Choosing a Digital Camera

What is the right camera for you? The answer to that depends very much of what your intended use(s) will be. The perfect, single camera for each and every application does not exist.

If you merely want a camera that is ready to shoot everyday scenes with friends and family and a bit of scenery from your vacations, you have a lot of pocket-size automatic point-and-shoot high quality cameras to choose from. You would probably not want to mess with imaging software except, maybe for printing, so just shop around and find a camera that provides nice sharp pictures, evenly lit and with a colour rendition and contrast that meet your personal taste.

However, you might want more than that - such as taking detailed pictures of nature, either close-up or far away. You may have special interests in arts or architecture. Maybe you would like to portray your collection of stamps or coins. You might fancy shooting  portraits of people; you might want to capture the beauty of the sky - day or night; you.......

In such events, there are more to consider before you make your choice. An easy - too easy - recommendation could be: "Just buy a quality digital single lens reflex (SLR) camera. With their huge list of interchangeable lenses and accessories you can meet any requirement". But is that so? What about portability? Will you carry it along wherever you go and use it as much as you could like to be able to?

Portability is obviously a matter of personal choice and should be put on our list to consider. Let it be stated right from the outset: I don't intend to discuss the merits of the highly sophisticated digital SLRs - wonderful instruments as they are. I care for the compact but more advanced camera that I can carry along under almost all circumstances; that is ready to shoot whenever I am and which I can use for both everyday point-and-shoot and for more special applications.

For such a camera there are certain things to consider that you do not find readily in the documentation you get along with your purchase. Nor can you expect the average sales person to be able to give you guidance on the more subtle ones of the issues that I am going to point to as things you should consider. Therefore, without being too technical (and I am not a technician) we should start with a brief look at what the camera actually does, when it takes a picture and stores it on your memory card for later use.

 
What does it do?
We all know that a camera consist of a lens and a shutter that determines how much light is sent to "some pixels" where the photographic image is generated and to a screen (the "LCD") at the back of the camera where we can see what our scene looks like. How much light reaches our "pixels" is determined by the ambient light available; how big the  aperture is set (the "f-number") and for how long time the shutter keeps the lens open, ("the exposure time"). Now, in most consumer cameras the "pixels" are small elements in a so-called charge coupled device (the "CCD") that have the capacity to build up an electric charge proportional to the light that hits them. After exposure, our camera scans each element and reads the amount of charge that was built-up during exposure. This can then be converted into a grey-scale image. However, we also want to have colour. This is - speaking again of the most typical consumer cameras - achieved by putting a filter in front of each pixel in one of the three basic colours, (red, green and blue) which, when blended in different proportions produce all of the millions of colours available in the final digital image. The sensor basic principle looks like this:
   

Each pixel builds up a charge that corresponds to the amount of light of its particular colour filter. You will note that there are two "green pixels" for each red and blue. This is partly for manufacturing reasons but also corresponds well to the fact that the human eye has it highest sensitivity at the green wavelengths of light. In the red end of the spectrum we don't sense infrared light (as some snakes do) and in the blue end we don't see ultraviolet light (as some insects do). And, by the way, most CCDs of this type sense infrared light very well. Just point your camera with lens open and LCD screen turned on towards your TV remote control and activate that control - then you will see the flashes from the control easily on the LCD.

The infrared pulses are readily recorded on "digital film". Here you see the flashes from the Olympus RM-1 remote control as recorded by a Minolta Dimage G400 compact camera.

Further, to make a long story short, there are standard mathematical methods through witch the information of incoming light on one particular pixel plus the information on the incoming light of the other colours on the neighbouring pixels are converted to colour information (wavelength = colour plus luminosity / brightness) on each individual pixel.

This information on colour distribution over the CCD, together with the overall information on the luminosity and colour of the ambient light, constitute the elements of our internal "raw" colour picture. The procedure and hence, the "raw result" is more or less the same for all cameras based on this sensor type, irrespectible of brand and other design details.

But before the camera can download the picture to the memory card for later viewing or printing it has to do quite a few more manipulations and here, different approaches by different manufacturers begin to come into play:

What does it do next?

In brief, and in layman's terms, the camera does some or most or all of the following, depending upon the final output format(s) that will be provided:

 

1. Substitution for defective pixels

No CCD is 100% perfect, and data from known malfunctioning pixels will be substituted by interpolation from neighbouring pixels.

   

2. Noise reduction

The sensitivity varies a bit from pixel to pixel and small fluctuations are normally smoothed out at this stage.

   

3. Transformation of image colour

    contents

Depending on the geometry of the CCD and other factors, the "raw" colour information will  be transformed to represent rows and columns of colour information (as what you have in the final picture).

   

4. Colour transformation

A sensor element will in general respond differently to light of the same intensity but of different colours and in this step, transformation of the colour information - as obtained in direct response according to the specific CCD's sensitivity characteristics -  will be transformed to an operational, internal colour management system ("a colour space").

 

Hubble Wide Field Camera CCD and Human Eye: Here is an example of the efficiency of a CCD measured at different wavelengths. By "efficiency" is here meant the ability to accumulate charge relatively to the amount of light of a given wave length that falls upon the CCD during exposure. For comparison, I have inserted a colour diagram for light in the visible part of the spectrum, which for the normal eye lies somewhere between 4000 nm (violet) to about 7000 nm (deep red) of wavelength. Also inserted is - in arbitrary units - a curve showing the sensitivity of the normal, human eye This is the lower curve that peaks in the green area around 5050 nm. (1 nano meter = 10 Angstroms = 0,000000001 meter).

Ordinary CCDs for consumer cameras may be somewhat different, but the CCD- curve illustrates well that the CCD responds differently to light of different wavelengths and that corrections have to made in order to get a non-biased representation of all colours.

It is also quite obvious that the CCD "sees" the world in quite a different hue than does the human eye. Our eyes respond most strongly to the green light; the CCD (at least in this case) most strongly to deep red into the infrared light.

 

5. White balancing

This is correction for the ambient light colour ("light temperature") that was used to take the photo. High clouds and shadows provide a bluish light; the sun a yellowish light; tungsten lamps a reddish light and fluorescent lamps a yellowish-greenish light. Most cameras have an auto setting plus provisions for pre-setting corrections for common light sources as per the above. More advanced cameras also have a manual mode for calibrating the camera response for any, particular light source.

   

6. Colour depth reduction

Most cameras work internally with 10 - 12 bits of colour  information (levels from 0 up to 4095) for each colour. This is reduced to the common 8-bit level (8 bits for each primary colour: 0-255, where 0 = no light at all and 255 = full brightness). You will recognize these levels from almost any colour imaging and painting software when you pick different colours in or for an image.

 

If you are unfamiliar with the representation of colours in 255 levels in an RGB colour space, you may click on the rose to see a practical example. Here shown in the PhotoImact12 digital imaging software).

 

Last Rose of 08

Pentax Optio 550 picture at 1/87 sec,

27.8 mm focal length, f/4.2 in macro mode

 

What do these figures about the internal workings mean to us, who just want our pictures out of the camera without the need to be software engineers? Well we end up with a picture that may at most contain 256 X 256 X 256 different combinations of red, green and blue levels - in other word some 16.7 million different colours. However, the camera started with 1024 X 1024 X 1024 (10 bits per colour) or maybe even 4096 X 4096 X 4096 (12 bits) of different combinations of levels. Were all these extra levels useless??? No, they represented a higher "true" dynamic range of the CCD from highlight to deep shadow. So, if done properly, by discarding (in a so-called non-linear manner) the "trivial" parts of the colour information but keeping the "right" parts we may end up with a picture much richer in detail in both shadow and highlight than would have been the case had we just started out with gathering colour on an 8-bit level. And that is what interest you and me as users - that this capability for improving the overall definition of the final picture is being utilized by the camera designer in an intelligent manner !

   

7. Colour transformation

The internal, working "colour space" c.f. 4. above is now being transformed into a standard colour space that is being used for the output picture file format, (e.g. the sRGB system for JPEG-files).

   

8. Compression

We are talking fairly big amounts of data. My trusted, old Olympus C50-Z boasts (we are talking a 2002-release) a 2560 x 1920 pixels = 4.915.200 pixels. Each pixel carries 8 bits (1 byte) of colour information for each of the 3 colours, meaning that the uncompressed output file will be 14.745.600 bytes = 14.44 MB. Actually, this camera does provide recording and output of uncompressed files as an option, but in the more normal SHQ (super-high-quality) mode output file sizes are only a bit more than 2 MB large. This reduction - at the cost of some loss in detail - is what compression is all about. It is done by built-in mathematical algorithms that represent colours of similar identity as groups of pixels rather than as individual pixels.

 

 

Anything else?

Certainly. Most manufacturers have additional features built-in, the effects of which we can see through experience and comparison, but of which we have little knowledge. There is limited or no information available in the public product documentation as to exactly what these features are and how they are implemented. All we can be sure of is, that they most likely differ from brand to brand. Some of the most common features are:

 

9. Contrast enhancement

Most people tend to favour pictures with vivid, rather brilliant colours and the right "snap" in contrast and clarity. The camera manufacturers tend to respond to that by guessing what degree of additional enhancement will sell the most to the average customer

   

10. Sharpening

Same as above. Standard algorithms for sharpening are applied before the images are being stored on the memory card.

   

11. Additional noise reduction

Longer exposure times produce noise in the CCD elements and noise will normally also be amplified if the electronic (ISO) sensitivity is increased. For that reason, built-in algorithms are often applied to reduce noise. In most cases camera documentation tells little or nothing whatsoever about "when", "how" and "how much".

 

Orion Rising, December 08

(Image somewhat excessively brightened for clarity's sake).

Automatic noise-reduction: The designers of my Minolta Dimage G400 clearly did not foresee that their special noise reduction procedure (algorithm automatically applied, but only for longer exposures) should one day be used on a 15 second exposure of Orion rising over the treetops one December evening around 8 o'clock. Click on image to see in more detail the artefacts introduced.

 

 

12. Other corrections

May be applied (again unknown to the user) for corrections relating to design particulars of the optical system (distortions, false colour fringes/ chromatic aberration, darkening in the corners / vignetting......)

   

13. Special ("scene") settings

Today there are often scores of settings available for "special occasions" such as night scenes, landscapes, birthday parties, weddings, water scenery, Christmas, Summer, Spring,.......you name it......). Some of these settings may influence the settings of the camera prior to exposure (e.g. Sports / Rapid Motion and Night scene settings) but most are presumably just various kinds of post-processing enhancements through application of some standard built-in algorithms in the firmware.

 

 

Output formats

Finally just a few words on output formats. There are basically three of the kind:

 

 

"RAW"

RAW is not a file format as such. Each manufacturer uses generic raw formats and file extension names. Anyway, "RAW" is an output just after step 1. or 2. above. Only, data is also transferred to a format where it can be exported to and processed in certain, external computer software. Thus, RAW is not (as of today) a standard format, but rather a family of output formats, each giving the user the opportunity to work on almost pristine data. That is, data that contain the maximum of information that you may ever get access to. Therefore, great detail and subtle tones and shades can be reached that are otherwise inaccessible in the normal in-camera processed images. But one has to work on each and every picture although some of the manipulations may be performed in a batch process. For good reasons, "RAW" is often regarded as the digital equivalent to analogue negative film.

   

TIFF

As discussed above under 8. this is the processed but otherwise un-compressed RGB-data. Files are large - take the megapixel-size advertised for the camera and multiply by three. Most modern cameras do no longer offer this output option but rather RAW as above together with JPEG as per below or - as is most normal for ordinary consumer compact cameras - just JPEG.

   

JPEG

This is the most common format used. It offers the possibility for quite efficient compression before quality loss becomes too visible. Normally, there are more options such as "Super High", "High" and "Normal" or "Standard" quality. The difference may be as well in the level of aggressiveness in compression as in actual reduction in the dimension (in pixels) of the image. "Increased Aggressiveness" here means that more and more pixels of different shades of are represented as belonging to the same group and as such, shown as a one, single colour when re-opened in imaging software or printed - that is: Increased loss of detail.

   

EXIF

This is not an image format, but both TIFF and JPEG standards include a standard for incorporation of camera and exposure information in the output image file. You do not see this information in your picture, but any reasonably useful digital imaging software is designed to read it.

RAW files include similar information plus a small thumbnail for preview of the picture.

   

Selection criteria for YOUR camera

With this "small and light" introduction we should be properly dressed to discuss selection criteria that go over and beyond the standard "megapixel sales talk":

 

A. Quality optics

This is simply the alpha and omega. Aggressive post-processing and so-and-so many pixels are no substitutes for mediocre optics. Quality optics involves the use of sufficiently many individual optical elements in carefully designed groups to effectively combat optical distortions at all zoom- and macro settings. Try it out on linear / regular geometric subjects and study the output critically. Quality optics also means essentially no chromatic aberration (colour fringes). If you add on additional optical elements as I do (for astrophotography purposes and for copying negative and positive analogue film), inherent problems may increase dramatically

   

B. Quality optics

It just cannot be stressed enough. When I got my Camedia C-900 Zoom in 1999 we had just crossed the magic 1 megapixel barrier and state-of-the-art was barely approaching 3 Megapixels as I recall it. Yet, I remember that my humble 1.3 Megapixel Camedia outscored competing cameras with more pixels in sharpness and clarity due to its very fine optics, (with 8 lenses in 6 groups) . By the way: My C-900 cost my family (it was a birthday present) around 800+ EURO - in 1999 price level !!! I could get quite a few deluxe cameras for that sum today.......

   

C. Optical Viewfinder

Yes, that denounced little thing that only shows 80% of the captured image and exhibits parallax at close range - I just wouldn't be without it. You will miss it in the glare of the sun (even if LCD displays have improved over time), you will miss it if you shoot in low light and you will definitely want to to turn the LCD display off in a number of special applications such as stellar constellation photography.

   

D. Optical Zoom

By today's standard, 3 times optical zoom is pretty much a (minimum) standard and so it should be. You don't have interchangeable optics for compact cameras and if you want to shoot anything more than just simple point-and-shoot wide-field scenes you will soon want more. If you want to shoot candid portraits, sports or similar, you should consider something in the 5-7 times optical zoom range. If you want to capture a bird at 30 yards distance - forget about standard compact cameras and go shopping for a camera system with genuine telephoto capabilities. (Which could be a combination of a decent compact camera and a spotting scope - see examples on the astro pages of this site.

Almost any zoom system will exhibit some distortion(s) at certain range(s), but some lenses are better than others. Preferably, you should try it out before you buy. (Se more under "macro capability" below).

   
E. Macro Capability Most consumer cameras have a macro setting. If macro shooting (of flowers, insects, stamps or whatever) is important to you you must try out your camera candidates before you buy. Firstly, the macro capabilities of some cameras simply aren't what I would call "macro" - the closest allowable distance is simply too large for you to get any decent "magnification". You should normally be able to tell that from the camera specifications (and if the manufacturer does not tell, you may have extra cause for suspicion); however, and secondly distortions may be very grave at close distances and/or large zoom ratios and this you should preferably test yourself before you buy.

When you test for possible distortions, you should try out the optics against subjects with straight lines and right angles. You will readily see if a stamp is distorted, whereas a close-up photo of a flower or an insect may cheat you at first sight.

   
F. Size does matter As cameras become ever smaller and in particular slimmer, the focal length of the optics usually also decreases - simply because the distance between lens and CCD becomes smaller. Now, big lenses (: large physical aperture) with short focal lengths are costly to produce in good quality and you will note that as the focal lengths have gone down, so have in general the physical sizes (aperture/diameter) of the lenses. For shooting of normal scenes this does not really matter, as the "fastness" or "slowness" of an optical system - i.e.: how long exposure times are required for extended objects of a given brightness - depends  not on the absolute physical size of the system but of the ratio between physical lens diameter/aperture and focal lengths. (The so-called f-ratio).

Thus, a typical 3 X optical zoom system of today has an f-ratio around f/3.2 in the wide-field range and around f/5 - f/6 at the 3 X zoom range - which is apparently - but not quite - the same as digital compact cameras had in the earlier 2000 years, even though those cameras were usually quite bulky by today's standards.

However, the smaller lenses are not completely as good as the larger in a number of other aspects:

  • Light gathering power: What is said above about exposure times only holds for extended objects. For point-like light sources such as stars a larger lens will record more stars than in a smaller one given that exposure times and CCD sensor sensitivities are the same.

  • Resolution: Larger lenses can split objects closer apart than small lenses can. Thus larger lenses will produce finer details and crisper images under otherwise identical set-ups and circumstances. Again, this can be of importance in astrophotography in splitting close double stars; it may also be of importance in macro photography and quite generally in greater enlargements of your pictures.

  • CCD sensor size: A simple geometric reflection will tell you that while you preserve the same field of view in your images the physical size of the CCD sensor may be reduced as the focal focal length is reduced - just as it took longer focal length lenses for medium format (6 x 9 cm film format) cameras to cover the same field of view as that of standard consumer (24 x 36 mm film format) cameras. This again has implications for actual pixel sizes and thereby for noise and light gathering characteristics, but this is a rather lengthy discussion that has therefore been referred to a page of its own.

   
G. After sales support Now that we have covered the most important aspects of hardware, namely the optics, let us divert for a moment and talk about something much more "soft" and immaterial: How well do the various manufacturers and their local agencies support their customers?

If you read camera reviews from camera owners (i.e.: not professional reviews) on the internet, you may quite frequently find statements such as "I love it" and "The best camera that I have ever owned" side by side with statements such as "I hate this camera" and "I shall NEVER buy a camera from XXX again". Are consumers really that different or, is it more likely that those who hate their cameras have had certain bad experiences when they needed after sales support?

Let us face it: There have been produced a few truly lousy cameras and a number of mediocre models that should never have been allowed to reach the market. However, in general it is impressive that so many hundreds of millions of cameras have been turned out at lower and lower prices (both relatively and in absolute figures) at such a comparatively high quality. But there will always be a few malfunctioning cameras that slip through quality control and there will always be some of us who need repairs after some time for one reason or another.

Here, the manufacturers warranty and service policies will be put to test and, from my own experience and from what I have heard from both dealers and owners, there are indeed substantial differences between the manufacturers (or is it rather their regional representatives???). When the need is there, it is not irrelevant whether it will take 4 weeks or four months to have one's camera repaired - or, whether it will be serviced at all even after just a few years of ownership.

My trusty old Pentax K1000 - bought in 1981 - was produced and supported for more than 20 years. I wonder if any of my current digital cameras (they are all about 5 years old now) will be supported 15 years from now ?!?

This is the downsize of the rapid development and the race for constant releases of new models with more and more impressive features: The camera model that you bought a few years ago has already been discontinued, spares and accessories are no longer available - you cannot even buy a proprietary battery or a charger from the manufacturer any more..............

On way of getting an idea about this aspect of after sales support is to visit the manufacturers' official homepages - and check against the local representatives' do., if applicable: Search for discontinued camera models and see, to what degree such are mentioned, what service is provided and what spares/accessories are available for these discontinued models. This site has not been made to recommend or rate one brand over the other, but the Olympus home pages are very good examples of, how this could/should be done.

Also, I have heard and read  that even within the warranty period, manufacturers tackle warranty claims very differently. Unfortunately, facts are not readily found on this subject and myths and facts are not readily discerned - my best advice is to try find an honest sales person. To me, after sales support is one of the most important specs of any camera brand and should be of concern for everybody: For manufacturers if they wish to maintain customer loyalty;  for reviewers if they want to present really good, unbiased advice to their readers;  for potential buyers before they end up "hating their cameras"........

   
   
 

PhotoHome

   
   
 

Copyright 2007-2008 - Steen G. Bruun