rather than circles there seem to be vertical traces in the same areas as the circles - sort of like if you picked a point in the middle and stretched it out on both sides then the circles would be there.
These are extremely processed to show the effect, so I don't think with your normal processing it would have been noticed. I would tend to agree you are on the right track as to it being DXO related and then whatever you are doing in SEP2 has just emphasized the effect and brought it to everyone's attention.

Now I don't know how software works, but isn't it possible that there are all sorts of things lurking in these pixels and it is just a matter of noticing it and then processing accordingly, which in this case might mean spending more time with SEP2 so you don't bring the artifact out - nothing was noticeable until you started using SEP2 and the quality and look of your images is always unique and exceptional. I wouldn't want you to loose that just to get rid of circles that are only visible when processed to the extreme in order to reveal them.

Another thing I am wondering is if there is/was any kind of vignette applied to these. The reason I ask is because from what I have seen with my own experiments (in Lightroom anyway) is that the vignette slider does seem to work from the centre in a circular pattern, and if the pixels had been arranged in this pattern and then processed in a certain way it might reveal that pattern that was created by the vignette. You know where I am at with PP and such, so I'm sure you understand that I'm just speculating and kind of throwing things out there just in case - This is not the voice of science talking

I would hate to see you throw the baby out with the bathwater and end up ruining the look of your images, so although I hope you find the cause, I really hope you don't change your workflow which will inevitably change the look of your images

Wendy

[Donald]

I would hate to see you throw the baby out with the bathwater and end up ruining the look of your images, so although I hope you find the cause, I really hope you don't change your workflow which will inevitably change the look of your images

Quite honestly, Wendy, that's what's scaring the hell out of me at the moment. I'm doing 'my thing' and suddenly I seem blighted by this problem. I'm now seeing circles everywhere. Will probably start having nightmares about them soon.

There are no vignettes on these images.

DxO Support has responded very quickly to me and the RAW and Sidecar file for 'A Wall and Two Trees' has been uploaded to their server, at their request, for their technical people to examine it. So, we'll see where we go from here.

I downloaded Elements 9 to use as a trial, just to see what would happen with images processed in ACR. Would it leave the artefact or not. The answer? No. So, it's got to be a DxO thing ... or something in the interface between DxO and SEP2.

The one other thing I'm wanting to try is to get SEP2 working as a plug-in to this trial version of Elements 9 I've downloaded. That's how it is meant to work. At the moment because of the problems I had a few weeks ago, Elements 7 stopped working. So, I've been using SEP2 as a stand alone. The problem is that, as a stand alone it can't open a Canon .cr2 or a DNG file, only a TIFF or a JPEG.

Don't know why I can't see SEP2 as a filter in Elements 9 the same way as I did with Elements 7 (is it because I've only got 9 as a trial?).

Why am I saying this? Because I want to see if the same thing happens if I take a DNG into SEP2 as when I take a TIFF in. (As you can see, I'm now getting desperate for explanations!)

And finally, if I can't solve this problem, then I may have to consider using ACR. However, I understand that the ACR that comes with Elements is a much lesser being than that which is hooked into Lightroom and Photoshop. Can anyone tell me what the differences are?

Most of my current workflow is (has been) done at the RAW stage in DxO. That's what i want to continue with.

[TobyMac]

Really love this picture, the grass in the front somehow emphasizes the oldness of the ruins

[JohnC]

Donald-I usually like your shots in this Silver Efex Pro 2 format but to my eyes it seemed almost too much - the texture isn't the same as the previous ones I've liked so much-I do keep harping back to the photos taken in France. I don't know the technical term I should probably be using.

I'll look at the Forth Bridge- I've had to cut down on time on here (and reading my newspaper) due to eye strain- ordered new specs. today -that was the problem misreading text on here - tired eyes so I skipped through it with unfortunate results .

Ok..I really like the Forth bridge, it's a great shot,not anywhere near as grainy as this one but I saw that Steve thought the sky was. I also liked Hound Point too and yes, it was worth getting up early for and in your Efex format it portrays, very realistically, the 'usual' (sorry lol) drizzly grey weather you suffer so often up there - the reason we decided to stay here.

By the way re. painting the bridge this will interest you,maybe you know about it already.
http://www.bbc.co.uk/news/uk-scotlan...-fife-14789036

[ScoutR]

Don't know why I can't see SEP2 as a filter in Elements 9 the same way as I did with Elements 7 (is it because I've only got 9 as a trial?).

Donald: I have Elements 8 and I had the same issue with some Topaz Lab programs that I downloaded. Turned out the download put a certain file in the wrong location - something to do with 32/64 bit folders if I recall correctly. Anyway I found the information I needed in the Topaz Read Me file and ended up having to copy and paste the files they named from one folder to another. The instructions were very specific and everything worked like a charm when I got the files copied over. Hopefully you can find something in the SEP manual or on their site. It must have been a common problem with Topaz because they had specific instructions for Elements, Photoshop and a number of other programs.

Good Luck with all of this, I know you will get it worked out and I hope DXO comes up with some solid answers for you so you don't have to muddle around with all this different software.

Wendy

[arith]

I tried to simulate it, but it isn't easy, it just happens and sometimes you have to go back to the drawing board. This quick effort is nothing near as good as the original because it showing something and but you should be able to see some lines, although not circular, I believe they are special to the light in your case.
I think it is posturisation due to 8 bit manipulation similar to this; if you can see it.



Sorry; it is not so easy to see and I normally go 100% to check for posturisation. But above the green building and in blue areas of sky are lines.

[Donald]

Just to conclude on this thread and the issue of concentric circles.

I think the person with whom I communicated at DxO Support has been as helpful as he can be. Our exchange has ended with:-

"This is very interesting. While SEP2 may be doing the accenuation, it seems that this may in part be coming from what makes DxO's RAW conversion superior.

Fact is that DPP, ACR, and DxO do their RAW conversion differently. Unless seriously accentiated in a particular way, you do not see this phenomenon at all.

A very curious phenominon that does indeed need to be looked at. I thank you for making us aware of it."

It came with a footnote that there could be no promise made about if and when DxO boffins would actually look at it.

In my final reply, I said:

"I agree with your assessment of the situation. I will, hopefully, have achieved something constructive if you are in a position to draw it to the attention of others within DxO. As for me, I will continue to use DxO as I think it is a fantastic tool. I will just have to go lightly with sliders in Silver Efex Pro 2 when I have images similar to those we have looked at, when this phenomena is brought out.

Many thanks for your time and your help."

[lizzy310]

Wait a minute...let me just hold my breath, tilt the screen and squint...ROFL!!!

Sorry Donald, but I see circles too...
They become more prominent if I tilt the laptop screen away from me, and squint with one eye and hold my breath for about 10 seconds.

Seriously though I do see them, their centre is mid-width of the image and about half way down the sky...but still a great image and I do like the tone/texture bands of grasses and clouds. Nice...