Skylight Publishing

Articles & Information

Helpful resources and information for all aspects of giclee printing

High Definition Photography
and it's application to photographing artworks

How does film relate to Pixels ? A bit of a background.

In the days when film photography was all there was, if you could talk about High Definition photography you would probably be referring to film format. This in effect refers to the physical size of the photosensitive film that light passing through the camera lens falls upon. It stands to reason that the bigger the film surface the more detail can be captured. This assumes the quality of the camera lens matches the possibilities made available with the larger surface area available on the film.

Most photographers would be familiar with the dominant 35mm film format, but there were several others, some smaller and some bigger. The ‘professional’ studio photographers would probably have used what is referred to as Medium format or Large format cameras. Medium format would have included the popular 2 ¼” square roll film cameras, while large format cameras used film plates like the 6”x4” plate, in relatively heavy cameras pretty much restricted to the photographers studio.

The point of using the medium or large format cameras was to capture more detail on a larger film surface, so that when printed the result would show more resolution and could be successfully enlarged further than 35mm film.

How does that relate to pixels?

Modern digital cameras have evolved to such a degree that now it can be argued that professional grade digital cameras can produce results that match that of film. That is largely due to greatly improved technology and increased sensor size (more pixels). Think of pixels as light sensitive points on the sensor at the back of the camera where traditionally the film would have been. The more of these light sensitive points (or pixels) you can pack into the available space the more information you can capture and then reproduce in print or on screen. It is in fact somewhat more complicated than that, and I can’t go into that level of detail in the scope of this article. The principle however is that the more pixels you can use the more resolution you can capture.

Current high end consumer digital cameras are offering around 16 Megapixels (a Megapixel = 1,000 pixels), and 8 to 12 Megapixels are more common. Professional medium format digital cameras use special digital sensor backs that fit on the back of the camera used, and these are now reaching up to 60 Megapixels (Phase One P64+ 60Megapixel digital back produces a 130 Megabyte file and costs around $40,000. And that does not include the cameras body or lens!). The cost of these top end cameras is out of reach of all but the most serious and successful professional photographers. [Note: with Nikon releasing the D800 camera in 2012 with a sensor size of 36Mpixels and a retail price of around £2,500 things have changed a bit. This is now the studio camera we use ourselves].

How to play with the big boys?

There is another way to produce 60 Megapixel images without spending in excess of $40,000. In fact it is possible to produce huge images in excess of 1 Gigapixel (a Gigipixel = 1,000 Megapixels). You would need a pretty powerful computer with loads of RAM for that sort of a project, but possible all the same. The world’s first Gigipixel image (of Bryce Canyon USA) was produced in December 2003 and was taken with a 6 megapixel digital camera. The final image is 40,784 x 26,800 pixels in size, and contains about 1.09 billion pixels...a little more than one gigapixel. And you can do the same.

How is it done?

How do you get a 1 Gigapixel image from a 6 Megapixel camera? You take many separate photographs of the scene, all overlapping by around 1/3 to ½ and then using special software these are then ‘stitched’ back together again into one big seemless image. The first 1 Gigapixel image of Bryce canyon required 196 separate photographs to reach that goal. But you don’t need to reach beyond 50 Megapixel stitched images to achieve what I refer to as High Resolution photographs, and they are far easier to produce.

Perhaps you will have spotted the one major restriction you will be faced with, which is the technique does not suit photographing scenes with moving elements, such a people, vehicles etc. While this is strictly not impossible to deal with, and I’ll explain more in a moment, the best scenes to use this process on are static scenes with minimal if any movement. Landscapes come to mind, or perhaps a Vulcan bomber parked being prepped for takeoff.

To produce high resolution images ideally you will need a sturdy tripod, and something called a panoramic head. A panoramic head is a piece of equipment which fixes the digital camera in a position such that the lens on the camera is situated directly over the central column of the tripod. Why do you need to do this? This is to overcome something called parallax error. I’ll explain what that is and why you need to deal with it in a moment, but I will say now that it is possible to produce successful results for certain kinds of scenes without either a tripod or a panoramic head.

Nuts and bolts.

The panoramic head allows you to move the camera from a fixed position to scan the scene in overlapping sections, and over multiple rows. A single row of overlapping images would produce a fine panorama, but we can use the same process over multiple rows, which are all stitched together later on the computer. There is a process required to set up the camera on the panoramic head so that light passing through crosses over at the point directly above the centre column of the tripod. This eliminates parallax error, which would otherwise result in foreground objects not quite lining up with objects in the mid and far ground between each adjoining shot, producing an imperfect final image. However I can say that if you don’t really have any foreground objects to worry about you can often get away with not bothering about parallax error.

When you take the photos you need to set the camera up in manual mode exposure or aperture priority. If in manual mode you need first determine the overall average exposure required for the scene as a whole and set that for all the shots. Take care not the set it so that some areas will end up badly over exposed or under exposed. If the scene has a high dynamic range this can be a problem, and you could try exploring the world of HDR photography to get around that – which is whole other topic in itself. I often combine the use of high definition and HDR photography to get the best results.

You also want to set the white balance yourself and not leave it to the normal automatic setting. This is to keep the individual photographs looking the same as possible.

Putting it all together again.

The process of stitching together again all the separate photos taken requires the use of software. There are several options available. If you have Adobe Photoshop CS4 or above you can try using the Automerge tool which can do a pretty good job, or you can use a specially written for the job package. There are a few on the market, but I would recommend you try Autopano Pro. It is beyond the scope of this article to provide a software user guide, but there is good on-line support available and the process is fairly straight forward. It’s certainly a lot easier now than it was when I started doing this 7-8 years ago.

If you don’t get a good result it will probably be because a) you didn’t allow enough of an overlap for the images you took (allow between 1/3 and ½ to be safe), or b) you had objects moving too much between photos, or c) you forgot to set white balance on the camera or d) you had parallax errors.

Pixel size versus print size.

People often get a little confused over the relationship between pixels and print sizes. A normal jpeg file out of a consumer digital camera will typically be output at 72 pixels per inch. If the sensor size was 2560 pixels wide by 1920 pixels high (a 4.9 Megapixel image) it would print 35.5” x 26.6” at 72 pixels per inch (2560 / 72 = 35.5). But you would not try to print at a print resolution of 72 ppi. That is a standard resolution set for computer displays. If you did you would see very obvious pixilation (ie: it won’t look good). Typically a good print requires 300 pixels per inch. So if you divide the pixel width of 2560 by 300 you arrive at a print width of 8” (in our example a standard 6”x8” print). If you halved the pixel per inch resolution (that’s 150 ppi while keeping the pixel width unchanged) the print would now be twice the size, ie: 16” wide. If you have the right software such as Adobe Photoshop you can make the changes to print resolution quite easily.

What many photographers don’t realize is that printers have their own optimal print resolution, and for Epson printer that is actually 360ppi. And if you halve the print resolution to 180ppi you will double your print size and still get a very fine quality print. In fact most people would never know the difference the quality is so good. But you have to keep the proportional change to exactly half, or exactly ¾ to get the best results. Using this method you can easily double the size of your prints with no loss in quality observable to pretty much anyone.

If I end up with a stitched High Resolution image that is say 13,200 pixels x 5,400 pixels (or 71 Megapixels) and I print it at 300 ppi then it will print 44” x 18” without needing any enlargement process, and so retaining its' full resolution even at that size. It would have been like having a film plate for a huge studio camera that was 44”x18” and making a direct contact print onto photo paper that size, without using an enlarger. And, there is no professional digital camera available on the market today I know of that can output a 71 Megapixel image for you (and when one does come along – which it probably will, you can bet it will cost a small fortune). So – there you are. You can play with the big boys and create stunning high resolution photographs with your consumer digital camera.

Have a go.

See an example from a collaboarative project here.

copyright: Gordon Burns

Resources Worlds first 1 Gigapixel photograph
www.tawbaware.com/maxlyons/gigapixel.htm


Software you can use very inexpensively or for free:
www.tawbaware.com/


Autopano Pro stitching software:
www.autopano.net/en/photo-stitching-solutions/autopano-pro.html

Inexpensive panoramic head for those on a budget:
gregwired.com/pano/Pano.htm
or in UK www.red-door.co.uk/pages/productpages/panosaurus-redshop.html


Artist and Photographers using our services...
The quality of the prints is excellent, with really good colour, and much better than anything I've had previously from other bureaus... Thanks once again Gordon for a truly superb job you made of my prints. We are so pleased
The quality of prints on the archival art papers I get from Skylight Publishing has transformed the way I feel about my work and the way others relate to it. I really appreciate the time and care Gordon takes to ensure I get the best possible results.