Digital Signal Processing Reference
In-Depth Information
and conditions for their processing vary over a wide range, adapting signal dig-
itizing to them represents a challenge. To face this, the used digitizing methods
and means need to be sufficiently flexible. The hybrid sampling approach might
be considered as a competitive option for developing specific sufficiently adapt-
able sampling methods and sampling driver designs. Hybrid double sampling
techniques discussed in this section develop the basic hybrid sampling idea fur-
ther. Specifically, organization of the sampling operation in a way ensuring that
the signal sample value can be taken at sufficiently short sampling intervals is
considered.
Hybrid double sampling is implementation of hybrid sampling based on the
use of at least two ADCs connected in parallel. It is aimed at obtaining the
right conditions for processing signals with both discrete and continuous spectra.
To achieve this, the distances between adjacent sample values often have to be
shortened and application of hybrid double sampling makes it possible to perform
sampling with small discrete sampling interval increments. This kind of sampling
is based on various combinations of periodic and nonuniform sampling elements.
Note that some of the hybrid sampling schemes can be realized either as a single
ADC version or as a hybrid double sampling scheme built on the basis of two
ADCs. For instance, this applies to the hybrid sampling model mixing pseudo-
randomly decimated phase-shifted periodic sampling point streams, considered
above.
A few specific models of hybrid double sampling models are considered in this
section. These models are considered to be generic and many variations of them
are possible. Their characteristics differ. Some sampling models are better suited
for the needs of special application conditions than others. Therefore it is possible
to adapt, to some extent, the hybrid double sampling mode and parameters to
specific applications. It also means that it is better if the microelectronic sampling
drivers used to implement these sampling techniques are reprogrammable.
10.2.1 Providing for Short Sampling Intervals
There is a typical drawback of scarce nonuniform sampling that leads to some
limitations imposed on essential applications of this kind of sampling. As the
distribution of the sampling intervals for nonuniform sampling has to guarantee
that the distance between two consequent sampling instants is equal or larger than
the shortest sampling time interval permitted for the specific ADC to be used, it is
impossible to obtain signal sample values located more closely on the time axis.
This limitation of scarce nonuniform sampling essentially impacts a particular
type of wideband signal processing. For instance, this applies to the spectrum
