Home
Contents
Search
Up

Pixels and Dots - Colors

 





Colours
(Colors for those below the 49th)

Richard C. Pitt

Second in a series on Digital Images for the Professional

The first in this series of articles was in answer to a friend's question about what PPI (Pixels per Inch) meant in the real world. Along the way to answering this question, I brought up some other subjects that many people don't understand when dealing with the new digital world compared to older times and processes.

I've decided that, since I'm at best a talented amateur in all the areas covered at this time (I've been a professional in many but am no longer), I'll stick to giving you pointers on where to look for problems, leaving the details to those I reference below if you are interested in a particular aspect.

The first thing to understand about getting excellent colour output from anything is that it is HARD!

Don't despair if you don't get it right the first time, or the second or third. It may take you many attempts to get some pictures to print correctly on your home printer - and then take just as long to get similar results via your local photo lab. I'm still working at getting what I consider to be an excellent print from some of the photos I've recently taken with my new Nikon D70 - and I've run through several packages of paper an ink.

Step 1 - The Original

One of the most basic problems I see with folks using digital images is a complete lack of understanding of how they are created and stored digitally. There are just too many ways to do this for many to really understand how they relate to each other.

I'm going to cover the three most prevalent ways of getting a digital image to start with:

  1. a digital camera
  2. a scanner
  3. photo-cd from "real" film

The Digital Camera

Most people realize that the number of "Megapixels" a camera can take relates to how good the picture can be as well as how much space each image takes on their chosen storage media, but few really have an appreciation for just exactly how they, the image creator, can influence the results themselves.

The first thing to understand is that regardless of the camera's actual imaging chip size, the contents of the final digital file have to do with the settings you choose prior to taking the picture. The size and cost of the storage media used in digital cameras has lead some owners to choose settings for higher quantity of pictures on small media over getting the best image they can but far fewer pictures. Today however, the cost of digital storage media is so much less than it was a few years ago, that this should no longer be an issue; but it is hard to get out of the habit.

Rule 1 - set your camera to give the fewest number of pictures, not the most, and then purchase larger media to get more pictures.

Most consumer cameras "optimize" the picture prior to storing it and there is nothing that the owner can do about it. These cameras for the most part will only store the image in one of the compressed image formats; typically JPEG (jpg). The optimizations typically performed include:

  1. color temperature adjustment - making whites look right no matter whether the primary light source is the sun, flash, incandescent lights (tungsten), or fluorescent (daylight or warm) or candlelight
  2. sharpness - making the edges of things just a little bit more definite
  3. final file size - compressing the image so that it takes up less space - which coincidentally also loses details and numbers of colors

Some cameras, and all of the better ones, can also be told to store the image with no optimization and either no compression or only "lossless" compression - so called RAW output.

Prior to getting my first "real" digital camera, the Nikon D70, I researched whether or not I'd be able to deal with its raw format files on my preferred computer system, Linux. I discovered a wonderful utility called "DCRAW" that allows me complete flexibility in dealing with this proprietary Nikon format. Along the way I have learned that in fact, DCRAW does things better than the programs that Nikon supplies with the camera and which only run under Windows. Note that this program is also available for Windows :)

These raw files take up much more space than a JPEG file - sometimes as much as 10 times or more. As an example, on my 1Gbyte flash card my camera says it will store 95 raw images or 235 "fine/large" JPEG (and 3.8 thousand small/basic JPEG) even though they are in fact compressed too; but the compression is lossless, unlike JPEG.

From the raw image, using either the vendor-supplied software or something like dcraw, you have the ability to get the most out of the original photo you so lovingly took. In section 3, we'll look at what choices you can make and why you might make them.

Scanners

Many people today have a scanner available to them of some sort. It might have been purchased specifically to do document or picture scanning (creating digital images) or may simply be the scanner portion of one of the latest crop of all-in-one scanner/fax/printer combinations now hitting the market.

No matter how you got it, there are several things about it that will affect your results:

  1. its "native" resolution
  2. its color separation method
  3. what file format you choose for its output
  4. what other things it does that you may not be able to influence

The "native" resolution of a scanner is a measure of how many distinct pixels per linear inch of the scanned material the scanner measures in its work. In many cases the number of PPI is greater in one direction due to the different technologies used for X and Y stepping. The X (length of the paper) axis resolution is governed by how fine-grained the control over the stepper motors driving the image head is, and can be changed easily up to a maximum limit. The Y axis (width of the paper) is governed by how close together the photo array segments are on the carriage assembly that is moved along the paper - something that cannot be changed physically.

To vary the resolution the system achieves, the manufacturer will interpose some image chicanery into the results in many instances; boosting "effective" resolution by interpolating between pixels, usually only on the Y axis, or lowering resolution by averaging. As noted above in the camera section - you should get the scanner to give you a raw file, and you should make sure that if there is any manipulation to be done, that you get to do it so you can control the results.

As you will see in such photos as those at www.clarkvision.com where a 35mm image was scanned at 6000 dpi, scanners can be used to produce excellent digital images from film - but the file sizes are huge (137 Mbytes in this case). Doing such scans requires a scanner much better than typically found at your local electronics warehouse.

Commercial Photo CD

Most film processing businesses today offer you the option of receiving your photos digitized on a CD. Many have only two choices: consumer and professional. These essentially denote whether the resulting image is good enough for web publishing (the consumer choice) or for printing (the professional choice). The basic difference between them as far as I can tell from the several services I've used is the amount of compression applied to each file. The "consumer" pictures show themselves as the same number of pixels and same color mode (24-bit) but the file sizes are less than half as large with a corresponding loss of detail and color fidelity.

The problem with both consumer and professional from most places is that the scans have undergone image manipulation in the same fashion that the picture printer does for printing them - boost the shadows to provide "some" detail, flatten the overall dynamic range, and sharpen the image. On the typical 4x6 print this is almost unnoticeable - or may in fact be fine for aunt Tilly's snapshots, but it does nothing for your use as a professional wanting to get the best from the film.

If you want a set of digital image files from your film, find a shop that will guarantee they will only just scan the film, not try to enhance it. Many shops do their CDs from the same equipment they use to print, and literally can't turn off the enhancement. You will also want the shop to store the files as uncompressed or at least lossless compression; something like compressed TIFF is fine as this is a lossless compression type. Expect your files sizes to be multi-megabyte in size, not the 200-800k or less that they typically are from generic shops. Expect to pay more too. The process takes time but it is worth it for the quality. What you should get back should rival or better the output from a 6+ Megapixel camera. If you can't find someone to do it this way, then invest in a good quality scanner with a film attachment and do it yourself, the results will be again worth the time and effort.

Step 2 - On Screen Editing

I guess the number one thing to remember about editing at all is; Never edit the original!

For those who are working at being a professional, the number two thing is; weed out the bad images early - and delete them! Unless the image is of a subject that can't be re-taken, or has some sentimental value or value as a documentary of you or someone you love's travels, get rid of it if it is not in proper focus or shows signs of camera shake, poor exposure or other debilitating error. Don't spend time trying to make it better - shoot it again, better the next time, or work on something that has a good starting point instead.

If you're working from a raw image, you'll probably have to create an image in a format that your editing software can use in any case, but be careful - see the section on long term storage below.

Using a computer to edit your photos can be either or both of rewarding and frustrating. It's rewarding in that you can do a lot of things quickly that we old darkroom types took hours to do, if we could do them at all (and you don't have to get your hands wet and breath nasty chemicals)

It's frustrating for any number of reasons but I've found the number one reason people tell me it is frustrating is that they wait so long for some pictures to change once they've decided what needs doing and pressed <enter>. The primary reason for the wait is their computer doesn't have enough RAM - digital images take up a lot of working memory when they're being manipulated. The software may keep two or more copies of each image in RAM (the "original" and the working copy being built up as the change is taking place) and if you have more than one image open at a time, this simply compounds the problem. Note also that even though the image may only take up a couple of Megs of disk space (you are using raw mode aren't you!) it may well explode to 20-30 Megs when it is pulled into the image software because of the uncompressing and creation of the various colour layers in memory. Scans at high resolution can take 100's of Megs!

The tools you use to edit your photos also have an affect on the final product. At least one tool I've been fond of over the years - one called "xv" that runs under Linux - has one unfortunate side effect - it cannot save an image larger than the screen size it is working under. This means that if I'm working (as I often am) on an image that is 3038x2012 using xv, when I save it the maximum size I get is 1280x1024 if I'm working on one of my LCD monitors, or 1400x1200 if I'm on my Optiquest V95 monitor. This is just one of the things I ended up learning the hard way as it is not documented.

Regardless of this particular failure, some programs will try to help you in subtle ways if you don't turn off some of their default settings. Others will only do their best work if you start with or convert your image to their favourite digital format but... beware of conversions from one image type to another; for example from TIFF to JPG and then JPG to BMP. You typically lose something each time a conversion is made. If you must output a particular type of file, make the conversion as late in the process as possible, working with the original file type and color mapping to do your manipulations. Compare the results of similar manipulations to the same original file with format conversions done earlier or later to get a feel for what happens.

Another thing to do is "save early and save often" - something that will save you time in the long run as you do more and more manipulations of your images looking for "perfection". Note that this can use up lots of disk space and if you don't manage the file name conventions you use, you'll end up with all sorts of intermediary files that clutter the disk once you've finished. Disk space is now relatively cheap, but that doesn't mean you shouldn't conserve it at least a little bit.

I use a number of methods of tracking my progress on an image, depending on what it is I'm trying to accomplish:

  1. working from a first generation copy of the original, I'll make changes such as minor crops and sharpens for the web and save the new files as original-name+letter; eg. dsc-3092.jpg -> dsc-3092a.jpg, dsc-3092b.jpg, etc.
  2. working from the original (in this case NEF files from the Nikon) I'll run a command line utility (see below for how and what) that creates a series of TIFF files at each of several gamma settings, and names the files depending upon what setting it is; eg. dsc-3092.nef -> dsc-3092-0.55.ppm, dsc-3092-0.65.ppm, dsc-3092.0.75.ppm, etc. and then use a file viewer to select which one I like the best for each image. I'll then remove the other ppm files (note that ppm is a standard full-resolution intermediary file type for Linux graphic programs, similar to TIFF.
  3. working with a single file that I want to do the best job possible, I'll copy the original NEF file to a new, descriptive file name and work from this toward the final which will end up with the same name, having had several intermediary, numbered or lettered iterations. Eg. dsc_0726.nef -> girl_on_bike.nef -> girl_on_bike.0.55.ppm -> girl_on_bike.0.55a.ppm -> girl_on_bike.055b.ppm -> girl_on_bike.ppm -> girl_on_bike.jpg (for catalog) and/or girl_on_bike.tif (for printing)

Step 2A - document, document, document

In addition to saving files as you progress, you should learn to keep notes - notes on everything. Photographers who make their living with their cameras will typically document all the way from the camera's setting through to the finished print or other medium.

There are a number of "online" image bank programs available that allow you to take notes and keep them with the image. I use one called gThumb under Linux. It allows categorization as well as notes.

When I'm out shooting commercially I carry my bound notebook and keep track on a shot by shot basis. When I'm shooting for pleasure I'll typically note only major changes from my typical settings (ISO 100 and my normal lens on the D70)

When I'm editing and perfecting an image I keep notes on what processes and settings I use to achieve each step along the way. There may be a time in the future when I want to do similar things to another picture and it is far faster to read about it than to re-invent the process.

The only way you will be able to reproduce an image via printing 2 years after you've sold or given away your last copy is to know what you did and on which equipment. Needless to say, this assumes you've documented the paper source (type, make, finish, etc.)  and printer settings too; as well as which of potentially several versions of the image you used finally.

Step 3 - Choosing the Reproduction Method

Are you going to show the image on a web site, print it on a home printer, print it on a laser at work, or hand it to a professional to print in a brochure via web-offset? Each of these and other forms of output require different manipulations to achieve pleasing and visually "equivalent" results. Even the final size of the image makes a difference.

As an example, I tend to sharpen fairly hard any image that will be a "thumbnail" on a web site. In the small size, the hard edges make the image leap out at you a bit, where in a larger version of the same image they would be obviously false or exaggerated.

You'll likely have to compress the dynamic range of images that will be printed - depending upon the method being used.

Step 4 - Long Term Storage

In process - watch for it

Conclusions

For those of you who are simply trying to get good looking prints at home on your own equipment, the easiest way is to purchase camera, printer and supplies from the same manufacturer. The colour characteristics of each of the steps is tested and matched before you get the product, and you're pretty much guaranteed that any picture that looks good on the screen will look good on paper.

For those of us who need to go farther afield than just our own home pictures, the problems are a bit more cumbersome but with a bit of work we too can get good results within a reasonable time and dollar budget.

For those who are in the business of making colour picture containing products across many different media, the best way to get things to work consistently is to hire a professional color editor.

References and Links

bullet http://www.kodak.com/US/en/digital/dlc/book3/chapter2/index.shtml Color Theory - Kodak's Digital Learning Center - Kodak knows this subject better than almost any other company IMHO, but this series is fairly basic in nature
bullet http://www.boscarol.com/pages/cms_eng/ a much more detailed discussion of digital color management systems by Italian Mauro Boscarol - the first sections under "General concepts" are excellent! The balance are very technical in nature
bullet http://www.popphoto.com/article.asp?section_id=4&article_id=437&page_number=1 Popular Photography's online magazine
bullet http://www.tech.purdue.edu/cg/facstaff/jlmohler/dcm.html an online converter for digital values in RGB, CMY, HSB/HSV, HLS - interesting toy
bullet http://davis.wpi.edu/~matt/courses/color/ Especially take a look at the biology of color perception -
bullet http://www.clarkvision.com/imagedetail/index.html Film vs Digital, discussion of scanners and all manner of topics - includes some stunning photos by the author.
bullet http://www.dpreview.com/learn/?/Glossary/Digital_Imaging/ Brief explanation of terms used in and around digital photography

richard
 

 


 

Home ] Contents ] Search ]
Up ]

Copyright © 1993-2007 Richard C. Pitt - all rights reserved
Updated June 17, 2005