Google

 

 

     
 

Photographs of Paintings

 

A few technical web questions answered in English.
by Terrence Foley  
Copyright Western News Company (Chicago) all rights reserved


This article is in progress as of 09/30/08

Return to Hoffman Estates Art League

 

There are two main difficulties to be overcome when taking pictures of paintings.  There is a third, but it has to do with getting the computer screen colors to exactly match the colors of the painting and is usually of interest to no one, unless the color mismatch is making a particular painting look funny on the screen.  Ordinarily, because all the relative hues, tones and values change together from painting to computer, the small change of any individual color goes unnoticed.  Editing individual colors in the photo of a painting is the subject for a book and I'm not going to get into it here.

The two things to watch for when photographing a painting are Glare and Trapezoidal Distortion.  And by the way, this discussion is intended for the do it yourself, first time, painter, at home, who wants to sell a picture before spending any more money on equipment.  In reproducing images, better usually means more light and better camera, both of which can be sport for those who love to throw money.  But you can get perfectly fine results just making the best of what you already have.

Glare

The problem of Glare should be obvious to anyone who has ever taken a self portrait photo in a mirror using the camera's flash.  All you get is the glare of the flash and the rest of the image is so dark you can barely make out who is in the photo.  As you can see in the second row of illustrations below, if the camera and flash are aimed at the center of the picture the light will just bounce back and produce a large blob of glare in the center of the picture.  Photographing an artwork covered with glass will produce glare almost as bad as a mirror if the picture is taken with the camera and the light squarely in front of the artwork.  A painting with glossy paints or one that has been varnished is almost as bad.  Surprisingly, even an image done in flat (no gloss) pastels on paper will give an unacceptable amount of glare.

The only thing to do is to move the light source.  You can't move the camera or you will induce the Trapezoidal Distortion discussed below.  Naturally, never say never.  Trapezoidal Distortion can be corrected, but usually elimination is the better solution rather than correction. A general maxim in most kinds of this type of work is that you get the best results when you start with the best pictures.  One generally has the easiest row to hoe and the  best luck if one tries to conserve and improve what is already very good.  Trying to salvage drek is best left for when you have no other choice. As you can see in the bottom illustration below, if you put the lights off to the sides of the painting with the camera centered to the painting, the glare will bounce off to the opposite sides beyond the camera and all the camera will see is the soft diffuse light from the painting.

Eliminating glare is easy but a bit of work.  

  1. How much light do you need?

    • It is best if you can start with bright sun light streaming into the room.

    • If you're working indoors in a dark room, if you have at least 100 watts on in the room, then you will need at least 2 150 watt bulbs, one on either side of the painting.  2, 200 watt bulbs would be better;  4, 100 or 150 watt bulbs, on on each side and one on top and bottom would be best.  

      • 4 bulbs at 600 watts total should be plenty for pictures up to 4'x4'.  

      • As the paintings get bigger than that, it would help to get more light on the subject, but with less light on a smaller painting or 600 watts on a significantly larger painting, it is possible to adjust the camera exposure or lighten the image later with the computer.  Remember, always start with the best photo you can manage.

  2. If you can see glare in the painting:

    • Move the lights further out so the reflection is not visible in the surface of the painting any longer.  This is not the best solution because as you move the lights you get less light on the painting and a dimmer picture.

    • A better solution is to put a translucent cloth or paper in front of the light to diffuse the light.  This might  be unavoidable if, for instance, the glare you see is not a large spot directly reflecting from the light but lines and areas on the picture where the light is reflecting off glossy ridges in the paint, or areas of impasto which are so thick they are actually tilted and produce a glare.

    • It is also possible to produce more indirect, more diffuse lighting by using flood lamp bulbs which will fit ordinary sockets and shine the flood lamps at sheets or white panels which will reflect a soft yet bright light at the painting.  This is what you see on a movie set when they have all kinds of screens and white panels hanging all over the place.  They are trying to eliminate shadow and glare.

    • As you can see, each step here is more complex and could require more and more professional lighting equipment.  But it is still very possible to get professional results by laying a painting on the floor, surrounding it with goose neck desk lamps and standing over it with a camera.  Professionals will always try to innovate before they go out and spend more money.

Here is a diagram of the process which you can click to enlarge.  Please excuse the poor hand drawing.  I'll try to redo it soon.

Trapezoidal Distortion

The top row of illustrations shows how to avoid Trapezoidal Distortion.  I know that's a mouthful, but there are lots of different kinds of distortion so .......

One of the reasons old cameras have a bellows between the lens and the film is because if you're taking a picture of a tall building from the ground in front of the building, for example, you will get very severe distortion.  The top of the building will be very much narrower than the bottom.  The sides of the building will converge toward the top just the way the edges of a road do at the horizon if you take a picture looking straight down the road from the center of the road.

This is exactly the same image we get when we look at that tall building, but for a more accurate, pleasing image, our brain compensates and unless we concentrate on the converging perspective, we will percieve the sides of the building to be straight, rather than converging.  Looking at a building it is much more obvious to our mind that the sides of the building are parallel than it is when looking at a severely distorted photograph.

With a camera that has a flexible bellows, the photographer can snap a lever and the lens can be raised until it is higher than the film on some cameras.  When the picture is taken the sides of the building are much less distorted in the photo and the picture is much more pleasing.

The same thing happens when you take a photo of a painting if you do not keep the camera aimed perpendicular to the painting.  For instance, if the painting were mounted high on a wall, say 10 feet or more above the photographer as they are in some castles or museums, in the photograph, the top of the painting will be much narrower than the bottom and the sides will converge accordingly.  So when you take a picture of one of your paintings try to get the camera as square and perpendicular to the painting as possible.  Remember, it doesn't have to be perfect, but the better your picture is to start, the easier it is to make a correction on the computer.

Resolution

Resolution has to do with how small a detail can be seen in a photograph.  For instance, regarding satellite photographs, it is very common to speak of resolution that allows one to distinguish individual automobiles on the ground.  Two meter resolution, meaning you can see objects as small as 2 meters across would allow distinguishing individual cars.  One meter resolution would begin to allow distinguishing the make and model of individual cars.

Note that we are now discussing a third way to improve your photographs and we haven't even yet moved the pictures from the camera to the computer.

Resolution is an information technology problem.  For us there are two considerations about resolution.  First, the more detailed our pictures are when we post them on the web, the more someone is likely to grab one and use it somewhere else without respect for our ownership of the rights to this image.  In other words they will steal our work and not pay us for it.  There's a very fine comment on this subject by Harlan Ellison called Pay the Writer.  Ellison is a writer not a painter, but rights is rights and money is money.  Be warned, nothing obscene but the language is a bit salty.  

Second, web sites charge by the bit both for storage and transfer.  That means the larger your image is, the more it costs to store on the website and the more it costs to transfer to the viewer.  Also, larger images take longer to transfer.  This doesn't matter as much with a high speed net connection, but if someone is viewing your pictures on a dial up connection, it can easily take 20 minutes to transfer a picture with a size of 20x30 inches, 300 dots per inch and 24 bit color depth.  The size, 20x30 inches means how large the image would be if printed out using all the dots available in the stored image.  Obviously, if you are storing 20x30 inch images on your website you are wasting 30% when they are viewed on even a 20 inch screen and you are begging people to steal from you.  300 DPI means that every inch across the screen and every inch down the screen there are 300 dots of varying color and brightness that make up the image.  Newspapers are produced at about 72 dots per inch and if you think about it, all you need for a website is the same quality as you find in COLOR photos in your Sunday newspaper.  300 DPI is a lot more wasteful than you think because pictures are 2 dimensional which means you are paying by the square of 300.  300DPI = 90,000 dots in a square inch vs 5184 in a 72DPI image.  That's a whopping 95% waste.  Would you go to the store and pay $17.25 for potatos if they were going for $1 next door?

Color depth means how exactly the color is printed to the screen.  8 bit, 16 bit and 32 bit color are common these days.  8 bit color is only 256 colors and while that may seem like a lot, it actually gives a shabby picture.  At least in a side by side comparison and beyond that the picture may not display well for technical reasons.  For instance all the blues might come out as the same shade and in a picture that is all sky it could look very bad.  16 bit color allows for millions of colors and is necessary for images as delicate as paintings.  32 bit color is good for actual production work but is unnecessary for display.  Your eye cannot readily tell the  difference on a computer screen.

 

Copyright Western News Company (Chicago) all rights reserved.


     

     
     

 

 

         
         
* Compressed files are used to make file transfer times faster and more efficient.  If you are using Win98 or lower, you can easily download a free copy of a compatible unZip utility at www.7-zip.org
If you are using WinXP or higher, zip file compression is automatic.  
In current releases of Linux, zip file compression is automatic but I also indlude links to .tar.gz files.

*.PDF files are Adobe Portable Document Format.  The PDF viewer is available free from www.Adobe.com.
The PDF viewer is probably included with your Windows system and is standard in current releases of Linux.

         
         
    Top of Page    

Western News Company (Chicago)  2008 all rights reserved.