Dynamic Range

Highlight clipping and noise are the two most important technical problems in digital photography.  They are opposite ends of the tonal scale and therefore they define the dynamic range of the camera.

Dynamic range is the range between light and dark your camera's sensor can capture in an exposure.  We as photographers can measure it in terms of stops.  The dynamic range of a scene is the number of stops from the brightest highlight to the darkest shadow in the image.

The highest image quality in digital photography arises from utilising the full dynamic range of the sensor to capture the full range of the scene.

Getting technical
As we already know, our eyes deal with light differently to what our camera's sensor does.  In general our eyes can handle 20-24 stops of light whereas the average DSLR has a dynamic range of about 10-11 stops.  This means that a single frame can capture brightness levels 10 or 11 stops apart without clipping the highlights.

From a more technical point of view the dynamic range for a sensor relies on limits.  Firstly the capacity of each photosite to hold electrons - full well capacity.  The larger the area of the photosite, the better this will be.  Modern DSLRs (high end) have capacities of 7,000 to 10,000 electrons.

Secondly, the limit set by noise and the point at which noise cannot be distinguished from real detail.  This is called the noise floor which has 4-8 electrons. If you divide the full well capacity by the noise floor you will get the dynamic range of the sensor.  

Losing information
If the dynamic range of your scene fits that of your camera, you will have an image that captures all the visible detail.  However, if the camera's range is less than the scene all the visual elements will not be captured and something will be lost in your image. This something usually results in the clipping of the highlights and noise in the shadow areas.

Working with your camera's range
The dynamic range is between the brightest and the darkest that can be captured we need to find the end points to work out what our camera can handle.  Highlights are the easiest to find.  The highlight clipping warning tells us when we are about to lose information from the brightest parts of our image.  We can use this to find the exposure that just captures the brightest highlight.

Noise on the other hand determines the darkest tone for the dynamic range.  This is because noise results in the shadow areas where not enough light has hit the sensor.  This causes a sampling error  resulting in speckles in your image.

Looking at the dark areas of an image there is a point where it becomes impossible to distinguish between noise and real detail, any darker and the noise drowns out all the detail.  To see this we can lighten the image  in Photoshop.  However, there is also an element of taste here and what one person would determine as noise and drowned out detail another would find acceptable in the image.