Processing order

I have to admit I’ve been doing more post-processing of my pictures since the eyesight failure. Most of it is cropping to make up for not seeing what in the viewfinder correctly, but some of it is the usual colour/brightness/contrast tweaking to get things just right. Then I usually shrink things down and sharpen them as a matter of course.

One of the photographers I follow recently did a very long video on all the rigmarole he goes through in processing, and at least I haven’t fallen into that trap. I say “trap” because I think a lot of people over-process their images, and not just the HDR ones. I probably will never do this because I now lack the ability to see subtle differences.

That said, I’m going to show you some subtle differences! The purpose of which is to demonstrate that what order you make adjustments (at least ‘automatic’ ones) has an affect on the final result. First the base image which has only been shrunken down and sharpened:

A somewhat poor image of lilacs, taken with a not-so-good lens.

This picture can benefit from both some colour enhancement and tonal correction as it is somewhat washed-out, lacking both richness and contrast. So let’s correct the white balance then enhance the colour, using ‘automatic’ functions for both, before shrinking & sharpening:

White balance, colour enhance, S&S.

But what if we colour enhance first, then correct the white balance?

Colour enhance, white balance, S&S.

There is a subtle difference, even to my eyes. Better? Worse? That’s a matter of opinion.

Now let’s look and see what happens if we shrink & sharpen first, then adjust white balance and colour:

S&S, white balance, colour enhance.

And in the other order:

S&S, colour enhance, white balance.

The last two images are the most alike, and you can see they are both different from either image which had colour changes before being reduced in size. This is because the computer program judges the whole image when making its adjustment ‘decisions’, and if the image is made smaller first there is less information to work with. Will that make the end result better or worse? That would depend on the particular image and your personal tastes. It is important to remember this especially if you are cropping segments out of the whole, as whether or not the removed portions are included in the program’s analysis for adjusting can have a significant effect of the final result (as in if large amounts of a particular colour or shadow/highlight are being left out).

There’s not only science to this, but art as well. You may find it helps to try such multiple changes in different orders to get the best results.

It’s still not a good picture, though.

Two from three

So much going on I didn’t know which way to turn. It has taken me days to decide just what the “weekend post” would be of/about. Thus it’s a sampling because I couldn’t make up my mind.

Canon T100 55-250mm lens

Canon T100 55-250mm lens

I’ve been ‘trying out’ the Canon lately to evaluate it as the next ‘main camera’ to use. I like using it but it simply can not do the things the Nikon P610 ‘bridge’ camera can, so I find myself repeatedly grabbing that instead when going out.

Nikon P610

Nikon P610

I have stopped using the Olympus for now. Not just because the battery is failing and I’m loathe to put money into it, but also because I need to force myself to evaluate the future of photography for me and it is unlikely the E410 is the answer no matter how much I like using it.

Olympus E410 40-150mm lens

Olympus E410 40-150mm lens

What with everything everywhere being as bad as it is (yes, my wife is still in England with no return date even guessed at) switching to “artistic” photography only is about all I can do. I loathe the idea of it as I do very much like taking wildlife photos. You know: pictures of birds I can’t actually see because they are small, far away, and hidden in tree branches. Do I need to mention the failing camera + failing eyesight thing again? No. Not going to say anything about the triple digit inflation rate around here either.

Just trying to keep my sanity together. Remind me again exactly why I should do that.

More analysis coming up. Er, camera analysis that is. I could probably do with the other kind as well, to be honest.

No more colours.

Jay Foreman’s No More Colours

As my eyesight continues to get worse this song becomes more significant.

(Taken with the Canon T100 and 24mm Vivitar.)

Colour and Black & White

A few images taken with the Canon G11, demonstrating how colour sometimes gets in the way of a good picture.

Last year’s leftovers.

Remove the colour and see only shape and texture.

Strange colour for a cement truck!

In monochrome the machine’s shape is emphasized.

Engines look good in full tones.

Still look good desaturated.

Boxcar graffiti.

Graffiti becomes less evident in gray shades.

You can decide for yourself which way looks better for each picture: I’m just demonstrating the possibilities to encourage people to try it. Only seconds are needed to desaturate a colour image – or switch it back.

The train was at a siding in Williams Lake on my most recent visit to that city.

One scene seen six ways

Manual images are fine, but sometimes need some post-processing to make them just right. Or maybe not.

The original. Underexposed, moody. Or is that ‘muddy’?

Automatic equalization applied in GIMP. Better exposure, odd colours.

Auto white balance correction applied in GIMP. Most people would say this is ‘right’.

Exposure corrected using brightness and contrast controls in GIMP. This is the most accurate to life.

Getting artistic: white balance correction with colours “toned down” by ‘fading’.

When all else fails try turning the colour off. Or even if the colour is good it doesn’t hurt to see what the monochrome version looks like.

Colour or not colour; that is the question

It won’t surprise anyone who has read even some of my blogs that I have my own views on whether to shoot colour or monochrome. You may have heard me express the sentiment that I generally shoot colour and then desaturate if I think the image would look better in B&W. It gives more choice, I feel. Other people prefer to shoot monochrome in the camera, and that is their choice. I do it myself occasionally, if I think it’s what the picture calls for. I’d love to be rich enough to have a camera “dedicated” to each, but that’s not very likely to happen.

Let’s face it: if the camera makers made it easier to flip between the two we wouldn’t be having the discussion, because a simple flick of a switch would give us either colour or B&W in a moment. Instead we have to paw through menus and push buttons, which rather spoils the fun. I’d like to see a combined ISO/Colour dial with speeds of 50 (please bring that back; the sun does shine sometimes) to 800 (above that doesn’t gain you much but noise) and half in colour half in B&W or maybe even thirds for high colour and low colour and B&W. I don’t know; if someone pays me to work out the details I will, okay?

Anyway the subject today is how and why I choose between the two. To start with, we have the Work Truck:

For me the colour is a distraction here, mainly because of the background blues competing with the subject which is practically monochrome (sepia) all on its own. In B&W the crazing on the panel is more prominent, and you might notice some of the smaller details of the form (such as how it sags on one side) because you’re not looking at the wide tonal range. I tried this in low saturation colour and didn’t like it. Shifting it to sepia (basically the colour of the dirt on the truck) however, works. Possibly the best version:

The background colours are no longer a distraction, and the monochrome aspect of the road dirt (the main point of the image) is emphasized. Although you could argue that the colour version puts the road dirt into vivid contrast with the rest of the scene.

Now let’s look at a picture which works either way:

In the colour version we see the nice brass of the candlestick as well as the red and green remains of previous candle wax in the base. The blue background complements all of it, including the shapely shadows. In the monochrome version the image becomes one of shape and texture, of which it has a lot. In fact you could say it’s more poetic as the snuffed candle contrasts with the long shadow of the daylight (side note: this is not early or late light, it’s just the very low sun angle we have at this time of year. It stopped me taking it direct-on because it glared back horribly. “Angle of reflection is equal to angle of incidence.”)  Now here’s two more versions, low colour and “sepia” (actually trying to match the brass tone of the holder), both of which “work” in my opinion:

If someone asked me to pick between the four I’d have a hard time of it. Perhaps I should do a large image with the four versions together, like Warhol? *LOL*

Now a picture which could only be in colour. If this were monochrome it would be gray on gray, as there wouldn’t be enough contrast to show the fine details. This basically is a picture of colour contrast:

Brush Strokes In The Sky

Finally here’s an image that only works in B&W. I thought this when I planned the shot, and so took it in monochrome to begin with which is unusual for me:

Lonely Bear

If this were in colour you would see the bright blue of the background cloth, the bright red of the bear’s scarf, and the contrasting browns of the bear itself. All of which would remove the sense of melancholy generated by the image of a teddy bear that’s been left behind for some reason.

We have a mixture of images from the two Kodaks in this series: the P850 is responsible for the candlestick and the bear, the V1003 took the truck and the sky (with some post-shoot help for its failing sensor).

Primary Colours

Some examples of selective desaturation for artistic effect. All photos taken with the good ol’ Kodak V1003 and processed in GIMP by turning down saturation on all but the intended colour.

Red

It is interesting to note that in “Red” some brown and orange tones were retained, because they are a form of red.

Yellow

This is my favourite. The old (’69 or so) GMC bus abandoned to use as a storage facility and looking forlorn indeed. There is some residue of colour in the weeds in front of it of course.

Blue

Here a couple of abandoned water pressure tanks lay in the weeds awaiting eventual recycling. This one brings up the issue of just what colours can be eliminated this way: the green of the small pine sapling is completely gone. Trying to retain green is another matter, as green is often not green but yellow and blue (or rather cyan in the weird world of photo colours) put together. Thus removing the yellow and/or cyan/blue results in a loss of green tones that you may prefer to keep.

This shows up in processing other colours too, if they are comprised of blending two tones. Sometimes removing the colours doesn’t make any noticeable difference:

In the second version magenta, cyan, and yes green have been removed. It’s hard to tell the difference, isn’t it? On a more complex pallet the secondary colours have more effect.

If you set out to do this yourself, you have to keep in mind what primary colours are in the image and try to envision what it will look like with some removed. Some cameras, by the way, have the ability to do this single-colour rendering within them. It is usually buried deep in the menu system, and of course shooting that way initially means eliminating any ‘normal’ version of the picture which might look better (although sometimes this is an in-camera post-shoot processing which retains the original; check your manual carefully).

The Chimney Series

Recently I had the opportunity of looking at an old farmstead (the story behind this is complicated). I didn’t have much time because of all the other things I needed to do that day, but I had an hour to look about. The one thing that struck me was the site where the newest house (built in the 1960s) had burned down leaving only the chimney. Well it was more than just a column of bricks, and I shot a few photos of it. If I’d had more time I would have shot more – and come ready with other gear. Different times of day or of year et cetera would have yielded more results just from this one mass of miscellaneous masonry. As it was I took a dozen photos with the camera I had, the T100, and about half of them turned out nice with little effort. I haven’t even tried desaturating any yet. But here are the ones I think came out good.

That last one I believe I’ll get a print made of it and hang it on the wall; it’s a great image of colour and texture and shape.

Perhaps in the future I’ll have the chance to return there and shoot some of the other old log buildings. Or perhaps not.

Hoorray for the Red, Green, and Blue

While sifting through the seemingly endlessly nested menus on my Nikon, I came across a setting that allows the camera to take multiple exposures (combining two or three shots into one composite image). We used to do this in the old days of film, often by accident (not all cameras integrated film wind and shutter lock, you know). There really isn’t much point to doing it on purpose, except for certain artistic expression or trying to fool people with “ghost” images. Nevertheless, once found the setting must be tried.

In my usual semi-instructive manner I decided to try it out by assembling the red, green, and blue aspects of a picture – much as our display screens do. So with the camera on tripod on a day that was still a bit too windy (causing some blur in the composite image) I took the same scene through the red, green, and blue filters:

And then the camera puts them together, and we can compare this to a straightforward “single shot”:

Note the composite picture has a somewhat magenta cast to it, owing to the filters not being “ideal” shades for RGB separation. But you get the idea of how it works. The camera even assembled a picture from the first two filters used, red and green:

You can see the lack of “blueness” in it, which is a strong indicator that the blue filter is the one which isn’t quite on spec because the RGB composite shows a similar lack, just not as severe.

In art class we learn that the primary colours are red, yellow, and blue. The secondary (or complimentary or ‘opposite’) colours are orange (made of red and yellow, opposite of blue), green (made of yellow and blue, opposite of red), and purple (made of blue and red, opposite of yellow). Then in photography they throw cyan (pale blue heading towards green) and magenta (pale red heading towards purple) at us, and explain it’s because light is yellow to begin with so everything has to shift a few degrees on the colour circle. Different shades of yellow too, depending on the light (the all-important white balance). Gee, did someone mention ROYGBIV – the seven visible colours of white light? Yeah we get in trouble picking which shade of Indigo or Violet we want to call purple. Oh and along comes the age of colour TV & digital imaging and suddenly we’re supposed to understand that because light is yellow (especially artificial light) we have to make colours from red, green, and blue instead. And this doesn’t touch on infrared, near infrared, and ultraviolet – which aren’t visible to us but are to our cameras.

Confused yet? If you are, you’re learning.

Pseudo-Panatomic-X

Kodak’s venerable black & white films were Panatomic-X with an ASA of 50, Plus-X with an ASA of 100, and Tri-X with an ASA of 400. Yes, they all started with lower exposure indexes, but those were the speeds they ended up at and thus the ones most people would be familiar with. You can probably see the relationships in the names: add one stop to Panatomic and you get Plus, add three stops and you get Tri.

There are many people experimenting with Tri-X simulations in the digital world today. There are even ‘filters’ built-in to some software to give your images that Tri-X look with a click of a button. There seems to be an obsession these days with ‘high speed’ exposure indexes, as digital cameras go up to ISO 6400 or more. That’s +6 stops from ‘normal’ daylight speed. Where are you planning on shooting? In the bottom of a mine shaft? Of course use of such high numbers then gives you the opportunity to complain about (and try and remove) the ‘noise’, which is the digital equivalent of film’s grain. In my day we tried to avoid that.

Here’s me, the old contrarian, complaining about the opposite; cameras’ minimum ISO setting seems to be 100. This is fine, but just as higher speeds have advantages so do lower speeds. You may find that hard to fathom because the problems of lower speeds are what drive us to use higher ones. Yet neutral density filters sell. Everything is a compromise. Just be thankful you aren’t using glass wet plates, okay?

So why would we want lower speeds? Not just for blurring motion or limiting depth of field (I will die before I use that stupid ‘b’ word). Just as high speed gives you more contrast, low speed gives you more range. In black & white we’re talking about defined gray tones. Think of it like this: let’s say Plus-X gives you ‘X’ (I wonder where I got that algebraic variable from, eh?) tones. Tri-X gives you less than ‘X’ tones, and Panatomic-X gives you more than ‘X’ tones. It’s like having more colours on your palette – only they’re all gray.

Mostly as an exercise in further exploring digital photography for myself, I set out to try and simulate Tri-X and Panatomic-X. The first was easy to achieve, and every bit as disappointing as the original film. Frankly I never liked using Tri-X because it was grainy and contrasty. It’s not my style. I’m really a boring, Kodacolor kind of guy. Nevertheless, this was an experiment for science!

There are two steps to it: determining what the difference is between how the camera shoots and how the film looks, then finding a way to make the camera output look like the film. It’s not as easy as it sounds: merely changing the camera setting to monochrome doesn’t work as it relies on what the manufacturer thinks is the right tonal quality and gradations. Slipping on a neutral density 2 filter to knock the ISO down to 50 doesn’t solve the problem either; it just lowers the EV. In fact, even in combination it comes up lacking. Why? Because your camera is biased about colours.

Black & white film (aside from specialty films) is made to be fairly even in its sensitivity to all of the visible spectrum. It doesn’t necessarily achieve this, and under certain lighting circumstances some adjustment is needed to render an image that is satisfactorily normal. If you’re familiar with shooting B&W you’re probably familiar with adding a K2 (yellow) filter to heighten contrast of the sky. In this case the bias in the camera’s sensor has to be overcome in a like manner.

I found my cameras all lean towards green. In converting the colour images to B&W (my preferred method – you’ll see why later) the various colours become gray tones based on luminosity or lightness or a combination of the two. If a colour does not reflect ‘equally’ it will slide into an incorrect tonal range depending on what conversion method is employed. Thus a colour may become lighter or darker gray than it should be. Admittedly this will not be noticed by most people, and probably doesn’t really matter. So long as you like the end result, it’s fine.

In my preliminary testing of that which I knew would happen, I shot the back end of my Xterra next to grass so I had red tail lights and green foliage to look at. Sure enough, changing to monochrome produced different results with different methods. All would have been acceptable by the end user, except where the ‘end user’ is me trying to simulate wide-tone Panatomic-X film. So it was time to apply some thinking, and some filtration.

Knowing which process turns which colour which way is a good place to start. My finding was: Lightness makes red tail lights lighter, green grass darker, yellow flowers darker; Luminosity makes red tail lights darker, green grass lighter, yellow flowers lighter. The second key to the puzzle was knowing the camera favours green. So we have to make the red tones lighter and use the luminosity function in order to get the most accurate conversion.

Enter the orange filter. Fortunately this is not a true ‘Wratten 21’ orange as that would be too extreme I think. Just enough to shift the colour, and incidentally require 1 more stop of EV – effectively lowering the ISO to 50. (The Wratten 21 would be 2 stops.)

I have to say I spent some time adjusting the camera’s “picture type” settings as well, to see if that would help. Subtle 1 step changes didn’t really show at all – or at least I couldn’t see them. I went with as ‘neutral’ as the settings could be, with an increase in sharpness to help the definition. I also reduced the MP to 4.5 for shooting, with a further reduction for Internet display. Yes, I realize this alters what you see. So does your screen and your eyes. You’ll just have to take my word for it that it’s right.

First, the picture as-shot. It’s a kind of snap-shot thing because I wanted to get a lot of variety in the composition; it helps with determining the quality of the final result as well as aiding the computer in doing the conversion (the more monochromatic the original, the more inaccurate the change to B&W).

Next, the black-and-white version. True blacks, true whites, and many gray tones in between:

Last, the colour-corrected version. This is why I like shooting in colour all the time; if this had been monochrome but would have looked better as colour, it would be lost. As it is the filter correction for B&W is subtle enough to be eliminated and deliver an acceptable quality colour rendition as well:

Perhaps a tad magenta. That could be fixed if I wanted to spend the time working on it.

Here’s a more dramatic series, same processing order. You can see what I mean about what happens when starting with a fairly monochromatic image.

Anyway, I plan to shoot some more like this and some starting with monochrome but ‘slowed down’ with a neutral density filter to see how that compares. It may be just as good, or better, or worse.

I also plan to try simulations of Kodachrome and Ektachrome once I’ve worked out the initial settings. But probably not Agfachrome with its muddy “European” colours. 😀