Agawa Canyon

[Manfred M]

Another image that I have never been quite happy with is now ready to be printed.

This is the Agawa Canyon, a fall favourite for tourists all over the world. This is the end of the line of the 183 km / 114 mile trip from Sault-Ste-Marie, Canada on the Algoma Central Railway. One can hike in to the park, but most people take the train ride.

[mugge]

I love the colours, green orange and yellow but I find the scene itself quite overdone for my taste. That is only my opinion.
Cheers Ole

[Wavelength]

I feel an ill-proportionate inclusion of the sky , a feeling of an overall imbalance from my personal aesthetic level. But i like the left bottom 1/4th of the total image....

[Manfred M]

I feel an ill-proportionate inclusion of the sky , a feeling of an overall imbalance from my personal aesthetic level. But i like the left bottom 1/4th of the total image....

Thank you Nandakumar - you are 100% correct. I've been giving other people that advice, yet in this example, I couldn't see it until you pointed it out. It is a much stronger image without the sky.

[DanK]

I hadn't commented because I am away from my good monitor, but I have calibrated the monitor I have where I am now, and it gives me the same impression I had looking at the image on an iPad. That is, it seems oversaturated to me. Then again, I wasn't there.

[Wavelength]

Now it is a really wonderful scenery!!!

[Manfred M]

I hadn't commented because I am away from my good monitor, but I have calibrated the monitor I have where I am now, and it gives me the same impression I had looking at the image on an iPad. That is, it seems oversaturated to me. Then again, I wasn't there.

I had another look at why this would be. As a general rule I never increase saturation or vibrance when I prep the image in the raw converter and I generally do all my dodging and burning using the luminosity blending mode so as to not impact the saturation.

I also tend to use a custom profile when I do my raw conversions and this shot was taken when I still had the D800. Somehow the custom profile went AWOL during one of the major Adobe upgrades. Not much of an issue with my new work when shooting the D810. The main difference between the stock Adobe Standard profile and the custom profile I made with my xRite Color Checker Passport was with the reds and to a lesser extent the yellows. They were more vibrant in the custom profile, which suggests that the Adobe profile is suspect.

As I have been revisiting old images during COVID, I've been working up the raw data and for a change used the Adobe Landscape profile. This is a lot more vivid and has higher contrast than Adobe Standard I've used in other images. When I use Adobe Standard, it becomes a lot more flat. The contrast and colour contrast seems to be the main difference between these two profiles

That all being said; this part of the province is considered to be the best fall leaf viewing area. A sunny, cool day where the humidity in the air is way down, the colours are absolutely unreal. That being said, I took this shot five years ago and am not likely to be there to check this year (10 hour non-stop drive from home). So no, the colours and saturation does not look over the top to me; but could come down a tad. People do visit here because the fall colours are "over the top".

Here is the image using the Adobe Standard profile.

[xpatUSA]

I had another look at why this would be. As a general rule I never increase saturation or vibrance when I prep the image in the raw converter and I generally do all my dodging and burning using the luminosity blending mode so as to not impact the saturation.

Must be time for some pedantic analysis ...

The two posted renditions are certainly different:



The first image saturation:



The second image:




There does appear to be some over-saturation in either image, more in the first ... well, in the HSV model anyway ...

[pschlute]

I like the saturation of the first image and agree about losing the sky. The trains make the image and the twin features of water and track leading into the distance.

[Manfred M]

Must be time for some pedantic analysis ...

The two posted renditions are certainly different:


The first image saturation:


The second image:




There does appear to be some over-saturation in either image, more in the first ... well, in the HSV model anyway ...

I would hope that there was some difference between the two image as they were built from a different raw conversion process.


I'm not sure why you do the analysis with the HSV colour model. It is interesting historically but really doesn't answer the question you seem to be implying that it does. Over saturation (and under saturation) is an opinion, not something that has a technical answer that can be measured.

[DanK]

I agree abut the sky.

Re the profiles: I think the rendition using the Adobe standard profile is a bit flat and could use some additional work, but I would still prefer that to the first, which is exaggerated and unrealistic.

[Manfred M]

This one uses the Adobe Color profile. Still too flat in my view and I still, by fall prefer the colours in #4. Far more realistic, in my view. The fall colours are general not dull.

[xpatUSA]

I'm not sure why you do the analysis with the HSV colour model.

Was wanting to measure the saturation distribution, and the HSV model is familiar to most of us.

I can do it in HSL or LCh - a different saturation measure or Lab Chroma. How about Y'CbCr and thereby the chroma axes?

It is interesting historically but really doesn't answer the question you seem to be implying that it does. Over saturation (and under saturation) is an opinion, not something that has a technical answer that can be measured.

100% saturation is easily detected in HSV. I am saying, not implying, that there are bottomed channels in the images. In HSV, a bottomed channel means 100% saturation as defined here. Since 100% saturated colors are rare in Nature, bottomed channels indicate over-saturation, by which I do not mean more than 100%.

I could take another at the first image out of interest, just to see the amount of "bottoming" ...

[Manfred M]

Was wanting to measure the saturation distribution, and the HSV model is familiar to most of us.

I can do it in HSL or LCh - a different saturation measure or Lab Chroma. How about Y'CbCr and thereby the chroma axes?



100% saturation is easily detected in HSV. I am saying, not implying, that there are bottomed channels in the images. In HSV, a bottomed channel means 100% saturation as defined here. Since 100% saturated colors are rare in Nature, bottomed channels indicate over-saturation, by which I do not mean more than 100%.

I could take another at the first image out of interest, just to see the amount of "bottoming" ...

Ted - sorry you missed the point.

Having saturated values has little to no impact on if someone considers an image oversaturated or not. It's an aesthetic argument and nothing more. It's like suggesting an image is oversharpend (or has not been sharpened enough), has too much or too little contrast, too warm or too cool, etc. You can't measure it; but you can give an opinion.

My personal taste is that I tend to like images that are bright and saturated, whereas a lot of photographers will go in the opposite direction and create dark and moody images. There is no right or wrong, just individual taste.

Because of the saturated colours in this scene (especially the greens and yellows), I have been working in the ProPhoto RGB colour space, so when I convert to sRGB for display on the internet, I am not at all surprised that we would see saturation when you look at the HSV values as the rendering intent would have driven the OOG values onto the hull of the sRGB colour space, so those colours should be showing up as saturation in your analysis.

As I was working to print the image, I am not constrained by the sRGB colour space. I do test the various papers that I use through the soft proofing process to identify which of my papers are best able to print the image. If I go with the Hahnemühle Fine Art Baryta, virtually the whole image is in gamut; quite a few of the hues are outside what the Adobe RGB colour space can handle. Unfortunately, I only have that paper in 8-1/2" x 11" so I'm going to have to wait until my order with larger sized paper comes in from the distributor.

[mugge]

Just my two cents worth. The last version is more pleasing to the eye whereas #4 is (merely) stunning.
Cheers Ole

[Acorn]

For me it's a too complicated picture with too many distractions.

[xpatUSA]

Ted - sorry you missed the point.

<>Because of the saturated colours in this scene (especially the greens and yellows), I have been working in the ProPhoto RGB colour space, so when I convert to sRGB for display on the internet, I am not at all surprised that we would see saturation when you look at the HSV values as the rendering intent would have driven the OOG values onto the hull of the sRGB colour space, so those colours should be showing up as saturation in your analysis.

" ... those colours should be showing up as saturated in your analysis."

That they are, Manfred.

0.6MP, over 20%, of the blue channel is bottomed which probably prompted Ole's and Dan's comments (posts #2 and #5).

So it would appear that it's OK to post images here that were developed in a wide-gamut color space but then converted to sRGB without bothering to adjust out-of-gamut colors.

As I was working to print the image, I am not constrained by the sRGB colour space. I do test the various papers that I use through the soft proofing process to identify which of my papers are best able to print the image. If I go with the Hahnemühle Fine Art Baryta, virtually the whole image is in gamut; quite a few of the hues are outside what the Adobe RGB colour space can handle. Unfortunately, I only have that paper in 8-1/2" x 11" so I'm going to have to wait until my order with larger sized paper comes in from the distributor.

Again, from that, it seems OK to post an image here without accounting for hues that are out-of-gamut for the average Viewer's monitor. I guess that's the point that I missed ...

[DanK]

0.6MP, over 20%, of the blue channel is bottomed which probably prompted Ole's and Dan's comments

It was the yellows and light greens that first grabbed my attention.

I think it's fine to post anything here for comment, and I also think that images may look fine in sRGB with some problems of this sort. My point was simply that the first image looks artificial to me, and in a way that is common in landscape photography one sees on the web.

[xpatUSA]

0.6MP, over 20%, of the blue channel is bottomed which probably prompted Ole's and Dan's comments

It was the yellows and light greens that first grabbed my attention.
...

Yes, yellows in particular are opposite to blues. Therefore if yellow saturation is increased by whatever means, blue does go down - as that image demonstrates quite clearly.