Aussie triplets - fixed yay!

NeilL

Got ' em back up, phew! As before, three pics all related. NW Tasmania, Australia. Thanks again to thapamd for his earlier comment, now unfortunately in internet heaven. Enjoy!






























Thanks for looking!

Canon 40D, 24-70 f2.8L

Neil

thapamd

I remember this! As I mentioned before, the 3rd shot is my favorite because of the composition and the nice lighting.

NeilL

thapamd wrote:

I remember this! As I mentioned before, the 3rd shot is my favorite because of the composition and the nice lighting.

And thanks again!

NeilL

Bumping in hopes of some more crit, especially of the post work. These pics probably represent my high point so far in post processing skills so I'd be grateful for any reactions and pointers. I was aiming for a forceful impact of feeling (serenity) and color (lush pastoral), as well as a variety of contrasts eg light/shadow, pasture/trees, FG/BG.

So if you are inclined, please cast a critical eye over them and let me know what you think!

rutt

NeilL wrote:

So if you are inclined, please cast a critical eye over them and let me know what you think!

I guess it says something that I chose the first of these when I was looking for practice with the Dan Marguls Picture Postcard Workflow which is supposed to add both contrast and color and achieve a happy picture postcard type look. It seemed too flat, lacking in pop. I ran through the steps of the PPWF and got this:



I was going for a somewhat conservative postcard here because I didn't think you were after totally the totally happy look it can make given its head. Still I think the better blacks, sky details, and deeper colors all make it work better.

There are no doubt other ways to get similar improvements, and learning the PPWF will take a while, but once learned it's very fast. In fact, that is its main mission profile.

If you look closely, you can see that I am still mastering some of its details, particularly using RGB curves to remove all the casts early on. Take that as an issue with me, not the workflow. This is only my third try.

thapamd

Nicely done, Rutt. However, to tell you the truth, I prefer the original version. In the modified version, the trees don't look "natural". Perhaps the midtone contrast was boosted too high? Yes, the skies and sheep look more saturated, which may be an improvement, but I'm not sure about that either. In addition, there is a slight halo where the sky meets the earth.

NeilL

Thanks rutt and thapamd. I appreciate the dialog. Experiments/versions in PP are always interesting and valuable, and if others do them then it saves you the work. rutt, the postcard treatment sure works, if "postcard" is the intention. I can't see it well enough to judge the consequences to finer points of image quality. But a nice job! Thank you!

I did these pics with great attention to pixel quality. Everyone knows the risk of using PS's readymade tools - they can quickly degrade pixel quality, especially when used repeatedly, and they work on the prime channel only, therefore they represent a kind of sampling, similar in effect to compression, and having similar degradation consequences, and they make changes to information in the individual RGB channels' information in the prime which might hinder rather than help. The posterisation caused by using curves is an example.

I did in fact use curves as part of my technique on these, but very conservatively, and typically with the luminosity blend mode. I used justthorne's development techniques as I know them to date. I also used plugins from Nik and onOne. The particular plugins I used passed my scrutiny for preserving pixel quality and the character of the subject and my intentions. Take the zoom up to 300% for example and you will see zero haloing. Saturation was enhanced yet there is no bleeding or hazing that you get from PS's saturation adjustment. It's clean and the contrast too is clean. The histograms were perfect.

Not blowing my own trumpet here, but really just giving you the inside story into my approach and attitude. In terms of pixel quality I think I won. I think LAB was originally developed by Margulis for graphic work, and my understanding is that its translation to some kinds of photography needs to be done with a great deal of caution and sensitivity.

However, there're more global things that could be discussed, like exposure, color, contrast, sharpness, DOF, lens choice, image orientation, composition, crop, lighting, appropriateness to subject, emotion and aesthetic effect... I take your point rutt that the first image has less punch, and I think I made one curves adjustment, to lighten it, too many! I made that call because it looked slightly underexposed.

As most of you know I am just going on 2yr old in my digital photography life, so I think I am at last walking... with some help that I am grateful to get from you guys, but a ways to go from running, let alone becoming a champ!:D

thapamd

Neil, I like your phylosophy about preserving as much of the original data as possible when doing edits. Personally, I don't use plug-in's at all and prefer to do all my edits either in Lightroom or as layers in PS (currenlty I use CS4). Whenever possible, I prefer to convert my original RAW files to smart objects before making any edits in PS. One must be careful, however, not to get lost in all this technical "mumbo-jumbo." In the end, I always look at the the image and have others look at it and ask for my or their first response to it...their visceral response, if you will. Sure we can analyze a photograph to death by looking at color, composition, and what have you, but I want to what your initial response to an image is. I hope I am making myself clear.

Also, I've noticed that much of the "flaws" we can see at 100%, 200%, or 300% on screen will never show up on print. I've made and sold many prints of various sizes (ranging for 3" x 5" to 3' x 9') so I can say I speak from experience. I think it's important to know about the technical aspects of photoshop and image editing, so you don't "ruin" your originals, but we should also keep in mind its artistic aspects. Just my $0.02.

rutt

Yeah, my treatment should be seen as a learning experiment for me, not a loss for the PPWF, which I haven't really mastered. But I think it does show a few areas for potential improvement in the original even if it doesn't overcome my flaws.

1. Better shadow depth with richer blacks and better detail
2. More detail in the clouds
3. Bluer sky (maybe less than I added)
4. More depth and variation in the water surface

NeilL

thapamd wrote:

Neil, I like your phylosophy about preserving as much of the original data as possible when doing edits. Personally, I don't use plug-in's at all and prefer to do all my edits either in Lightroom or as layers in PS (currenlty I use CS4). Whenever possible, I prefer to convert my original RAW files to smart objects before making any edits in PS. One must be careful, however, not to get lost in all this technical "mumbo-jumbo." In the end, I always look at the the image and have others look at it and ask for my or their first response to it...their visceral response, if you will. Sure we can analyze a photograph to death by looking at color, composition, and what have you, but I want to what your initial response to an image is. I hope I am making myself clear.

Also, I've noticed that much of the "flaws" we can see at 100%, 200%, or 300% on screen will never show up on print. I've made and sold many prints of various sizes (ranging for 3" x 5" to 3' x 9') so I can say I speak from experience. I think it's important to know about the technical aspects of photoshop and image editing, so you don't "ruin" your originals, but we should also keep in mind its artistic aspects. Just my $0.02.

Yeah! I totally agree that there is a "saleability", or at least "convincing", bottom line to every photograph. However, I think you don't get there unless from the pixel level. You're right that I am a bit fixated at the pixel level at the moment, but I think that every beginner in photography should be at some stage in their development. It's like there is no shortcut to maturity.

Also, I think pixel quality is the photographer's credentials. Does it matter that Leonardo could draw? Of course it does. Do people ask themselves how well he could draw when they react to his pictures? Of course they don't.

thapamd

Neil, I think you and I are speaking the same language. Get the basics and fundamentals down so you don't stumble across them. When that becomes second nature, let your art do the talking.

NeilL

rutt wrote:

Yeah, my treatment should be seen as a learning experiment for me, not a loss for the PPWF, which I haven't really mastered. But I think it does show a few areas for potential improvement in the original even if it doesn't overcome my flaws.

1. Better shadow depth with richer blacks and better detail
2. More detail in the clouds
3. Bluer sky (maybe less than I added)
4. More depth and variation in the water surface

I take all these points, thanks! I am impressed by the variations you found in the water! They were a surprise. I didn't even notice them before. I would still want to protect my pixel quality and the overall naturalness.

thapamd

rutt wrote:

Yeah, my treatment should be seen as a learning experiment for me, not a loss for the PPWF, which I haven't really mastered. But I think it does show a few areas for potential improvement in the original even if it doesn't overcome my flaws.

1. Better shadow depth with richer blacks and better detail
2. More detail in the clouds
3. Bluer sky (maybe less than I added)
4. More depth and variation in the water surface

Your points are well taken and they are completely valid. Perhaps a combination of your version and the original will produce the "best" image.

thapamd

Re-reading this post, I'm reminded of my own progression in photography. A little self-reflection from time to time is good for the soul.

When I first started doing digital photography, I was obsessed with lens sharpness, noise, exposure, and color balance. I would literally spend hours and hours analyzing the technical aspects of an image to death.

Then came my composition obsession. A tilted horizon became my pet peeve. I trained my eyes to detect even a 1/3 degree of tilt. If an image didn't comply with the 2/3, 1/3 rule or didn't have something "interesting" in the foreground, midground, and background, I considered it a "bad" image. I became overcritical of centered images or images where the lines didn't necessarily lead the eyes anywhere.

Now, I think my obsession is with light. If the light isn't "good" then I don't even bother taking my camera out. I'm sure I've missed a lot of potentially good shots because of this stubbornness.

God only knows what my next "obsession" will be.

I think we all have to go through a similar process as I've outlined above before we can truly appreciate photography. As you said, there really are no shortcuts.

NeilL

A little explanatory note is in order...

I referred earlier to the 'prime' channel. justthorne has alerted me to the fact that this term is one chosen by him in his development system for the RGB luminosity channel, and wouldn't be familiar.

Here is his comment:

"most people won't know what the "prime channel" would mean. If you called it the Luminosity (or even more exactly, the "RGB Luminosity"), they would. I elected to name it differently to minimize confusion with LAB's Luminosity channel (or with HSL's), which is discernibly and qualitatively different."

rutt

You know, I don't really know what "pixel quality" is or why you would want it. Please enlighten me.

Richard

rutt wrote:

You know, I don't really know what "pixel quality" is or why you would want it. Please enlighten me.

:lurk

NeilL

Richard wrote:

:lurk

Come down from the popcorn gallery, R, and join in! Afterall you are more experienced than me and could probably make a better answer than me.

NeilL

rutt wrote:

You know, I don't really know what "pixel quality" is or why you would want it. Please enlighten me.

stand by...

rutt

Tried another of these for PPWF practice:



Original here.

Richard

NeilL wrote:

Come down from the popcorn gallery, R, and join in! Afterall you are more experienced than me and could probably make a better answer than me.

Well, I don't really know what pixel quality is either, but I would assume whatever people mean by that, it is not to be taken literally. Smarter people don't have higher quality neurons, AFAIK. I suppose they really mean image quality at a low radius-- color accuracy, sharpness, detail, contrast, lack of noise/aberrations, etc.

I hesitate to comment on the differences between your processing and Rutt's because I am not using my usual machine and I don't quite trust this monitor. FWiW, on a Macbook, Rutt's renderings look oversaturated to me, while the originals have a more interesting, unusual look--sort of HDR-like but without the typical over-the-top colors. There's something a little unnatural looking about the local contrast, but the effect is not unpleasant. On the contrary, it makes me want to look more closely, which is usually considered a good thing.

thapamd

rutt wrote:

Tried another of these for PPWF practice:



Original here.

I agree with Richard. The reworked version looks a bit over-saturated on my monitor as well. However, the added contrast really gives the shot some needed "punch." Again, I think the "best" version is probably somewhere between the original and the reworked. Just my $0.02.

rutt

thapamd wrote:

I agree with Richard. The reworked version looks a bit over-saturated on my monitor as well. However, the added contrast really gives the shot some needed "punch." Again, I think the "best" version is probably somewhere between the original and the reworked. Just my $0.02.

Well, that's actually a part of the workflow. At the end of contrast enhancement, save a duplicate. Then color enhance to get those bright colors. Finally blend the contrast enhanced version into the color and contrast enhanced version to cut back the brightness. When I did it, I was greedy and kept all the color enhancement. Looked good to me, but maybe I'm too used to the June Massachusetts colors and am not thinking Australia. So take my effort as an example of just how far this can be taken, not the proper endpoint for this particular image.

As before, though, I think you just have to like my version for the deeper shadows, the details in the trees and brush, and the water surface. The deep green grass and deep blue sky might really not be Australia, but they certainly are picture postcard.

NeilL

rutt wrote:

You know, I don't really know what "pixel quality" is or why you would want it. Please enlighten me.

The story as I know it -

The word pixel is used to refer to the unit photodetector in a sensor array, and also to the unit representation of the output of the digital camera system on a digital display.

This dual use reflects the two main stages in the production of a digital photograph.

A photodetector is a device for sampling light, it collects certain information from the light which hits it. It reacts to this information by translating it into electrical values. The sensor chip further processes these electrical values into binary information in some format (RAW, jpeg etc). This binary information is stored in the camera's memory. This is your digital photograph waiting for your viewing.

Obviously, up to this stage the information from the light source has undergone a number of samplings and transformations, and each of these adds to and deletes from the source information. This is where the concept of quality comes in. It is how well or badly, in physical and aesthetic terms, the source information fares through these samplings and transformations, additions and deletions - how these affect what you finally see. This quality is a major influence on your choice of gear.

The second main stage is post production. The digital information in your images is transferred from the camera to your computer. In the case of a RAW file the information is further translated into tif or jpeg formats. Jpeg files are compressed, which involves another set of sampling and transformation, and binary information from each pixel is changed to create alterations in brightness, saturation, and contrast. Again, this is subject to a physical and aesthetic quality assessment.

Subsequently each tool that you use in an editing application allows you to alter further the source information which was originally obtained by the photodetector. You are seeing this altered information in a representation in a digital display via, let's say, liquid crystals, called pixels. The ability of these pixels to represent the original data is pixel depth. How much of the original information is still there, and how well or badly what information you have represents the original information, and how much of the data sent to a display pixel it can display - physical meaurements, and how much you like what you see on your screen - aesthic judgments, all of that is pixel quality as the term is used in assessing the final photographic product.

I used the posterisation caused by curves earlier as an example of degradation of pixel quality, so let's use it again. The curves tool uses the image information to increase dynamic range. This necessitates spreading the information and consequently reducing the information in any one tone value. Less information to any one display pixel means reduced pixel quality.

How 'bout that?!

DISCLAIMER: any or all of the above could be wrong!!

rutt

I guess I still don't know the difference between pixel quality and just plain good looking images.

We see very differently than a camera does and the most photographically accurate possible rendering of an image may not be very much like what we remember when we shot it. White balance is an example of this. We tend not to see the casts that the camera picks up. Later on, when the image is displayed or printed, our eyes don't neutralize the casts to the same extent as they do when we were immersed in the scene.

We are also really good at dynamic range, seeing detail in each lighting level. We can look into a window on a sunny day and see an interior scene while also registering a bright exterior. No current camera and display device can do this, at least not well.

So, my goal is to make images that match the viewers' memory and perceptions. This is a much more interesting project than just trying to deliver the physical reality of the pixels the camera captured.

Or perhaps I still don't get what "pixel quality" means.

NeilL

Richard wrote:

Well, I don't really know what pixel quality is either, but I would assume whatever people mean by that, it is not to be taken literally. Smarter people don't have higher quality neurons, AFAIK. I suppose they really mean image quality at a low radius-- color accuracy, sharpness, detail, contrast, lack of noise/aberrations, etc.

I hesitate to comment on the differences between your processing and Rutt's because I am not using my usual machine and I don't quite trust this monitor. FWiW, on a Macbook, Rutt's renderings look oversaturated to me, while the originals have a more interesting, unusual look--sort of HDR-like but without the typical over-the-top colors. There's something a little unnatural looking about the local contrast, but the effect is not unpleasant. On the contrary, it makes me want to look more closely, which is usually considered a good thing.

Very interesting comments, R!

Regarding smartness - it is probably best thought of as an emergent property of the body's systems. As far as smart people's neurons are concerned, they do differ. In general, there are more of them, they have different response profiles, many more dendrites, and also many more supporting glial cells.

NeilL

rutt wrote:

I guess I still don't know the difference between pixel quality and just plain good looking images.

We see very differently than a camera does and the most photographically accurate possible rendering of an image may not be very much like what we remember when we shot it. White balance is an example of this. We tend not to see the casts that the camera picks up. Later on, when the image is displayed or printed, our eyes don't neutralize the casts to the same extent as they do when we were immersed in the scene.

We are also really good at dynamic range, seeing detail in each lighting level. We can look into a window on a sunny day and see an interior scene while also registering a bright exterior. No current camera and display device can do this, at least not well.

So, my goal is to make images that match the viewers' memory and perceptions. This is a much more interesting project than just trying to deliver the physical reality of the pixels the camera captured.

Or perhaps I still don't get what "pixel quality" means.

Simply, good images depend on good information and good representation of that information - physical values.

Whether you like what you see involves more than the physical values in the image.

Both make pixel quality.

Some styles of degradation of physical pixel quality, as in grunge, are liked or disliked.

I like to have the best pixel quality (physical) to start with, and I like to "degrade" (alter) that pixel quality only to the extent that it pleases me (aesthetic). Eg no one likes noise (physical), but some add grain (aesthetic).

rutt

Since I still don't really understadn the pixel quality concept, perhaps you'll tell me whether my last (more careful) PPWF attempt improved or degraded it. Both?

Maybe I'm being a little willfully obtuse here. It doesn't seem that you are in pursuit of something you can actually measure. I'll try to reach for the concept and you can tell me if I've got it.

1. Good initial capture with good focus, exposure, no camera motion blur, high resolution
2. Post introduces no visible artifacts, perhaps even not when highly magnified
3. Post loses no information from the original capture
4. Oh, and we like it

Is that basically it? For example, we might decide to plug some shadows to get rid of noise and add drama. But this might also lose some noisy detail from the capture. Does this degrade pixel quality? I'd guess so, eh? We'd prefer some sort of perfect noise removal move.

The LAB color moves, so long as the result stays in gamut of the ultimate display device, shouldn't degrade pixel quality; they drive colors apart, but they don't create color ambiguity where there was none. A B&W conversion is a huge loss of pixel quality, no matter how successful. S-shaped curves risk pixel quality in the highlights and shadows as the price for a possible increase in the midtones.

Am I sort of getting it?

NeilL

rutt wrote:

Since I still don't really understadn the pixel quality concept, perhaps you'll tell me whether my last (more careful) PPWF attempt improved or degraded it. Both?

Maybe I'm being a little willfully obtuse here. It doesn't seem that you are in pursuit of something you can actually measure. I'll try to reach for the concept and you can tell me if I've got it.

1. Good initial capture with good focus, exposure, no camera motion blur, high resolution
2. Post introduces no visible artifacts, perhaps even not when highly magnified
3. Post loses no information from the original capture
4. Oh, and we like it

Is that basically it? For example, we might decide to plug some shadows to get rid of noise and add drama. But this might also lose some noisy detail from the capture. Does this degrade pixel quality? I'd guess so, eh? We'd prefer some sort of perfect noise removal move.

The LAB color moves, so long as the result stays in gamut of the ultimate display device, shouldn't degrade pixel quality; they drive colors apart, but they don't create color ambiguity where there was none. A B&W conversion is a huge loss of pixel quality, no matter how successful. S-shaped curves risk pixel quality in the highlights and shadows as the price for a possible increase in the midtones.

Am I sort of getting it?

In short, no.:D

Re 1, what you list here are aesthetic values (except the last), not data values (measurements), you need both

Re 2, artefacts become an unavoidable and integral part of the data as the original data is processed, at whatever stage. The issue is the influence of the artefacts on the original data, the kinds of changes to the original data they cause, and the effects of their combined presence, in physical measurements and in aesthetic acceptability

Re 3, both loss of data and the effects of change of data influence pixel quality

Re noise, it is an addition to the original data, not a degradation of that data, but in effect is typically disliked, so affects pixel quality badly. NR masks noise by altering original data and added data (artefacts), so it degrades original data while increasing aesthetic pleasure (if noise is unwanted), the net result is the subject of a very lively debate

Re b&w conversion, there is no loss of data involved since there is no color data in the original data, it is all tonal data

Any change to original data is a degradation of that data, in physical terms. In judging pixel quality that has to be weighed against aesthetic gains.

You can't escape the facts that digital photography is highly sophisticated physics-optics-electronics, and also a form of pleasure and art.

Richard

NeilL wrote:

Re b&w conversion, there is no loss of data involved since there is no color data in the original data, it is all tonal data

Sure there is, you're just not looking at it from the correct level of abstraction. The tonal values generated by the sensor are specific to the Bayer filter wavelengths for the individual cell, which can be red, green or blue. The pixel is determined algorithmically on the basis of the values of adjacent cells. Information is lost when converting to B&W, as all channels end up with the same RGB values for a given pixel. If this were not the case, you could do a B&W to color conversion algorithmically.

NeilL wrote:

Any change to original data is a degradation of that data, in physical terms. In judging pixel quality that has to be weighed against aesthetic gains.

Hmmm....seems to me you might be making too big a deal about the physical side of things. I don't think of the time I spend in post as degrading data, but as enhancing an image. I don't see any value in thinking about individual pixels. The image is also emergent.

NeilL

Richard wrote:

Sure there is, you're just not looking at it from the correct level of abstraction. The tonal values generated by the sensor are specific to the Bayer filter wavelengths for the individual cell, which can be red, green or blue. The pixel is determined algorithmically on the basis of the values of adjacent cells. Information is lost when converting to B&W, as all channels end up with the same RGB values for a given pixel. If this were not the case, you could do a B&W to color conversion algorithmically.

Hmmm....seems to me you might be making too big a deal about the physical side of things. I don't think of the time I spend in post as degrading data, but as enhancing an image. I don't see any value in thinking about individual pixels.

Not sure I can cede you the first point just yet!

Re the second, you enhance (or ruin) an image by manipulating individual pixels. Don't see how you can do it otherwise!

There is not an artist of value who doesn't understand the physical characteristics of their paints, and no sculpture of value who doesn't understand the molecular structure of their stone, and no woodturner of value who doesn't know the same about their wood... and no digital photographer of value who doesn't take loving care of each of his pixels!:D