Digital Signal Processing Reference
In-Depth Information
DCT values. However, in many image blocks many of the DCT
values are near zero, or small, and can be presented with few bits.
This can allow the pixel block to be presented more efficiently
with fewer values. However, this representation is approximate,
because when we minimize the bits representing various DCT
coefficients, we are quantizing. This is a loss of information,
meaning the pixel block cannot be restored perfectly.
Search WWH ::




Custom Search