I Took That!

Blog by Jim Everett

http://www.itookthat.com

Blog Support Notes



CREATING LARGE PRINTS WITH DIGITAL IMAGES

Presented by Jim as a guest on The Lab With Leo Laporte TV Series

G4 Tech TV Canada, and How To Channel Australia

http://www.labwithleo.com


The latest digital cameras are packed with megapixels, and have sensors that provide very high resolution and quality. This means that it's possible now to make very large, high-quality prints from digital images.


Why make a large print?


- Posters for display

- Exhibiting and selling artwork

- Personal pieces for home decor

- Show a small subject in the context of a much larger image

- The overall subject needs a larger size to create the effect you want


Here are a few thoughts, tips and techniques on how you can get stunning large prints:


1. Determine print size


Megapixels give you resolution. This means how big you can enlarge an image before you see pixels. Definition is the clarity of the image, and includes both the number of megapixels and the sharpness of the image produced by the lens and interpreted by the sensor.


Generally speaking, a good quality 10 megapixel camera will produce an image that can be printed as a razor-sharp 13x19 without enhancing or boosting the size. With current image enhancement and size boosting software, it is possible to take images of good quality to much larger sizes.


One way to create the effect of a camera with a much larger megapixel sensor is to take three overlapping shots of the subject. To end up with a horizontal image in a standard 2:3 format, you would turn the camera on the side and take three overlapping vertical shots side-by-side (exactly the same as you would do when taking a panorama).


If you want a much bigger standard 2:3 image in portrait (vertical) format, you'd do the same but you would stack 3 horizontal images one above the other of your subject. It is important that you zoom in a little before taking the images to avoid lens distortions, and to create the effect of a "normal" lens.


Then using Photomerge in Photoshop (www.adobe.com), or a panorama program such as DoubleTake (www.echoone.com), you create a single larger image. Three images side-by-side or stacked will give you the correct proportions.


Photomerge in Photoshop also allows you to create a large number of stitched images, resulting in one enormous image. Of course, you would need a supercomputer and a massive industrial printer to do anything with it.


Another factor in determining the ideal print size, or how large you can make a print, is viewing distance. Generally speaking an 11x17 print is the largest print that someone would hold and look at as a handheld print. Any bigger, and people tend to step back. If people will be viewing the print from a great distance, you can make it large without worrying so much about definition. But if people are going to inspect a print up close, then the print can only be as big as the resolution and definition allow.


For large prints, the choice between RAW and jpeg leans strongly towards RAW. RAW gives you much more flexibility in processing, and more information and tonal range to work with.



2. Calibrate your display


Especially for large prints, it is vital that you calibrate your display with a device such as Huey from Pantone (www.pantone.com). You also need to understand and set your color management process in a way that works for you. We won't go into color management here - it's covered very well in other segments.  



3. Reduce noise


But before you start enlarging your image, it is important that you reduce any digital noise (grain) if you need to. You'll find digital noise in shadow areas from lower quality sensors will way you have used a very high ISO rating for dark or fast-moving subjects. Noise reduction is the first thing you do. Be very careful not to overdo it, or you'll lose details. You're better off having a little grain or noise and preserving the detail, then eliminating it altogether and having an image that looks plastic.


Photoshop CS3 has a good noise reduction function, but my preference is noise Ninja from PictureCode (www.picturecode.com). Noise Ninja has quite a range of sophisticated functions, and was developed by scientists, but still has a few simple and easy things that anyone can do to get rid of most unwanted noise without losing detail. It can be used as a standalone application or as a Photoshop plugin.



4. Correct lens distortions


Now that you have cleaned up your image, you can adjust for any lens distortions. My preference is LensFixCI from Kekus (www.kekus.com), and I use it as a plugin with in Photoshop. It has a database of just about every lens on the market, and automatically adjusts for distortions and vignetting (those pesky dark corners).  Photoshop CS3 also has lens correction tools which you can use manually (Filter>Distort>Lens Correction).



5. Correct perspective distortions


The next step is a matter of personal preference, for the order you do them. If you have a very powerful system, you can probably get away with doing your enhancements after you enlarge the image. Or you can do your enhancements before you enlarge the image. In particular, if you're distorting or correcting perspective in your image, that is probably best to do first. Or if you're using Layer Adjustments in Photoshop, or Layer Masks, then it is probably quicker, easier and more effective to do these before you enlarge.


With street scenes, buildings, or interiors, it is almost always worth correcting perspective and getting your vertical and horizontal straight. Photoshop has good tools for doing this. There is a whole lot to this, and it will be covered elsewhere.



6. Enlarge


Since you need to print at a resolution between 240 and 360 dpi (300 is the default), you may need to use image enlargements software to bring your image file up to this resolution. The two methods I use are: the Image Size function in Photoshop, and Genuine Fractals from OnOne Software (www.ononesoftware.com). Both have their own ways of handling lines, texture, and sharpening.


Genuine Fractals tends to do a better job with edges and straight lines, but requires more care with preserving texture. Photoshop tends to handle texture quite well, but does not have quite the same definition with lines and edges. However, both do an excellent job, and require a highly-trained eye to detect the difference.


Now, choose either Photoshop's Image Size, or Genuine Fractals, to enlarge your image. With Genuine Fractals, you can do it in one action. With Photoshop, I strongly recommend increasing the Image Size in 10% increments. This gives a much better interpolation, and finer end result. If you're printing at posters size, you can probably get away with 240 dpi resolution. Otherwise your image file might be too large.



7. Sharpen the image


The very last thing to do before you print your file is to sharpen it. There are a range of third-party sharpening applications. But I stay with Photoshop's Unsharp Mask, or even better, Smart Sharpen. Determining the size of the radius and the amount of sharpening is an art. Most people tend to over-sharpen, and this creates artifacts around the edges. It is better to be a little conservative at first, and build your skills in knowing how to sharpen.  Getting the right sharpening is vital to a good looking print. However, once you have sharpened, you cannot go back and make any major adjustments. That is why sharpening is the last thing you do.



8. Print


Now to get your image printed. Your image resolution should be close to 300 dpi (technically it is ppi). Home printing up to 13x19 is about the sweet spot. Larger than this, you need to be a professional or serious photographer, or send it out to a lab. There are many labs that of on-line services. One that offers excellent services in Canada is Silvano Imaging of Toronto (www.silvanoimaging.com). In the US, mpix.com seems popular, although I have not tried them Large prints are OK at 240 dpi, and this makes a smaller file to upload than 300 dpi.


An alternative to making a single large print at home is tiling. DoubleTake allows you to take a single larger image and split into tiles, which can then be printed on small sheets and assembled into a larger image.


You also need to choose the material that you will print on. Generally, the choices are glossy, pearl, matte, velvet, and canvas. My preference is Epson Velvet Fine Art Paper. Because the paper is porous, there is no ink dispersion. This means you end up with very high definition. However, the surface is a rather delicate and can be easily damaged by scratching or water. I use a photographic lacquer with a UV inhibitor - Suregard McDonald ProtectaCote Matte. This is a spray product, and the can is covered with health warnings. Use only outside in a well-ventilated area. There is a real art to using this product, and it takes practice and care to avoid getting a "frosty" finish on dark parts of the print. 


9. Next steps


Check out my June '07 blog post on www.itookthat.com, "Framed Decor from Your Prints" and download the PDF.

Listen to my audio podcast, "Frame Your Prints as Home Décor", in the series "Digital Snaphots" on iTunes.

-----------------------------------------


© 2007 - James Gordon Everett

James Gordon Gallery

www.jamesgordongallery.com