PDA, Notebook and Phone Reviews and buyers guide
PDA Phone Notebooks Gaming Gadgets iPhone & iPad Shop Discussion



Image Quality in Camera Phones – What Matters?

By Feisal Mosleh


The image quality (IQ) of a picture is determined by many factors: the level of illumination, quality of the subject and its background, the direction of the light, the quality of the lens, aperture setting, speed of the shutter and many image processing features such as demosaicing and anti-vignetting.

Here we will not discuss how to increase the IQ of an image but what really matters to human perception and review a few examples of how the camera sensor may improve its apparent image quality. What factors make the picture more pleasing to the human eye? What really matters?

Humans are fundamentally analog, not digital, that is, unless you count their fingers when considering ‘digital’. As such, the human eye is an interesting device that helps the brain ‘see’ an image.

For many years, photographic images were judged by the quality of analog cameras and chemical based film on which the images were deposited. Only recently have digital cameras come close to capturing the detail and depth of old analog photography. And these digital cameras have been large, with bulky CCD based sensors in many cases.

But CCD sensors have set the standard by which most consumers judge a digital picture.
CMOS sensors which are preferred in camera phones have to meet these standards if they are to be perceived as producing good image quality pictures.

Charge Coupled Device (CCD) sensors are

  1. Bigger
  2. More resource intensive to design and take longer than CMOS
  1. Power hungry
  2. Expensive

And exhibit a certain high image quality feel, especially in low light conditions.

Complementary Metal Oxide Semiconductor (CMOS) sensors are

  1. Smaller
  2. Much easier and faster to design than CCD
  3. Lower power consumption
  4. Less expensive

Hence CMOS is the choice of most high volume cell phone applications. In 2005, CMOS sensors made the breakthrough in matching the perceived image quality of CCD sensors, becoming responsible for over 97% of the world’s VGA and 1.3 megapixel phone camera volumes. Such a high volume market as cell phones has created an enormous economic pressure on sensor makers to develop better and better performance at reduced prices. In fact VGA sensors which cost $5 in 2003 are now selling for under $1.60 and yet the image quality and low light performance of these VGA sensors is in the same ball park, with slightly worse low light performance, due to smaller pixels.

Why Bother About Better Image Quality?

According to InfoTrends, outside of the voice function, the camera is the most used feature but users are not necessarily utilizing their pictures. If they utilize those pictures i.e. email them or print them, a tremendous amount of latent revenue is unleashed for many players in the digital photo ecosystem, from printer companies and ink makers to imaging software vendors and wireless carriers offering email and imaging services for camera phone pictures.


And why are people not printing their camera phone pictures?

Low IQ is seen as being one of the inhibitors for printing pictures.

Low IQ is also preventing users from sharing or emailing their pictures.

So What Really Matters?

Through a number of tests and evaluations and research it is apparent that the following five factors greatly influence the perception of good image quality.

Image Quality - The Five Elements that Matter

Enhanced perceived image quality can be obtained through appropriate:

  1. VIVIDNESS – color uniformity and richness
  2. BRIGHTNESS – low light performance
  3. CLARITY – no noise or distortion
  4. SHARPNESS – great detail
  5. CONTRAST – dynamic range


I use the word ‘appropriate’ because there is a fair amount of subjective leeway from person to person and even from region to region as to what constitutes the most pleasing image.

What does this mean for achieving a high perceived image quality?
Sharpness of subjects in a picture is probably one of the strongest determinants in the perception of high image quality. This is followed closely by the quality of the colors and the vividness or richness of the subjects. Well adjusted brightness and contrast and the lack of any noise in the picture also make it look better. But there are some inherent issues that phone camera makers must tackle to achieve better perceived image quality.

Challenges faced by camera sensors in camera phones:
-Produce great images in low light conditions with very little apparent noise.

-Ensure there is color uniformity so that the different colors in images appear distinct. There should be no blurring of colors.

-Reduce inconsistency or variability in the quality and look of colors between different sensors from the same production line. i.e. increase the manufacturing yield rate by minimizing the need to ‘throw away’ sensors that are outside of the acceptable range of variability.

-The ‘thin is in’ trend is causing image quality to deteriorate by demanding smaller camera modules for thinner, sleeker phones. Some clamshell and candy bar phones are now only 7-9mm in thickness! What does this do to the image quality?

Well it means the height of the camera module must be closer to 4mm or 5mm. To get that ‘z’ height, you must have a smaller optical format, i.e. the diameter of the lens in the camera module must be smaller. This is why new resolutions introduced to the camera phone market (say the new 3MP) typically start at 1/3” optical format and then progress to ¼” and finally to 1/5” optical format.

The diagram above shows how this reduction in size helps create thinner phones but means you have lower IQ. Why? Because of several factors:

For the same camera resolution, smaller camera modules, with smaller optical formats have smaller sensor arrays created by using smaller pixels. E.g. a 1/3” 2MP typically uses a 2.8 micron square pixel, but a ¼” 2MP uses a 2.2 micron square pixel. Smaller pixels absorb less light because of a smaller cross sectional area and in darker conditions their signal to noise ratio goes down, resulting in higher noise levels and lower signal.

Also, a smaller module and optical format means a smaller and thinner lens, which usually means fewer elements, resulting in a lower quality lens.

Examples of Enhancements to improve Image Quality
Improving image quality by improving light capture:

Some examples of recent developments that have made CMOS sensors of comparable performance to CCD: All CMOS sensors have a microlens situated above each single pixel to focus the incoming light on to the photodiode at the bottom of the long ‘straw’ that is the pixel. To make CMOS sensors even more sensitive to light in dark conditions, Avago technologies introduced a second microlens below the first, to bend the already focused light even further, ensuring that no light is lost through the edges to crosstalk, entering adjacent pixels.

An example of how the apparent IQ can be improved by adding the Adaptive Tone Mapping (ATM) function in the image processing pipeline: In the example below, Avago uses ATM in its I-Pipe to create richer (more vivid) looking pictures with greater contrast and enhanced brightness.

The Challenges for Better Image Quality in Camera Phones

In Camera phones, image quality can be improved through better

    1. Low light performance
    2. Good colors and white balance
    3. Sharpness (no ‘blurriness’ of colors)
    4. Noise in dark conditions
    5. Ensuring manufacturing consistency – this last requirement must not be overlooked in super high volume products like camera phones, where even a 1% return rate can significantly impact manufacturers’ profitable margins.

The perceived image quality of phone cameras is tied closely to the areas of low light performance, low noise, vividness, contrast, brightness, sharpness and clarity.

As image quality improves, the degree of utilization of the camera phone will also increase.

But we must remember that IQ is subjective and depends on the eyes of the beholder. If enough adjustments are made to the image, it’s perceived quality can be enhanced.

By introducing innovative methods to improve light capture and shape that information in the image processing pipeline for improved vividness, brightness, clarity, sharpness and contrast, images can be made to look better and become closer to what the human eye perceives as good image quality.



Feisal Mosleh is head of worldwide marketing, business development and systems engineering for Avago technologies’ mobile imaging business in San Jose, CA. He can be reached at

Back to Home Questions? Comments? Post them in our Discussion Forum!