Lossy compression
ABSTRACT
The requirement that no information be lost in the compression process puts a limit on the amount of compression we can obtain. The lowest number of bits per sample is the entropy of the source. This is a quantity over which we generally have no control. In many applications this requirement of no loss is excessive. For example, there is high frequency information in an image which cannot be perceived by the human visual system. It makes no sense to preserve this information for images that are destined for human consumption.
In short, there are numerous applications in which the preservation of all information present in the source output is not necessary. For these applications we relax the requirement that the reconstructed signal be identical to the original. This allows us to create compression schemes that can provide a much higher level of compression.
In this section we describe a number of compression techniques that allow loss of information, hence the name lossy compression. We begin with a look at quantization which, in one way or another, is at the heart of all lossy compression schemes.
INTRODUCTION
Lossy data compression is the converse of lossless data compression. In these schemes, some loss of information is acceptable. Dropping nonessential detail from the data source can save storage space. Lossy data compression schemes are informed by research on how people perceive the data in question. For example, the human eye is more sensitive to subtle variations in luminance than it is to variations in color. JPEG image compression works in part by rounding off nonessential bits of information. In many applications this requirement of no loss is excessive. For example, there is high-frequency information in an image which cannot be perceived by the human visual system. It makes no sense to preserve this information for images that are destined for human consumption. Similarly, when we listen to sampled speech we cannot perceive the
The requirement that no information be lost in the compression process puts a limit on the amount of compression we can obtain. The lowest number of bits per sample is the entropy of the source. This is a quantity over which we generally have no control. In many applications this requirement of no loss is excessive. For example, there is high frequency information in an image which cannot be perceived by the human visual system. It makes no sense to preserve this information for images that are destined for human consumption.
In short, there are numerous applications in which the preservation of all information present in the source output is not necessary. For these applications we relax the requirement that the reconstructed signal be identical to the original. This allows us to create compression schemes that can provide a much higher level of compression.
In this section we describe a number of compression techniques that allow loss of information, hence the name lossy compression. We begin with a look at quantization which, in one way or another, is at the heart of all lossy compression schemes.
INTRODUCTION
Lossy data compression is the converse of lossless data compression. In these schemes, some loss of information is acceptable. Dropping nonessential detail from the data source can save storage space. Lossy data compression schemes are informed by research on how people perceive the data in question. For example, the human eye is more sensitive to subtle variations in luminance than it is to variations in color. JPEG image compression works in part by rounding off nonessential bits of information. In many applications this requirement of no loss is excessive. For example, there is high-frequency information in an image which cannot be perceived by the human visual system. It makes no sense to preserve this information for images that are destined for human consumption. Similarly, when we listen to sampled speech we cannot perceive the