What is Interpolation at Digital Camera?
A new advertisement about digital camera said that this new digital camera
have 3 megapixel resolution until 6 megapixel (interpolated). This information
make a customer become confuse about this product, is it 3 megapixel or 6 megapixel?
Ok, this article will explain about what is digital camera interpolation.
Interpolation is a technique where the spatial resolution of an image is increased
from its original size to a higher (larger) resolution. The spatial resolution
of an image is simply its horizontal x vertical pixel count. (i.e. 1600 x 1200).
In extreme explanatory and simple terms it can be summed up as that interpolation
refers to software programs that can effectively enlarge image resolution beyond
the actual resolution by adding extra pixels using complex mathematic calculations.
Now, this is a feature that can be absolutely useful in providing a good success
to a user.
When your image is 400 x 300 pixels, and you want to enlarge it to 400%, it
becomes 1600 x 1200 pixels. This means that from a mere 120,000 pixels in the
original, you go to almost 2 million pixels in the enlargement. So your computer
software has to "guess" what all those new pixels are going to be,
based only on those 120,000 original values. This just a sample of Interpolation.
The interpolation have some methods like bicubic and lanczos look to more known
pixels in order to estimate the empty places. Interpolation methods alway guess
an unknown pixel by taking some sort of average of the surrounding known pixels.
How this average is exactly calculated is what makes one interpolation technique
differ from another. These differences can be quite dramatic.
The big difference between all "classic" interpolation methods (such
as bilinear, bicubic, lanczos and b-spline), and "adaptive" methods
like GF, S-Spline (or the techniques that forum contributors Ruzinsky and Darian
use), is that the classic methods always take the average (of surrounding known
pixels) the same way throughout the whole picture. That is, the way it takes
the average depends only on the pixel's position in the image, and not on the
actual pixel values.
Interpolation by Software and Hardware
Software Interpolation.
Software interpolation can be performed on a digital image using an one of
a number of image editing programs (PhotoShop, MGI PhotoSuite, etc.) This is
often known as "resizing". This is done on a computer, performed on
an image file (from a digital camera or scanner) that already exists in a file
format such as JPG.
Hardware Interpolation.
Hardware interpolation involves the resizing of an image - but it differs from
software interpolation in that the image is resized algorithmically inside the
camera during the image processing sequence - before the image has been saved
as a JPG image file.
JPG is a compressed image file format. Whenever a file is saved as a JPG file,
image data is lost in the compression process. Image data is sacrificed (thrown
away) in exchange for file-size efficiency.
Software interpolation is a process performed after the JPG losses have been
applied. Hardware interpolation occurs inside the camera prior to JPG compression
- before JPG losses have been applied. The resulting hardware-interpolated image
will be of superior quality to a comparable image interpolated in software.
HOME ARCHIVES
