Digital Experiments

From visit to France in 2003, black+white darkroom prints were scanned and digitally re-worked as a means to learn some of the tools available. The works have been grouped into:

• Monochrome Rework
  
• Simple Tints and Tones
  
• Hand Tints and Minor Effects
  
• Major Effects
  

Monochrome Rework

In the station picture the telephone cable in the top left corner was removed with the background eraser. Having gone white sky on the shot dodging it out in the enlarger would have been possible, but this would have complicated other dodge and burn steps. The essential 'antique' feel was achieved via a plug-in filter – this allowed both the colour cast and the blurred overlay to be applied in a single transformation.


Digitally edited pictures can quickly become ‘muddy’. The best approach is to experiment to find the required 'feel' and then to start over and to achieve that 'feel' in the minimum number of steps: The principal of minimum transformations. Of course, sometimes when starting over you find that the precise effect cannot be re-created – but that is irrelevant, providing the final image is both acceptable and reproducible. If the effect cannot be reproduced then the photographer is at the mercy of the tool, the end result is some abstract approximation to that which was (attempted to be) crafted. If other people’s software is used to edit the image, but those edits are not reproducible then a degree of the ‘artistic merit’ belongs to the creator of the software not the creator of the image.

The tower was handled identically, in approximately 1 hour (as opposed to 3 hours spent on picture 1). Because the starting images shared a common style and subject type it was pre-known that the final image would suit the treatment. Learning, and sticking with, a finite set of digital techniques not only reduces the image creation time (to something manageable) but also allows the photographer, not the software, to claim ownership over the result: The principal of developing aesthetic style.



The group picture was simply retouched, with many dust speckles spotted out. At a zoom scale much greater than that at which the image is intended to be viewed the clone stamp tool was used with the sample point a fraction outside of the brush area. The spotting was done with single ‘dabs’ and never with brush strokes. The proximity of the sample point to the brush head means that great control can be exerted in picking precisely the right de-spotting tone (i.e. from neighbouring pixels). The de-spotting process took around 30 minutes and should be the start point for all digital manipulations, i.e. before other effects are layered in – since any new layers from selections must not propagate dust speckles (Start with a clean base-layer) and it is only with the starting layer that we can guarantee the tonal selections will match the rest of the image (i.e. prior to any tonal or colour balance manipulations).

The final image was very challenging due to the heavy background detail, and brought all of the above methods into play. The subject was lit with strong side lighting, placed inches from a wall, and sealed in a glass fronted room – there was no way to deal with the wall at time of shooting. It was removed with the background eraser tool – attenuating the rigging in the process. Creating a layer copy of it and blending that layer using ‘Multiply’ blend mode restored the rigging. This seems to go against the principal of minimum transformations, to take away and then add back in, however there was no choice. By minimising all other transformations there is room in the image data-quality to ‘get away’ with this tricky manipulation.

Simple Tints and Tones

 The purpose of tinting and toning black and white images is not to make-up for the film’s inability to capture colours, but rather to complete the photographers interpretation of the subject. In the Beaurainville fields I wanted a strong panoramic effect, by creating an unusually wide shot. By capturing some detail in the sky on the horizon but letting most of the sky-part of the frame ‘white-out’ I was able to create the desired shape. But then the whiteness of sky meant that the picture ‘bled’ into its background. A frame and mount can fix that, but I do not always frame my pictures. A border in the print would also have worked, but I didn’t want to make the picture feel artificially constrained, so colour in the sky was called for. A very simple magic-wand selection was made along the horizon and the picture thus split into a ‘sky’ layer and a ‘land’ layer. Separate curves-adjustment layers were linked with each to provide the colour casts. It is exceptionally easy to split-tone a shot like this. The power of adjustment layers lies in the fact that they impact the image quality only once, on final rendering, it is possible to return to the adjustment layer settings and to experiment with the effect without degrading the image, irrespective of how many times the parameters are tweaked.
Colour replacement techniques (over-painting) have to be right first time, adjustment layers do not.


The ferns were even easier, the work was simplified by virtue of the fact that the original image supported the technique to be applied. The small patch of sky was printed totally white, so that within Photoshop it could be selected by a ‘colour range’ selection set to ‘pure white’ only (and in fact, within a rough-cut polygonal selection area that was applied first). The selection was then inverted (i.e. selecting everything but the sky) and a new layer generated. A gradient fill layer was inserted behind the ‘ferns’ layer so created. The fill angle was set so that the whitest part of the sky would coincide with the brightest patch of ferns to the rear, the scaling of the gradient was also set to bring the white-ish part of the sky to that point also.

The church window was the easiest effect of all. A simple radial gradient fill layer was set behind the image and the ‘blend if’ attribute of the window-layer modified to ‘do not blend if this layer is black’. Often this mode of blending is undesirable because leaks occur through the layer in unexpected places, leading to the nastiest of digital artefacts. Once again, work during the analogue stage  help to enable the final effect, processing and printing for high-contrast ensured that the containing wall and window leading were a true black so that the leak-through could be precisely set.

The tower used another plug-in, Digital Film Tool’s 55mm Split tone. The limitations of fill layer bleeds finally led me down this path, although I believe that heavily feathered split layers with curve adjustments could achieve the same. What is important is developing predictable, reproducible techniques. In the same way that polarising filters on-camera is not ‘cheating’, neither is using a selection of high-quality Photoshop plug-ins.
 

All of the foregoing gives rise to the next digital workflow principal: Ideally the end result is planned from the point of exposure, for the analogue preparation can support the end result.

Hand Tintsand Minor Effects

The real-world window that gave rise to this picture presented two main problems: I was not tall enough to shoot it without converging the verticals; The colouration was too vivid and variable for isolation of any subject. The entire base-layer was adjusted using the edit->transform tool. This is the first example of using Photoshop to ‘fix’ a ‘dodgy’ image. Note that the potential fix was anticipated whilst shooting. This is why the central panels (containing the queen/warrior figure) were not cropped more tightly – space was required for the adjustment. Less than anticipated has been lost and I could have cropped the final image tighter, but I didn’t so that the comparison with the original would be more evident. The colouration was added by over-painting; since each painted area was small I knew the risk of performing multiple transitions on a single pixel would be small. The paint mode was ‘Colour’ and a graphics table with stylus was used to draw-in the effect.

  Having  used the split-tone plug-in for the earlier Basilica shot the same method was used here for the corner² and aisle³ shots. The colours used are identical in all three shots. The next Photoshop principal being Plug-ins provide preset-save options in order to support the principal of developing aesthetic style, so use them. For the aisle shot I also over-painted the stained-glass windows as per picture 1. For the corner shot I found that there was an unexpected distracting highlight at the centre. Here another plug-in (Mystical Light) was used to add a single light-source beam-cast which not only tones-down the distraction but is also angled to meet up with the natural highlights on the floor.





 
The gate is a much more complex split-tone, there is no linear boundary between the two desired tones. I considered using magic-wand to precisely define the gate (or in fact a path selection would have worked well), but I did not want to simply colour the picture in, I wanted a toning effect with some degree of bleed. I also noted that the tree-trunks, which should be brown not green, would benefit from picking-up the rust-brown toning of the gate. So a deliberately rough (and feathered) selection was made by adding multiple marquee-tool square selections to subtract portions of the greater gate-square. Mostly the bleed works well, possibly less so at the top-left of the gate, but I still believe that a very accurate selection would have rendered a less ‘believable’ result. The intent is not to recreate some reality, but rather to add mood. Digital manipulations work best when enhancing, rather than ‘correcting’.

However, the next section takes digital manipulations into the realm of ‘creating new realities’ where the intent is neither to enhance nor correct what has been caught, but rather to construct a new graphic image. Here the digital manipulations predominate rather than the original source materials…

Major Effects

These images are designed as graphic statements composed from photographic elements, rather than representational photographic images, and as such a much freer reign is afforded. They are the result of experimenting with the source materials and manipulation methods available, rather than the execution of a pre-planned workflow. Much of the foregoing has concentrated upon the need to pre-plan in order to maximise the final image quality whilst attaining the desired result. Image quality cannot be sacrificed for the sake of it, and having experimented it still makes sense to rebuild the final desired image from scratch with the minimum number of steps. However, a great advantage of the digital workflow is that ‘undo’ is only ever a keystroke away and so experimentation is greatly facilitated. I noticed whilst de-spotting some of the earlier pictures (a process taking maybe half an hour) that the convenience of the digital workflow has a big impact on what can be achieved. The difficulty of de-spotting a 35mm negative, or of recovering from mistakes, is so great that it is really only something done in dire emergencies; not so for the digital workflow. So the final operating principal has to be: Master the techniques; develop a strict, workable, repeatable aesthetic skill; But
take time out to experiment, be prepared to throw away the rules and any vestiges of reality – at least, on occasion.

There is little point in deconstructing the execution method for these images, so rather I shall list the principals raised throughout:

• The principal of minimum transformations
  
• The principal of developing aesthetic style
  
• Start with a clean base-layer
  
• Ideally the end result is planned from the point of exposure, for the analogue preparation can support the end result
  
• Plug-ins provide preset-save options in order to support the principal of developing aesthetic style, so use them
  
• Digital manipulations work best when enhancing, rather than ‘correcting’
  
• Master the techniques; develop a strict, workable, repeatable aesthetic skill; But take time out to experiment, be prepared to throw away the rules and any vestiges of reality – at least, on occasion
  

Exposure Control

The difference between a ‘good’ and ‘fine’ print
is in the retention of detail in the deepest shadows and highest highlights.
Once the basic exposure for a print has been determined it is usual to apply
several steps of dodge or burn in order to balance the detail in the shadows
and highlights. This can become a most compulsive obsession in the darkroom.
It is not unusual to execute a dozen or more dodge and burn steps for a given
print.
Each step is applied for a given time, and if the print
is to be reproduced all those times must be recorded, and each individually
recalculated if any of the exposure conditions change. One reason exposure may
change is if the lens aperture is adjusted to give longer exposure times, so
that the dodge and burns are more manageable.

Longer dodge and burn times are easier to manage because
there is usually a short (1 second say) ‘error’ when synchronising the
position of the mask with the operation of the enlarger. This error has a
smaller impact when the dodge or burn time is longer.

So if you decide to perform a
dodge for five seconds you may well stop the aperture down to give a ten
second dodge – remembering at the same time to double the base exposure.
Sounds simple enough, but with up to a dozen such steps the calculations
interfere with the artistic thought processes that are on-going at the same
time.
The whole story is similarly
complicated if the enlargement factor is changed, or the contrast filtration(s),
or the paper batch. Considering Ansel Adams’ ‘how to’ series,
Adams
states

‘…it is useful to think of a dodging or burning-in time as a percentage of the total exposure. We can become quite proficient at estimating the effect of dodging or burning an area for, say, 20 percent of the total exposure time.’

Adams, A (1995), The Print, Vol III The Ansel Adams Photography Series, (pp 102). Little, Brown & Company

In Adam’s ‘Clearing Winter Storm’ example his ten
dodge and burn steps use only six different times. There are two –2s dodges,
one 1s, one 2s, one 3s, one 9s, and three 10s burns, with a 2 to 4s
corner burn. The 9s burn is performed in three 3s sweeps and two of the 10s
burns are performed through a hole rather than by a simple mask (see fig. 1).
The image of Adams’ hands literally dancing in the enlarger light stream through all these steps
is compelling and helps to illustrate his philosophy of the print as
performance.

Figure 1 Burn Map for the Ansel Adams example
(showing duration and types of dodge/burn steps)


The performance of the print, the subjective decisions
about exactly where to place the mask, by how far to overlap the surrounding
area, and exactly when to end the operation, are the factors that should be
consuming the attention of the photographer. It is this process that makes each
print unique, and it is uniqueness that is the wet darkroom’s unassailable
strength over the digital workflow.

It is worth then creating a method of work that simplifies
the numbers, mental calculations, and feats of memory required in the print
performance, so that the creative mind is at greatest liberty to realise that
performance.
Using percentages helps in this respect, especially when
the exposure must change, but there are still many different actual numbers to
remember, the duration of each dodge or burn step. Fortunately the process can
be further simplified.
First of all, in practice there is a minimum dodge-or-burn
time that produces an effective result. This will vary between different
material combinations but there are only a relatively few such combinations in
use by any given photographer. The minimum effective dodge or burn time for a
given paper/developer combination can be determined through experience. I
generally constrain dodge and burn times to a minimum of 5% of the base exposure
time with Ilford Multigrade IV™ RC Gloss developed in Multigrade Developer™.
Secondly, the precise dodge and burn time is not decided
rationally and absolutely whilst reviewing the proof print but rather is part of
the subjective performance of the print, within bounds. What this means is that
instead of thinking about three separate burns with three separate burn times of
1s, 2s, and 3s we can think of three separate burns that are around 2s. The
actual difference between the burns is decided on-the-fly during the exposures.
All three burn steps are applied with the enlarger set to 2s. The first burn can
then be given ‘about half’ the 2s exposure time. The second burn is given
‘most’ of the 2s exposure time. The third burn is definitely given one full
2s burn followed by ‘about half’ of this standard 2s exposure time.
The above scenario reduces the number of times that have to
be considered and the need to constantly change the enlarger timer setting
(reducing chance for error).
When creating a ‘fine’ print the subjective terms
‘about half’ and ‘most’ are counter-intuitive. It seems that such
approximations lack accuracy and so cannot give rise to a controlled final
result.
The subjective approach does however work. The absolute
‘inaccuracy’ is controlled by the size of the standard burn or dodge step
chosen. The 3s exposure above will be given at least 2s and no more than 4s, and
so will not be that inaccurate.
Also, it is unlikely that a difference between, say, 3.9s
and 4s would yield a perceivable difference in the print. The dodge and burn
steps need not be applied with any accuracy greater than, say, 5% of the base
exposure time. For a 10s base exposure the effect of a 3s burn can be achieved
anywhere between 2.5 and 3.5s.
Most especially, the subjective approach works because the
precise, final decision of how much of the second 2s exposure we give to the
third burn is taken as the exposure is being applied. The photographer is not
concerned with resetting the enlarger timer, only that the third burn should
receive one full standard exposure and one partial standard exposure. During the
partial standard exposure the thought processes are entirely concerned with what
was perceived from the proof print, and the subjective effect desired, and how
well the operation has been carried out so far – in real time.
This constrained-subjective method is further supported
when one considers that there is no such thing as ‘the final print’. The
printing of a given negative will always change as it is informed by the
development of the photographers’ style. A constrained-subjective approach
creates a workflow where the print is given the opportunity to reveal something
new that has not been previously, consciously considered – by its very
subjective nature.

Figure 2
Fixed Increment example burn map (note
regularity in time & subjective language)
The ‘standard exposure time’ for dodge and burn steps
is determined on a per-print basis and is a function of the base exposure time
and the minimum effective dodge or burn time. Performing all dodge and burn
steps with a single enlarger timer setting (or multiples thereof), instead of
perhaps a dozen different timer settings, greatly simplifies the printing
process.
Once the proof print has been evaluated a decision is taken
regarding the relationship between the different dodge and burn amounts required
and the responsiveness (i.e. sensitometry) of the paper/developer combination.
For a base exposure of 60s with a 50% sky-burn and a 25% central dodge, for
example, the enlarger timer would be set to 15s. The base exposure would be
achieved with 4 ‘standard exposures’ for the print, during one of which the
dodge would be applied. Two further ‘standard exposures’ would then be made
to burn the sky area. The entire print is made with only one timer setting.
The principal is also useful for split-grade printing. If
the entire print may be given 80% at grade 5 followed by 20% at grade 0 the
timer would be set to 20% of the base exposure and applied 4 times at grade 5
and once at grade 0. Eliminating the need to reset the timer reduces the
opportunity for error.
Such a method should not be applied thoughtlessly or
slavishly. As the number of ‘standard exposures’ required for a dodge or
burn step increases the chance of miss counting also increases, so eventually
the method becomes more complex than continually adjusting the timer. The
important principal is to establish some method that aids and simplifies
repeatability within the confines of the print’s sensitometry and the
photographer’s creative freedom. As a worked example the figure below shows
the dodge and burn map for the photograph at the head of this article.

Because of the extreme latitude I originally found the
exposure for the main window through one test strip and the left-hand wall
through a second. These suggested 45s for the window and 9s for the wall. I
decided to work at a fixed 9s exposure increment, with one increment for the
base exposure. A series of additional 9s fixed increment exposures were applied
as per the map.
The base exposure need not have been a single increment. In
this case the smallest burns were 100% of the base exposure. Had they been 20 or
50 percent then the base exposure may have been made from 5 or 2 increments,
respectively. Also if I had needed to dodge an area I may have arranged the
fixed increment in order to provide an entire exposure step for the burn
operation and one step with no planned dodge or burn. Planning to have a dodge
and burn free single exposure increment creates space for last minute
adjustments. Usually applied as the final exposure increment this allows for
last minute tweaks based upon the printer’s immediate reaction to the print as
is has just been performed.
The fixed increment exposure method is much simpler in
practice than in theory. By design the method is imprecise. The purpose is to
give rigour to the numeric process of exposure so that there is a well-defined
framework for the creative process. This means that some liberty can be taken in
setting the fixed increment value. If the numbers aren’t as convenient as in
the example they can be tweaked. Each increment will be used in an ‘about’
way. The base exposure however should be an exact number of fixed increment
exposures.
The benefit of the fixed increment exposure method is that
during printing we need only count exposure increments whilst concentrating on
the precise nature of the dodge or burn. In the example, the +4 burn of the main
window illustrates this. How far into the image this burn extends, and for how
long the burn maintains its maximum reach, is the on going subjective thought
process during the burn operation. The outcome affects the floor in a major way.
The feel of the image can be entirely controlled by the amount of burning
to this floor area.