Friday 10 December 2004

Miniature photography - part 3: post processing

After parts 1 & 2 about [lighting](http://www.nirya.be/snv/ttm/archives/000035.html) and [camera settings](http://www.nirya.be/snv/ttm/archives/000038.html), this final part in the miniature photography mini series will deal with post processing your pictures in a photo editing program. While some might think of editing photographs of painted miniatures in a photo editing program as cheating, I hold the opinion that as long as you only improve the _photograph_ and not the _subject_, one is still within the realm of the acceptable. So, improving the tonal range and general quality of the photograph is alright, editing away a painting mistake or blurring parts of the miniature to hide imperfections are not. Enough said on that subject.

Processing pictures in a photo editing program needs a photo editing program, obviously. I use [The Gimp](http://www.gimp.org), but other popular choices are [Adobe Photoshop](http://www.adobe.com/products/photoshop/main.html) and [Paint Shop Pro](http://www.jasc.com/products/paintshoppro/?). All of these three programs offer largely the same functionality but hide it away in different menus and toolboxes, under different names. As I realise that the Gimp is not the most used program out there, I'll try to keep the descriptions in this entry general and not use Gimp-specific terms so that people with other programs can find the equivalent tools in their programs. Unfortunately, that also means that this will not be a step by step tutorial, but rather more of a hazy overview - if you want a tutorial, you can do much worse that hopping over to [Cool Mini or Not](http://www.coolminiornot.com) and having a browse through the articles section (_after_ you've read this entry, of course).

To illustrate the operations described in this entry I will use a photograph of a recently painted figure of a Celtic warrior, as featured on my [Flickr Photostream](http://www.flickr.com/photos/robartes) [here](http://www.flickr.com/photos/robartes/2061940/). Throughout the article, I will illustrate the various manipulations I usually do on the photographs, starting with the original photograph straight from the camera (clicking on the thumbnail, and any thumbnail in this article, will lead you to a larger version of the stage):

The original image

This is the original image, taken against a white paper background with a single overhead light source. This photograph needs some work: obviously, the cork the figure is mounted on as well as its unpainted base need to be snipped off, and the single light source also leads to some problems, notably an uneven distribution of highlights and shadows, and some noise as a result of a general low lighting. We'll correct most of these problems in due course, starting with framing the figure through cropping.

The image cropped

This step is where you frame the miniature. It is as simple as selecting a rectangular region around the miniature and using the 'Crop image' (or some such) function - note that this is different from the regular crop function, in that it crops the image and resizes the canvas so that the image is no larger than the cropped selection. Regular cropping cuts away the remainder of the image, but leaves the image at the same size, resulting in large amounts of nothing around the cropped region. In this step, the cork is removed, but the figures' base is still present - that will be edited out further down the road.

The image despeckled

The next step I usually take is to apply a despeckling filter. This filter removes specks and grains from the image - for a digital image, these specks result from a too high ISO setting or too low light, causing image 'noise'. Not much noise was present on this particular image, but the despeckle filter did reduce the contrast at the edges of the figure a bit, notably around the spearshaft, which is nice too. The general effect of this filter is now to 'smooth' the image a bit (which is only logical, as the despeckle filter blurs some pixels, only less so than a full Gaussian Blur or some such filter).

The figure isolated from the background

In the next step, the background is deleted so that we can add a different one later. To do this, I have used the 'Magic Wand' tool, which is a tool that selects contiguous areas of similar pixels (both Photoshop and Paint Shop Pro (I think) have this tool, although possibly named differently) and is one of the most useful and magical tools in the Gimp toolbox. This is also one area where having only a single light source can bite you - a single light source means that the background has some variation in shading, not much but enough to make the magic wand tool not select it all at once. Luckily, you can easily add to a selection (in the Gimp by holding the shift key while clicking). After most of the background is selected, the freehand selection tool is used to adjust the selection a bit, and the selection is then inverted (to get the figure instead of the background), feathered (to get a border), and floated and anchored to a new layer. The original background layer is deleted, so what remains is the figure on a transparent background. Those of you _not_ using Internet Explorer will see this transparent background on the picture to the left (well, they will see their browser's background, of course), but IE does not do transparency in PNG (boo hiss), so IE users should see a grey background.

Colour tools magic

This step is where the _real_ magic happens. Althoug you would not tell on first sight, there are some problems with this photograph, mostly related to the use of a single light source: there is an uneven distribution of highlights and shadows, with the highlights at the top of the figure being almost washed out, and the lower half of the figure being generally too dark (the light source was overhead). The colours themselves are mostly fine and represent the actual colours, thanks to the setting of the white balance on the camera to match the lighting. In this step, we will correct the colours and value (lightness) distribution of the image. This is done through adjusting the 'levels' and 'curves' of the figure. Going into the detail of this would take us too far, but I basically lifted the shadows and midtones up a bit and toned down the extreme highlights, both through the curves tool, and stretched the value distribution across the full range of values using the levels tool. The result is a much improved image, I hope. I realise that this is all a bit theoretical, and might post an entry in the future with a detailed step by step rendition of this process, but for now, just saying that the curves and levels tools are among the most important in image post processing will be enough, and any aspiring miniature photographers should familiarise themselves with their use.

The final result

Finally, I add a new background by simply adding a layer below the figure layer, and filling it with a gradient from a medium grey to a light grey, with the light grey at the bottom. The resulting image is then exported to a jpg and is the final image to be published.

That's the end of this mini series on miniature photography; I hope it has been helpful to some. Finally, as with painting itself, the most important ingredients in miniature photography is practice, practice and practice. So don't be afraid to experiment - it might not look like anything but will help you improve your next pictures!

__Update__: turns out this is not the end of the series. There's a [part 4](http://www.nirya.be/snv/ttm/archives/000132.html).

* [Part 1 - Lighting](http://www.nirya.be/snv/ttm/archives/000035.html)
* [Part 2 - Camera settings](http://www.nirya.be/snv/ttm/archives/000038.html)
* [Part 3 - Post processing](http://www.nirya.be/snv/ttm/archives/000054.html) -- You are here
* [Part 4 - Stage setup](http://www.nirya.be/snv/ttm/archives/000132.html)

No comments:

Post a Comment