Calculating time between high state

Question

In python I want calculations of high / low states of a signal in chunks of a given size (samples_to_process).

The two required calculations are number index between rising edges (length_between_high_states) and number of index's the signal is high for (high_state_length).

The calculations must be stateful across chunnks.

Lets take a small reproduceable example:

data = np.array([0,1,1,0,1,1,1,0,0,0,0,0,1,1,1,1,1,1,0,0,0,0,0,0,0,1,1,1,0,0,0,0,1,1,1,1,0,0])

If this array is read 8 items at a time, the first iteration is

high_state_length = 2 
length_between_high_states = 3

then

high_state_length = 3
length_between_high_states = 9

I believe I have the correct logic to read in the first state of the array and signal changes, but subsequent state changes in the signal and carrying the state across chunks are not yet implemented:

import numpy as np
#total array size = 25
data = np.array([0,1,1,0,1,1,0,0,0,0,0,0,1,1,1,1,1,1,0,0,0,0,0,0,0,1,1,1,0,0,0,0,1,1,1,1,0,0])
#size of samples of data array to process in each read
samples_to_process = 8
samples = np.zeros(samples_to_process, dtype=np.complex64)
threshold = 0.5

index = 0
#slice(index, index+samples_to_process)
for index in range(0, data.size, samples_to_process):
    samples=data[index:index+samples_to_process]
    print(samples)
    while True
        start_index = np.argmax(samples > threshold)
        stop_index = np.argmax(samples[start_index:] < threshold) + start_index
        next_start_index = np.argmax(samples[stop_index:] > threshold) + stop_index
        length_between_high_states = next_start_index - start_index
        high_state_length = stop_index - start_index
        # how to calculate remainder state and pass into next itr
        start_index = next_start_index
        print("next loop")

The question is how to pass the signal state between iterations to be included in subsequent calculations.

Answer 1

You just need to track the last rising edge, and the state of the last sample. Then, you get an event on each rising edge and each falling edge.

(I've now reworked this to use a class to maintain the state data.)

import numpy as np
#total array size = 25
data = np.array([0,1,1,0,1,1,0,0,0,0,0,0,1,1,1,1,1,1,0,0,0,0,0,0,0,1,1,1,0,0,0,0,1,1,1,1,0,0])
#size of samples of data array to process in each read
samples_to_process = 8
threshold = 0.5

def islow(v):
    return v < threshold
def ishigh(v):
    return v >= threshold

class Checker:
    def __init__(self):
        self.last = 0
        self.lastrise = 0
        self.index = 0

    def feed(self, samples):
        for j,n in enumerate(samples):
            i = self.index + j
            # Is this a rising edge?
            if ishigh(n) and islow(self.last):
                if self.lastrise:
                    print('since last rising edge:', i-self.lastrise)
                self.lastrise = i
            # Is this a falling edge?
            elif islow(n) and ishigh(self.last):
                print('pulse width', i-self.lastrise)
            self.last = n
        self.index += len(samples)

checker = Checker()
for index in range(0, data.size, samples_to_process):
    checker.feed( data[index:index+samples_to_process] )

Output:

pulse width 2
since last rising edge: 3
pulse width 2
since last rising edge: 8
pulse width 6
since last rising edge: 13
pulse width 3
since last rising edge: 7
pulse width 4

Followup

Really, like most algorithms, this is just doing what you would do on paper. last holds the previous signal level. lastrise holds the index of the last rising edge (0-to-1 transition)

You asked about this statement:

for j,n in enumerate(samples):

The enumerate function is a better way to handle running through a list of things when you need the index of the item as well. It is functionally identical to:

for j in range(len(samples)):
    n = samples[j]

except that it is more efficient. I need to have the index of the sample so I can compute the current position in the chunk. So, within that loop, j has the index of the current sample, and n has the level of the sample.