Exercise: Sharpening for print

Aim: The aim of this exercise is to take an image and produce 4 versions; one with no sharpening and the others with varying degrees of sharpening from mild to aggressive.  Then compare these at 100% on screen and printed

Approach and results:

I chose a busy picture of Hong Kong for this exercise.  There is a lot of detail in the image including faces.  I then produced 4 versions from RAW varying the sharpening as follows:

Image 1 – No sharpening, image 2 – 13, image 3 – 70 and image 4 – 120.  The figures are on a scale of 0-150 in ACR.

Comparing these on screen at 100% (actual size) image 1 was clearly soft, I could tolerate image 2 and 3 but image 4 looked grainy and pixelated in comparison.

Next I printed the same section of each image at 100% and compared these to each other and their on screen versions.  It was quite obvious that the printed versions were far less grainy than the on screen versions.  In the first instance I preferred images 3 and 4 with the others looking rather soft by comparison.  Using a magnifying glass you could again see the grain appearing on images 3 and 4 but this is pixel peeping.  If I had a printer large enough to print the whole image at 100% I certainly wouldn’t be standing close enough to see the grain and I wouldn’t be checking it with a magnifying glass!  The point is that prints are far more tolerant to aggressive sharpening and in fact often look better for it when viewed from a proper viewing distance.  My overall preference was for image 3 which had a degree of sharpening that I would not normally go to making this a very useful exercise.  Edges were sharp and clean and face detail was clear.  Also there are other settings available when sharpening such as radius and also, if a plain area looks grainy you can apply some luminance noise reduction to improve this.

hk70sml

Learning points:

Prints have a greater tolerance for sharpening even up to 100%.  I should sharpen my images more if I’m going to print them and my personal tolerance is far higher than I thought.  I should wear my glasses when I’m doing this!

Exercise: Addition

Aim:  The aim of this exercise is to create two different images producing a similar result but in quite different ways.  For the first image, take two shots of the same scene from a tripod so that they are identically framed.  The first should expose correctly for the sky and the second should expose correctly for the foreground.  Then merge the two images so that the resulting image is correctly exposed for both.

For the second image, do the same as above but using two different scenes so that a completely different sky is merged with a different foreground.  You need to take care that the lighting situations and camera angles are similar in order to get a believable result.

Approach and results:

This first image was taken at Abney Hall.  I put the camera on a tripod and chose an angle away from the sun so that I could get a deep blue sky.  Taking two shots, I exposed the first for the sky and the second for the hall.  I chose this subject as there is a nice clean and unfussy line to the building.  I made sure the angle excluded difficult edges such as trees.  I then used the quick selection tool to select the correctly exposed sky and copy/pasted it into the image with the correctly exposed hall.  You can do this any way round; it doesn’t matter which area you cut and paste.  The easiest option is to use the image with the best defined edge (usually the darker of the two images).

Image 1

Sky

Sky

Foreground

Foreground

This resulting merged image gives an even exposure.  The main point of this is that this is only making up for the lack of range in a digital camera and would be what you would have seen had you have been there.

Merged

Merged

Image 2

For this second image I used the same picture of the hall but then used the sky from a shot I took on the Denmark coast; the lighting and direction of cloud being very similar.  For this image I cut and pasted the hall into the beach image.  This is definitely altering reality but fairly innocently.  If you think about the original image of the hall, thirty seconds later the sky would have been different anyway.

New sky

New sky

original

original

Learning points:

There are definitely degrees of intervention.  This exercise is quite simple and innocent, after all, you are changing a subject that would have changed in a short time anyway.  It definitely helps to have clean lines and particular attention needs to paid to the lighting conditions and direction if the result is to be at all believable.

Exercise: Enhancement

Aim:  The aim of this exercise is to enhance a close-up portrait and make a series of adjustments; general facial brightness and contrast, lightening eyes and changing the eye colour.  At which point do you feel you have tampered with reality?

Approach and results:

I thought I’d add a little twist to this exercise by using a self portrait.  Photographers can be quite ambivalent about changing other people but how would I feel about changing myself?  I sat by a small window with a black background behind and took a portrait that included the front and side of my face; the side being more in shade.

original

original

The result (above) is a fairly average portrait that is a little underexposed for my liking.  Next I used the brush tool in Adobe camera RAW to lighten up my face, dragging the brush over the area I wanted and then adjusting the brightness and contrast sliders to change the highlighted area.  This brightened up the portrait and gave it more depth giving an overall improvement (below).

brightness and contrast changes

brightness and contrast changes

Next I zoomed in and highlighted just the eyes (iris, pupils and whites of the eyes) and brightened the highlighted area.  This gave a further improvement to the image (below).  At no point did I feel like I was tampering with reality.  Why?  Because the human eye picks up a wider dynamic range than a camera so in my mind I was only showing what the eye might see anyway.

eyes brightened

eyes brightened

In the final image I have changed the hue of the iris to a fairly strong blue.  At this point it looks digitally enhanced and not particularly real (below).  This kind of technique is used in advertising shots all the time but the shot has to be a more stylised, brightly lit shot producing a more graphic style image.  To do this on an ordinary portrait looks contrived.  Am I saying this because it is me?  If it was someone I didn’t know would I think differently?  I don’t think so.  This is a very unsubtle change that could have be acceptable if I had toned the effect down slightly.  This is, however, definitely tampering with reality for me.  No matter what the subject, changing a colour so that it no longer resembles the original is going beyond reality.  I’m not saying it’s wrong in any way at all, just that it crosses that boundary for me.

eye colour change

eye colour change

So what it I changed it to B&W; is that tampering with reality?  I think the answer is that someone viewing a colour image generally assumes the colours are faithful.  B&W images do not try and represent a false reality, they just leave colour interpretation to the viewer.

Learning points:

I think I have a clear idea of when I consider I am changing reality (rightly or wrongly).  I also don’t have any issue with doing it from an artistic point of view.  I don’t want to confuse artistic intent with the ability to use Photoshop.  Manipulating a woman’s image down to a size 6 and smoothing the skin to almost blank for the cover of a magazine is marketing not artistic ability.

Exercise: Improvement or interpretation

Aim:  The aim of this exercise is to take a portrait, isolate the subject and enhance the subject in any way that you feel is appropriate.  Following this, consider the limit that you would accept for this to remain an innocent, legitimate adjustment and reflect on this.

Approach and results:

For this exercise I have used an image I captured in Shanghai of a small boy struggling with a large map in the wind.  To isolate the subject I tried two methods; the quick selection tool and the magic lasso tool.  I found the quick selection tool to be the easiest as it is easy to add or subtract from the selection.  You can also zoom in to an area to select it accurately.  With the lasso tool it is important to add anchor points as often as possible in case it deviates from the course you wanted.  If the area you want to select is too similar to its surroundings then you can temporarily adjust contrast etc. to create a differentiation.

Having made the selection I used the levels adjustment and the saturation slider to bring out the boy against the background.  On top of this I used the inverse selection to select the background and added a 3.9 Gaussian blur to very slightly soften the backdrop.

How far would I go in changing this image?  As I’m creating an artistic image here, I would have no real issue with making major changes; changing colours or content to suit my aims, keeping it realistic if needed but even this could be ‘pushed’ if I was after a more graphic style image.  If I was reporting something factual I would have more concerns about adjusting the image beyond what you would have naturally seen if you’d have been there.

Map problems 1

Map problems 1

Map problems 2

Map problems 2

Learning points:

Unless I am presenting facts I have no issues with manipulating an image to get the result I’m after.  I have taken shots in the past that I knew weren’t quite what I wanted, knowing that I can make the necessary changes.

If you wish to keep realism then subtlety is the key, especially with colour images.

Exercise: Alteration

Aim:  The aim of this exercise is to take a photograph and to deliberately alter it by removing something from it.

Approach and results:  For this exercise I’ve used a shot I took in China.  This is a good example as I waited for some time to get the shot I wanted but couldn’t get it because of the number of other people in view.  This isn’t surprising as the Forbidden City is a major tourist attraction in Beijing and the person I was photographing was a tour guide and unlikely to be alone.  She was only ever going to stay for a short period so I made the decision to take the shot and clone the people out later.  There are 3 extra people and an arm sticking out from behind a pillar in the bottom right hand corner.

I started by straightening the shot and making suitable colour and contrast adjustments.  I then used the clone tool to remove the people.  I did this by selecting the most appropriate areas to clone so that it was not obvious.  Another technique I used was to select the area I wanted to remove using the magnetic lasso tool.  If you clone into a selected area like this, the clone tool does not go outside the selection.  The advantage of this is that you can select intricate edges and clone around them if necessary.  Using this technique for the person in green meant that I didn’t clone over the sharp edge of the stone wall that she is stood behind.  This gave a clean result without needing a hard-edged brush or a steady hand.

The other people had a busier background.  Fortunately the repeating architecture allowed me to clone a similar area to the left of the people.  I was careful to leave all the right shadows intact.  I didn’t find it necessary to use the cut and paste options.

Before

Before

After

After

Learning points:  The main learning point for me is that you can take a shot that you know isn’t quite right and make it right later.  Ethically I have no problem with this image as it is what I wanted in the first place and it is only changed for artistic reasons.

Exercise: Correction

Aim: The aim of this exercise is to use the clone tool to make corrections to dust and lens flare and to discuss the justification of the corrections.

Approach and results:

Image one was taken in the South of France.  There are a couple of dust marks in the sky on the top left.  I Cloned them out using the clone tool set to normal and 100% opacity (image two).  The background is consistent so it was not difficult.  I chose to clone the area just to the right of the marks and chose an area of similar size with a feathered edge.  Are they really dust marks?  Well, yes; I know this because I took many shots and the marks are on all the images.  Would I remove them if they were not dust?  Yes again; the image is artistically better for it and that was my reason for taking the shot.

Image 1

Image 1

Image 2

Image 2

Image three has lens flare to the right of the sun’s reflection.  I removed this using a two step process.  First I used the clone tool set to colour to clone an area just to the right.  This left the flare spots slightly light.  I then used the clone tool again but this time set to darken.  I eventually repeated this, dropping the opacity as the areas initially became too dark.  I think it is necessary to experiment according to the individual image.

I have only removed the flare as part of this exercise as I prefer it left in.  I new it was there when I took the shot and the shot looks slightly flatter without it.  You associate flare with sunshine and warmth which adds to the image.  Also the small colour accents break up the large area of water.

Image 3

Image 3

Image 4

Image 4

Learning points:

Using the different blending options and opacity settings meant that I could be very precise.  With a few attempts you could probably remove the lens flare completely.  I tend to use the clone tool as a blunt instrument (left on normal with some opacity changes) but the blending options give it far more versatility.

I can draw a distinct line between removing things that are purely aberrations and not part of the scene; dust being the obvious example.  If an artifact is not dust but is insignificant and blurred enough to resemble it then I have no problem removing this either. There would be rare occasions where I would think twice such as a war reportage photo being used as evidence.  In fact any photo being used as evidence should be unaltered.

Lens flare is not in the actual scene; we do not see lens flare when we look towards the sun.  However, removing it is more of an artistic decision as often is is deliberately captured.  This depends on the individual image, sometimes it can be more intrusive or just in a bad position.

Exercise: Strength of interpretation

Aim:  The aim of this exercise is to demonstrate that you can be much more aggressive when interpreting the tonal range in B&W when compared to colour.  To do this take two images and treat the first to a strong increase in contrast and the second to a high key or low key adjustment.  Create versions in colour and B&W and you should find that the effects can be produced far more strongly in B&W.

Approach and results:

Contrast:   This first shot of the Rhino shows how pushing the contrast also exaggerates colour.  This becomes unnatural after a point whereas in B&W the tones become lighter or darker but the viewer has no perception of any exaggeration until much later when mid-tones start to disappear and you are left with extremes of black and white.  In this example I have been able to push the contrast much further in the B&W image.  You can see this by the very dark shadow and loss of detail around the Rhino’s feet.  It suggests a less evenly lit environment where a sneaky peak underneath the Rhino will go unnoticed – it didn’t!

SONY DSC SONY DSC SONY DSC

High Key:  Again here creating a high key image has an earlier effect on the colour shot.  Colours start to become less saturated which is far more obvious to the viewer than the lightened tones of the B&W image.  The B&W image has been processed much further than the colour.  Because of the processing the B&W image suggests much stronger sunlight and helps visualize the brightness and heat of the Midday sun in Southern China.  It feels like you’ve just walked into the bright sun from indoors and your eyes need to adjust.

original brightness colour brightness

BW brightness

Learning points:  Unfortunately the effects do not look as extreme on the low-res versions of the images on the blog.  The original hi-res versions show more extreme results.  It is however quite clear that B&W images are far more tolerant to strong interpretation and that you can use this to create or change perceptions of the image.