The problems associated with lesser-quality image processing engines begin with data processing times which can slow down camera performance substantially. This is not only annoying insomuch that it can potentially mean that a photographer misses a once-in-a-lifetime shot but will also impede the sequential shooting speeds. Moreover, the delivery of good photos also relies on the image processing engine to accurately compute the large amount of image data so that colours are faithfully reproduced, edges appear smooth and natural, and noise is reduced to a minimum.

Today’s high resolution cameras mean that more data must be processed by the camera – making the incorporation of a powerful image processing engine more critical than ever.

Background information The photodiodes employed in an image sensor are colour-blind by nature: They can only record shades of grey. To get colour into the picture, they are covered with different colour filters – red, green and blue (RGB). With most sensors there are two green for each blue and red diodes. As each photodiode records the colour information for exactly one pixel of the image, without an image processor there would be a green pixel next to each red and blue pixel.

The solution:
The image processing engine comprises a combination of hardware (processors) and software (algorithms). Among its tasks, it gathers information about the luminance and chrominance from the individual pixels and uses it to compute the correct colour and brightness values for each pixel. If it does this well, the result is an image with natural and pleasing colours, balanced contrast and sharpness.
This process, however, is very complex and involves multiple operations. Its success depends largely on the “intelligence” of the algorithms applied.

Getting the colours right
As stated above, the image processor evaluates the colour and brightness data of a given pixel. It then compares them with the data of its neighbouring pixels and processes these using a complex algorithm to get the correct colour and brightness value for this pixel. But the image processor also assesses the whole picture to determine the correct distribution of contrast. By adjusting the gamma value (heightening or lowering the contrast range of an image’s mid-tones) subtle tonal gradations, such as in human skin or the blue of the sky, come out much more realistically.

Getting rid of noise
Noise is a phenomenon found in any electronic circuitry. In digital photography its effect is often visible as random spots of obviously wrong colour in an otherwise smoothly-coloured area. Noise increases with temperature and exposure times. When higher ISO settings are chosen the electronic signal in the image sensor is amplified, which at the same time increases the noise level, leading to a lower signal-to-noise ratio. A good image processor separates the noise from the image information and removes it. This can be quite a challenge, as the image may contain areas with fine textures that when treated as noise could potentially lose some of its definition.

Getting smooth and sharp edges
As the colour and brightness values for each pixel are calculated, some image softening is applied to even out fuzziness that may occur in the process. In order not to lose the impression of depth, clarity and fine details, sharpening of contours and edges is needed. The image processor therefore needs to be able to detect edges correctly and to reproduce them smoothly and without over-sharpening.

Getting the job done quickly
Photographers don’t want to wait for the camera’s image processor to complete its job before they can carry on shooting – they don’t even want to notice some processing is going on inside the camera. With increasing image sensor resolution image processing algorithms need to be optimised to cope with the higher data volume in the same or even a shorter period of time.

How it works (The Olympus TruePic III image processing engine):

With the introduction of its latest image processing engine, Olympus has moved another step forward in regard to perfect image quality. Coupled with fine-tuning the Live MOS sensor, the new engine’s ability to reproduce colours even more naturally has been improved. This is in part due to the Advanced Proper Gamma III technology featuring independent control of luminance and chrominance difference signals for faithful reproduction of even pale colours. Now individual colours can be corrected without affecting the reproduction of other colours. Also, colour reproduction is fine-tuned so that they are not just correct but also appear pleasing to the human eye. Olympus engineers therefore paid special attention to faithfully reproduce human skin colours and the blue of the sky.

The new Advanced Noise Filter III contributes to the faithful reproduction of images through the reduction of noise by isolating the image and noise signals accurately. It replaces the real space (real image) with a frequency space and extracts the signal component. It then smoothes out all signal components while preserving the edges.

To reproduce edges smoothly but still sharply, the Advanced Detail Reproduction technology accurately detects edge direction and applies a Low Pass Filter (LPF) in the edge direction and a High Pass Filter (HPF) in the edges normal line direction. This way edges become smooth and false colours can be completely eliminated.

Find here some older information regarding this topic