Home

Please support mflenses.com if you need any graphic related work order it from us, click on above banner to order!
SearchSearch MemberlistMemberlist RegisterRegister ProfileProfile Log in to check your private messagesLog in to check your private messages Log inLog in

Millom and Black Combe
Post new topic   Reply to topic
Home   Digital Gallery MF
View previous topic :: View next topic  


PostPosted: Sun Jul 21, 2013 3:36 am    Post subject: Millom and Black Combe Reply with quote
Just before sunset I took a walk to the top of one of the hills in the middle of town with my NEX-3 and Konica Zoom-Hexanon UV 4/80-200.

Looking due north to Black Combe with Millom cricket club in the foreground.

HDR stitched from 16 HDR shots, a total of 96 exosures.



PostPosted: Sun Jul 21, 2013 7:27 am    Post subject: Reply with quote
Very nice view with lovely houses, which I think would be better without HDR, that creates noticeable halos.


PostPosted: Sun Jul 21, 2013 11:57 am    Post subject: Reply with quote
HDR give an extra look, do you have non hdr version too ?


PostPosted: Sun Jul 21, 2013 12:22 pm    Post subject: Reply with quote
I'm afraid I didn't make a non HDR version, it wasn't possible, this was as the sun was setting and the difference in brightness between the sky and the land was too great, you either had a blown out white sky if exposed for the land, or a nice sky with black silhouetted land if exposed for the sky. The sun was just to the left of frame, I cropped the picture after stitching to remove the flare on that side.

Hi Orio, where are the halos you see? Just so I know what to look for and can think about how to avoid them next time.


PostPosted: Sun Jul 21, 2013 12:41 pm    Post subject: Reply with quote
Lovely shot Ian, very nice! Smile


PostPosted: Sun Jul 21, 2013 1:00 pm    Post subject: Reply with quote
Cheers Ed, I'm trying to make the most of this unusual weather we're having.


PostPosted: Sun Jul 21, 2013 1:20 pm    Post subject: Reply with quote
Very nice, I can't see any obvious halos.. could just be a default response from someone that doesn't like HDR (in the same way some people make public statements about bokeh!). Or he's confusing haloing with the haze effect of layered mountain/hillside scenes.

It is, however, just a touch flat - which is a common thing with HDR when the shadows/highlights become averaged. I've found that layering one of the darker exposures over the top at a moderate opacity (around 40%) helps add a bit of depth back into an HDR scene. That might be tough though if you're workflow was tone map then stitch, rather than stitch and then tone map.

But otherwise, a pleasant composition and very restrained HDR to keep it naturalistic.


PostPosted: Sun Jul 21, 2013 2:04 pm    Post subject: Reply with quote
GeorgeSalt wrote:
could just be a default response from someone that doesn't like HDR (in the same way some people make public statements about bokeh!). Or he's confusing haloing with the haze effect of layered mountain/hillside scenes.


Or just say that I'm an idiot, it's simpler.


PostPosted: Sun Jul 21, 2013 2:19 pm    Post subject: Reply with quote
iangreenhalgh1 wrote:

Hi Orio, where are the halos you see? Just so I know what to look for and can think about how to avoid them next time.


Hi Ian, follow the line of the green Hills: behind it, there is a white halo over the bluish Hills behind. Haze appears at a level; when there is haze, you can see it at the same height level,
haze does not change altitude following the contour of the in-front hills as your bright line does.
You can also notice the reciprocal artifact, a darker halo following the upper contour on the green Hills on the left, and on the bluish Hills too.
Those halos are local contrast artifacts, they are produced when a large dark area is near a large bright area, they are easy to replicate in Lightroom (by pushing the Clarity slider to large amount),
and also in Photoshop with abundant use of local contrast enhancement (gaussian sharpening applied at very large radius).

Image created by applying Photoshop's Smart Sharpen, Radius 64px, Amount 500, Gaussian mode, to a simple White/black bicolour original:



PostPosted: Sun Jul 21, 2013 2:46 pm    Post subject: Reply with quote
Thanks for the info guys, most useful. I try to make my HDR realistic if possible, I see it as a tool to use when the light conditions are extreme, usually this means bad light, with the typical bad weather here, but it is also useful for situations where there is too much contrast due to low sun angle late in the day, which is the case here.

I understand what George meant, I've run into a fair bit of anti-HDR prejudice, but Orio isn't one of those Luddite types, so it's all good. Smile

Orio, thanks for the explanation, it makes sense to me but I'm still not sure I can see it, so I cropped out the hills:

Left edge, centre then right edge of the frame:





BTW I posted this picture on Facebook and it's been one of my most popular images, people like it because it's not so realistic and looks a bit like a painting. I've taken note of that, in case I want to make some HDRs for commercial purposes, seems to many 'punters' the less realistic is appealing.


PostPosted: Sun Jul 21, 2013 3:01 pm    Post subject: Reply with quote
Orio wrote:
GeorgeSalt wrote:
could just be a default response from someone that doesn't like HDR (in the same way some people make public statements about bokeh!). Or he's confusing haloing with the haze effect of layered mountain/hillside scenes.


Or just say that I'm an idiot, it's simpler.


Perhaps I am.. because I still can't see it even with your description - I can't see any HDR artefact halo that distracts from the image. Although I can see a sensor spot repeating itself four times across the sky that would be easy to clone out.

I have the same faint haze line effect following the line of hills in a layered landscape like this in HDR and non-HDR images of similar scenes, so it's not going to be something that's fixed by unwinding the HDR process.


PostPosted: Sun Jul 21, 2013 3:05 pm    Post subject: Reply with quote
By the way, I think it was missed that I was actually having a friendly nudge at your own signature line equating bokeh with bad photography.. although it's true that there's no visible bokeh in this scene.. Cool

And thanks for the crops, although I still can't see it. If there was any halo artefact from the tone mapping I would expect it to stand out clearly where the chimney cowl penetrates the distant blue mountain. And it's as sharp as a tack with no halo.


PostPosted: Sun Jul 21, 2013 3:45 pm    Post subject: Reply with quote
GeorgeSalt wrote:

I have the same faint haze line effect following the line of hills in a layered landscape like this in HDR and non-HDR images of similar scenes, so it's not going to be something that's fixed by unwinding the HDR process.


Like you know, HDR is basically a tone remapping technique, handled by software which does not use selections, rather it works by ranges of colours/luminance.
Obviously this type of software has a difficult moment when it has to deal with high contrast transition areas, since he does not "know" what happens in the image, it can only "measure" it.
The local contrast ("clarity")applied to the image to enliven it after the tone compression creates the haloes. In Ian's photo is not that much visible, in fact.
Perhaps only exigent users can see it. Ian is good in making credible HDRs.
However, let me comment one of Ian's crops:

First I have to say that in order to analyze the contrast in the scene, it is possible to do it in colour, but it is easier to make it grayscale.
I have developed a luminance-to-zone scale that allows me to *roughly* analyze the contrast in an image.
This scale is NOT scientifical, only empyrical. Since the 256 values do not divide equally by 9, I have slightly emphasized the middle range (zone 3 to 7),
which is what most B&W photographers tdo anyway when evaluating a scene. This is my corrrespondance scale:

0-28 Zone 1
29-56 Zone 2
57-84 Zone 3
85-112 Zone 4
113-141 Zone 5
142-170 Zone 6
171-198 Zone 7
199-227 Zone 8
228-256 Zone 9

So here's Ian's crop after the "desaturate" command in Photoshop:



If you hover Photoshop's eyedropper tool over the image by keeping the left mouse button clicked, you can see the luminance values
change as you move the mouse around.
The analysis of this crop shows that there is a full stop difference between the lower and upper part of the green hill. The lower part falls in Zone 5, while the upper part falls in Zone 4.
Now, if you look at the whole image, you can see that there is no reason why there should be a full stop difference. Textures are similar. There is no cloud-cast shadow (besides, the sun is quite low on the horizon).
One stop difference means double (or half) the reflectred light, that is a lot of change to happen under identical lighting condition.
You say, it's because of the haze. That may be, but how is it possible, then, that the height of the haze follows the contour of the hill?
If you scan the green hill's contour from left to right on the Whole image, you see that there is a significant change in height as you go right. And the supposed haze line follows
that contour exactly: it raises where the hill is higher, it lowers where the hill is lower.
This does not happen to haze in nature. Haze is water suspended into the hair; it belongs like water pour into a pot: the water edge follows the force of gravity even if the pot is inclined.

The same blending effect happens with the background hill, although there the difference is less than a full stop due to the lower original contrast.
Yet the upper and lower parts of the background hill still fall into two different zones.

Even more evident is the altitude analysis. If we suppose that the brighter areas are due to haze, then it should follow a logical altitude pattern.
What happens, in stead, is that in Ian's image we would have a lower area with no haze, then a mid-low area with haze, then a mid-top area with haze, then a top area with no haze.
This of course is not possible. There fore it must be something else that causes the difference in brightness.


PostPosted: Sun Jul 21, 2013 4:11 pm    Post subject: Reply with quote
I think I'm starting to understand what you're talking about now, and it is very slight in this image. It's clearer when I look at zone 4/5 that you've marked on the right. I'm not convinced by those on the left, but purely because I don't think they're comparable sampling points (I was an ecologist, and I can in the colour image see that the groundcover is too different to compare luminance).

But, without it, even more depth would be lost from the image, so is this extremely slight effect good or bad? I think the aesthetic result often over-rides the technicalities.


And on a technical note, why is this effect seen on non-HDR images as well?


PostPosted: Sun Jul 21, 2013 4:15 pm    Post subject: Reply with quote
GeorgeSalt wrote:
If there was any halo artefact from the tone mapping I would expect it to stand out clearly where the chimney cowl penetrates the distant blue mountain. And it's as sharp as a tack with no halo.


Yes, but look at the size of the chimney and compare with the size of the whole image. Remember we are talking of large radius contrast adjustment.
The chimney is very small in the frame, most probably smaller in size than the radius cut-off size.
Anyway, I think that the only possibility to understand more of the scene is to see one of Ian's 16 originals, preferably one in the middle of the exposure ranges.
By comparing the contrast of the HDR image with the contrast in one of the originals, it will be possible to figure out a lot more than we actually know.


PostPosted: Sun Jul 21, 2013 4:25 pm    Post subject: Reply with quote
GeorgeSalt wrote:

And on a technical note, why is this effect seen on non-HDR images as well?


Sharpening, most likely.
Try taking two JPG shots of the same scene, one with standard sharpening and one with sharpening off.
Or, simpler, a RAW shot, so you can adjust sharpening level in RAW processor.
As you know, sharpening is a micro-contrast technique. It enhances the contrast within a small radius of pixels. So in small size transitions, what is bright becomes brighter and what is dark becomes darker.
This tricks our mind into thinking to see more resolution than what is actually there.
The "clarity" command of Lightroom, and similar tools, also present in HDR programs, use the same technique, but apply it to very large radius of pixels.
So, small transitions are left untouched, while transitions between large areas (areas larger than the radius of the effect) are affected.
That is why you see the halo on the hills, that are large, while you don't see it around the chimney - chimney is too small, it falls under the radius cut-off.
If you applied, instead, "traditional" sharpening (usually in the radius range from 0 to 2 pixels), you would see the halo around the chimney too, at high amounts of sharpening of course.


PostPosted: Sun Jul 21, 2013 4:34 pm    Post subject: Reply with quote
Many thanks for the explanation Orio, I won't lie, I haven't absorbed the wisdom fully, but I think I'm following you!

Sadly, I didn't keep my originals, but no matter. I'll go out again this evening and shoot a similar scene, it's hazy again today, perhaps more so, humidity is a little higher according to my dad's weather station thingie.

I will post both normal and hdr shots and will keep all the raw files so we can investigate this more thoroughly.

I suspect that it is my sharpening settings in Nik Shapener that are responsible, but it would be good to know exactly what is going on, it may well be a property of the HDR processing I did (with Photomatix 4).

Thanks again, very illuminating discussion. Smile


PostPosted: Sun Jul 21, 2013 4:52 pm    Post subject: Reply with quote
iangreenhalgh1 wrote:

I suspect that it is my sharpening settings in Nik Shapener that are responsible


That is absolutely possible, since NSP has a "clarity" slider too!
So HDR software might not be the cause in your case - although "clarity" adjustment in HDR software is common.

iangreenhalgh1 wrote:

Thanks again, very illuminating discussion. Smile


Yes!
I tried finding an image in my archive, to replicate the effect that George reports in regular images.
I could not find one that works well. The one I show here is too blurred in background and too flat in lighting to be useable for transition in line of trees.
But - with the front dead branch it can show somehow how radius as well as amount does influence the halo.

First crop is from RAW image with zero sharpening:



Second crop has Sharpening amount of 90 and radius of 2:



Thrid crop has Sharpening amount of 90 and radius of 3:



It is interesting to note that both halo and noise artifacts are perceived at radius=2, while only halo is perceived at radius=3
with the same amount.
It is sufficient only 1 pixel more in radius to eliminate most of the noise but still leave halo visible!
Since "clarity" usually works at radius of about 90-100 pixels, it is clear that the halo can be huge (as it only affects large areas), and the small parts unaffected.


PostPosted: Sun Jul 21, 2013 6:56 pm    Post subject: Reply with quote
This has been a very enlightening discussion, thanks for the explanation Orio.
Post new topic   Reply to topic
Home   Digital Gallery MF
Page 1 of 1