I know, this whole RAW or not to RAW debate has been hashed and rehashed on this site, so many times.  But I am still often asked by new and budding photographers on this site.  I shoot in RAW, I have my reasons and the simple answer is that it gives me better results, and the newbie often replies, well I get Great results with JPEG, so why do I need to bother with RAW and all the extra work it involves.

The explanation is fairly complicated, as one needs to know the whole process of how a camera goes from the sensor to the JPEG image file and where and how shooting in RAW will make a difference.

I finally made an effort and tried to summarize it logically and trying to keep technicalities to a minimum.  (I would be happy to expand on any question or technical aspect that I have further knowledge about).  First I will explain the internal processing of the camera, because without knowing that, you wont understand the real advantages of RAW  and then present my reasons why I shoot in RAW and where it affects my photographic results.  If you go through my portfolio, you will see that most of my photography is done in less than ideal conditions of lighting and time.


Inside the Camera ..

Here it comes .. I will try and make this as simple as I can and as general as I can.  

The Sensor

The sensor of your digital camera consists of an array of "cells".  The function of these cells is to measure the quantity or intensity of light that falls on it.  The cells measure or record only the intensity of light, and have no "colour" information.   This essentially says that your sensor is your camera monochrome.  (Here I am talking about the CCD / CMOS sensor found in most cameras)

So How is colour implemented ?  

Over each of these cells is placed a filter (generally) of Red, Green and Blue (RGB) .  Only one filter can be applied to a cell, either R,G or B.  
These filters are applied in a special pattern, usually its called a Bayer Pattern.  These filters allow Red, Green or Blue light to pass thru and the cell measures the intensity of Red, Green or Blue light falling on it.

Many people do know that the computer screens are clusters of R,G,B light emitters and they combine to form a single pixel of colour, however  the camera sensor is not like that.

Just an example of a bayer pattern is where R is Red, G is Green and B is Blue.  You can see that there are equal numbers of Red, Green and Blue cells.  The colour information is made is a slightly different manner than simple triplets of Red Green and Blue that combine to make a simple colour element.

GBGBGBGBGBGBGB
RGRGRGRGRGRGRG
GBGBGBGBGBGBGB

The information of intensities of different colours falling on cells is collated by the camera.  This information is was is contained in the RAW file of your camera.  This in itself does not make a viewable image.  This information is contained in usually 12-14 bits of data per cell.
A lot more processing is done after the shutter is clicked and before the next image can be taken.

First Step : Demosiacing

From the RAW  image the camera then extracts three images of RED, GREEN and BLUE.   These Images are just partial images of Red, Green and Blue.  X's mean that the colour value is unknown.

Green Image

GXGXGXGXGXGXGX
XGXGXGXGXGXGXG
GXGXGXGXGXGXGX

Blue Image

XBXBXBXBXBXBXB
XXXXXXXXXXXXXX
XBXBXBXBXBXBXB

Red Image

XXXXXXXXXXXXXX
RXRXRXRXRXRXRX
XXXXXXXXXXXXXX

The Camera then employs mathematical guesswork  and approximation  (algorithms) to fill in the unknown values of the imcomplete images.

These three images are then combined back into a simple image consisting of RGB triplets
(R,G,B): (R,G,B): (R,G,B): (R,G,B):
(R,G,B): (R,G,B): (R,G,B): (R,G,B):
(R,G,B): (R,G,B): (R,G,B): (R,G,B):

This process is demosaicing

Step 2: Colorimetric Adjustment

Colometric adjustment, basically is compensation of the measured values of colour against the values of the colour filter as compared to absolute RED, BLUE and GREEN. A simple mathematical adjustment that is specific to the actual frequency and bandwidth of the filters used.  (further mathematical explanation is long and tedious) and then the image is mapped into colour space.  (Usually Cameras have two colour space schemes sRGB and Adobe RGB)


Step 3: Gamma Correction

Gamma correction is often confused with "brightness".  Its more better explanation is the way the human eye or camera sensor responds to varying light intensities.  When double the amount of light falls on a sensor it generates double the voltage or double the brightness.  This is called a linear response.  The eye on the other hand has a logarithmic response, that means to double the brightness perceived by the human eye one has to increase the amount of light falling on the eye 10 times.   This allows the eye to see more detail in very dark regions and very bright regions.  Gamma correction is applied to make the sensor's captured image appear closer to how the human eye sees it.

Step 4: White Balance

Different types of illuminating sources, eg sunlight, flash, tungsten lights, fluorescent lights, make colours appear different.  The human eye has the capability to see and adapt to changing light sources, a Camera sensor does not. Cameras need to be told what is the source of light.  Modern cameras do have a good built in algorithm that works MOST of the time to detect the type of light and adjust the colour values to compensate.  This is a very complicated process to undo if done incorrectly.  The biggest problem is when using mixed light sources, Fluorescent and Flash, daylight and Flash etc .  This needs to be done correctly to have an accurate representation of colour.

Step 5: Sharpening and Noise Reduction

All cameras will then perform Image Sharpening and Noise reduction on the image.  The amount of sharpening or noise reduction is predetermined by the camera and one may have settings for the mount of sharpening or noise reduction, but it may not the best or optimum for each individual image.

Step 6: JPEG Compression and File saving

Finally after all this, the camera performs JPEG compression.  This makes your file smaller, it doesn't just simply shrink the image file size, it performs what is called "Lossy" compression in that it DISCARDS information in the image.  You lose some information in the image, with JPEG the losses in compression are more loss of colour detail than intensity detail as the human eye is less responsive to colour changes than intensity changes.

Advantages of Shooting in RAW

So if you have reached so far.. you will have a better understanding of all what happens to RAW Image data to convert it into Jpeg. Post processing all RAW files may be tedious and laborious however it does give one certain advantages.

When you perform all these post processing from RAW data to Image data, it gives you better control of all these processes during almost all of the conversion stages.  This Extra Control can often lead to Better Results.  

When these processes are run on a computer, the computer doesn't rush to perform these processes like the camera does.  Better algorithms (that run slower) can be used in the RAW converter software, these algorithms can often produce better results.

 Raw processing algorithms are constantly improving one can always get better software or the latest version of software on one's computer, but the algorithms on your Camera remain the same, they won't improve as they are often made in hardware and not firmware.  You are stuck with what you have bought. With improved algorithms on your PC one can always go back to old RAW images and maybe get a better image or the image you wanted.

RAW images collect image information with say 12-14 bits per colour of each pixel, while JPEG only allows 8 bits per colour of each pixel.  Just to give an idea, going from 8 bits to 9 bits, doubles the amount of information.  Once the camera has given you a JPEG image it has already thrown out heaps of colour information, that gives you much less room for editing and images quickly develop "artifacts" and other unwanted imaging effects.

 Shooting in raw, allows you to set the white balance of your image "after-the-fact". For shooting in tricky lighting situations, and for certain types of color corrections, this feature affords you a level of flexibility you can't get any other way.

  A camera allows to you to configure certain adjustments of picture style, sharpening, noise reduction, contrast, saturation etc. Once these adjustments are applied, it's very difficult to remove them. If your camera is a little too aggressive with its adjustments, you may face some difficult editing. These settings consume much of the post processing freedom. Raw mode, gives you complete control over your image.