Digital Signal Processing Reference
In-Depth Information
−
−
Figure 7.6
ECG waveform reconstructed from a digital signal obtained as a result of sampling
performed on the basis of sine-wave crossings
On the other hand, while the listed advantages look attractive, the fact remains
that the data acquisition process is indirectly randomized, which may in some
cases discourage applications of it. To show that this potential disadvantage is
not always damaging and that using unfamiliar specific algorithms for handling
nonuniformly represented data is not always necessary, consider an example
related to encoding and reconstruction of a typical signal for biomedicine, an
ECG signal, illustrated in Figure 7.6.
A picked-up and conditioned cardio-signal was sampled according to the
scheme of sine-wave crossings given in Figure 7.3(b). The frequency of the ref-
erence signal was chosen to be 500 kHz and the mean sampling rate was set close
to 1 kS (kilosample)/s. This means that only one half-period in 1000 was used
as a time slot within which the crossings of the input and the reference signals
were taken into account. The values of the reference signal corresponding to the
crossing instants occurring during these time slots were taken as the sampled
input signal values. Under the specified sampling conditions, the randomness
of the sampling intervals plays an insignificant role. Therefore it was assumed
that the sine-wave crossings occur in the middle of sampling time slots that are
1
μ
s long and that the sampling process is consequently periodic. The waveform
