Last Sunday, in a manifestation of the arrogance for which I am known and loved, I entered a painting competition. For a £15 entrance fee, I went with a friend to Pittville Park in Cheltenham, and spent six hours painting. The competition brief was to paint anything inspired by what you see in the park, representational or abstract. Here’s the view that caught my eye:
That is the Pump Room framed by the trees, one of the most regal buildings in the town. Since I had plenty of time to work — I’ve never spent more than two or maybe three hours on a painting before — I started out by doing a page of quick studies to feel out the space. I wanted to figure out three things: could I represent trees adequately? Could I capture the essence of the Pump Room in paint alone without doing a drawing first? And could I capture the white band-stand in negative space?
As you can see, the answers were no, no and no.
I started with a foreground tree on the right of the sketch page. (This is not visible in the photograph above, being further to the right.) I was reasonably happy with the trunk and the darker lower branches, but it all fell apart as I ascended the tree, with all that white space looking very silly. At this stage I didn’t really know what to do, but I pushed on by trying to make a contrast with the tree behind it. In the past I have always struggled to differentiate different trees, since they are all basically green; this time, I consciously went for a faded, misty look. It sort of works — the trees are distinct from each other — but it looks too distant to be part of the same picture, and putting a more well-defined building behind it looks a bit silly.
The second study was of the building itself. It quickly became apparent that it’s too geometrical to paint on the fly, and that the columns particularly need care and attention. The general proportions of the building here are all wrong: it’s much too tall relative to its width. And the whole thing is far too big compared with the trees. On the positive side, I was happy with the dome and especially with the small colonnade that crowns it.
Finally, on the left, is my attempt to paint around the negative space of the bandstand, and it can only be termed a catastrophic failure. Maybe it could be done with a good enough preliminary drawing, but doing it directly with paint was not going to work.
As weak as all three of these studies were, I did learn from them. Most obviously I learned that I needed to draw the building before painting it. I also learned that I needed a composition that omits the bandstand, and to minimise the depth of trees that I include. That led me to choose a fairly tight framing of the Pump Room itself, as in this crop of the photo above (lightly edited to make it better match the proportions of the images that follow).
As you can see, the bandstand is out of the picture to the left, and there is only one tree on each side. Having figured out that this was what I felt reasonably confident trying to paint, I made a simple sketch just to get the shapes and proportions about right:
It’s not perfect for two reasons. One is that my eye was off, and I made the building relatively longer and lower than it should have been — though not by a great amount. Secondly, I stupidly forgot to bring a ruler, so the best straight-edge I had was the lid of a Chinese-food tub that I’d brought my sandwiches in. For that reason, I had to draw long lines as sets of short lines end-to-end, which means they are not as straight as I would have liked. Still, it got the job done.
Finally it was time to start painting! First up, a basic wash for the building:
Here is where watercolour, in some ways such an unforgiving medium, really comes into its own. A rough, uneven wash like this conveys the essence of an old building much better than an even one would, so I actively welcome the variations in colour intensity across the walls.
As you can see, I avoided painting into the areas that I had reserved for the trees to the left and right, and I erased the lines within those areas.
I also avoided painting into the lower windows. I now think this was a mistake, as I created extra work for myself and made things messier than they needed to be, given that I was going to overpaint those lower windows much darker anyway. (Of course, not painting into the upper windows was the right choice.)
The next stage is where things start to feel real:
There are new five things going on here:
Most important, I think, are the overlays on the original wash. There are several protruding ridges running horizontally across the front of the building, and rather than pedantically picking them out I suggested them with shadows underneath them.
Second, I suggested the upper-floor windows with minimal reflection/shadowing, deliberately ignoring the internal lattice structure of the windows. Learning from my overworking in the previous project, I carefully avoided revisiting the windows in the later stages, so they retain this fresh look in the final piece. I also suggested window sills and lintels, and parts of the frames.
Third, I added the two statues. (There is actually a third one on the right, but it’s obscured by the tree.) I got lucky with the central one: accidents of paint left it looking really nicely three-dimensional. When I did the one on the left, I failed to replicate the success of the first one. Here’s a close-up of the good statue:
Fourth, I put in the two roofs on the side parts of the building.
Finally, I filled in the shadows around the pillars in the bottom half of the building, still leaving the windows blank for now. This did not work out at all well. I originally intended to have the shadows strong at the top, just below the overhang, and fade out further down. But that was hard to do well — and of course in reality they don’t just fade out, they have specific hard edges where specific shadows fall. And painting that accurately would be a nightmare, because the shadows would move even as you were painting them. In short, my whole approach to this was wrong.
In the next stage, again, several different things happened:
Maybe most obviously, I painted in the dark windows in the lower half, leaving only a few suggestions of reflection and a few sparkling lights in the window to the left of the door. Once I had done this, it became even more apparent that what I’d done with the shadows between the pillars was inadequate, so I extended the shadows to ground level. Because I did these in two separate passes, they ended up even messier than they would otherwise have been but it actually looks sort of OK as part of an old building. Part of what brings this alive is the slight shadows under the capitals at the top of each pillar, which also went in at this point.
I also sharpened up the top left corner of the building, added the plaque below the central statue, added some more depth below the horizontal ridges and started out on trying to apply a bit of texture to the top part of the frontage, before deciding it wasn’t really working and leaving it.
But the part of this stage I like most is the dome, which I think I captured pretty well. The slight darkening at the left and right edges, the subtle shading between different tones, the hints of highlights, and the colonnade on top all work pretty well:
At this point I had a real stroke of luck. One of the judges, Roger Dellar, was wandering around the park, chatting to competitors and seeing how they were getting on. As you can see from his website, he is a proper artist — the kind I’d love to take lessons from. As we chatted about my painting I said that I knew I was going to ruin it when I added the trees that were pretty much the next step. Roger took ten or fifteen minutes to show me how to paint trees, by far the most useful bit of instruction I’ve ever had. (He also told me to get an easel and to use tube paints instead of pans.)
As a result, I was able to do this:
I’m not claiming that’s a great tree over there, but it’s definitely a big step up from the unstructured blobs of foliage I came up with last time. And an even bigger step up from the coloured-pencil effect of the fern/tree from my 1995 Greek holiday.
To summarise the key lessons: use bolder, less dilute colours; keep changing the colour; push the paint outwards with the belly of the brush rather than laying it down with the tip; take advantage of wet-on-wet effects to create indistinctness; use lighter, yellower greens towards the edges and the top, and darker, bluer greens further in. I even used reddish or brownish greens to suggest the trunk and low branches. All in all I am pretty pleased with it.
Next I put in the other tree, the concreted area in front of the Pump Room, and the lawn:
I like that the left-hand tree uses a recognisably different, browner palette than the right-hand tree, even though (as you can see in the photo above) that isn’t really the case in the real trees. This is a case where I think it’s good to edit reality to make a better composition.
In reality, the concrete area in front of the Pump Room has steps leading up to the middle section of the building, but I felt that trying to reproduce that would be fussy, so I went for a simple broad area, and a similarly sketchy lawn in front of that. The lawn is rather well maintained and an even vivid green, but I felt that if I painted it accurately it would both look boring and overwhelm the actual building with its highly saturated colour. Again: editing reality.
I could have left the painting at this point, and honestly I was very tempted to. I can’t paint people to save my life, and they’re really the only thing that was missing, The problem is made worse by the fact that a lifetime of looking at people leaves everyone with finely tuned radar for human shape, so if you get it a bit wrong people can immediately tell. But I really needed people to convey the scale of this very impressive building.
By experimenting on a spare bit of paper, I satisfied myself that it was possible to make adequate far-off human figures very simply. It suffices to use one colour for the lower half, another for the top half, and a suggestion of a head. Sometimes you might throw in some shoes or a backpack of something. So I held my breath and threw in fifteen people of different shapes and colours, one of them wheeling a bicycle.
If you look at the people close up, they don’t stand up at all, but I think that may be OK, You’re not really supposed to be looking at them, you’re looking at the Pump Room.
And that is the final version of the painting as I submitted it for judging. Needless to say, it didn’t win anything — I didn’t think for a moment that it would, given that there were professional artists taking part. But I think the discipline of knowing I was going to have to hand it in at a specific time was helpful.
So that’s it, another watercolour challenge done, and more lessons learned (see below). I’m keen to push on with this and see if I can get better at landscapes especially: I don’t want to be always majoring on buildings.
- When painting structures, you really do need a drawing.
- You can simplify your task by zooming in on the part of the scene you’re interested in.
- When painting a wash, go right across areas that you’re going to overpaint darker anyway.
- Think about shadows before you start painting them, and have a plan.
- Practice trees! (Details above.)
- Feel free to edit reality as necessary.
- Less is more when painting far-off people.
- I have looked around online for where I might take lessons locally, but most artists don’t use watercolour; and among those who do, they mostly produce abstract art of a kind that is of no interest to me, and which doesn’t demonstrate that they have the technique that I want to learn from them.