From aeafc969a5242d59d6e08f8054ac1824102f8bec Mon Sep 17 00:00:00 2001
From: Shreya123714 <yadavshreya400@gmail.com>
Date: Thu, 31 Oct 2024 01:08:05 +0530
Subject: [PATCH] Added a Exp. moving average filter to smooth audio signal
---
audio_filters/ema_filter.py | 67 +++++++++++++++++++++++++++++++++++++
1 file changed, 67 insertions(+)
create mode 100644 audio_filters/ema_filter.py
diff --git a/audio_filters/ema_filter.py b/audio_filters/ema_filter.py
new file mode 100644
index 000000000..2786cd110
--- /dev/null
+++ b/audio_filters/ema_filter.py
@@ -0,0 +1,67 @@
+"""
+By @Shreya123714
+
+https://en.wikipedia.org/wiki/Moving_average
+"""
+
+import numpy as np
+from typing import List
+
+class EMAFilter:
+ """
+ A class for applying an Exponential Moving Average (EMA) filter
+ to audio data.
+
+ Attributes:
+ alpha (float): Smoothing factor where 0 < alpha <= 1.
+ ema_value (float): Stores the most recent EMA value
+ for the ongoing calculation.
+ """
+
+ def __init__(self, alpha: float) -> None:
+ """
+ Initialize the Exponential Moving Average (EMA) filter.
+
+ Parameters:
+ alpha (float): Smoothing factor where 0 < alpha <= 1.
+
+ Raises:
+ ValueError: If alpha is not within the range (0, 1].
+ """
+ if not (0 < alpha <= 1):
+ raise ValueError("Alpha must be between 0 and 1.")
+ self.alpha = alpha
+ self.ema_value = None
+
+ def apply(self, audio_signal: List[float]) -> np.ndarray:
+ """
+ Apply the EMA filter to a sequence of
+ audio signal values.
+
+ Parameters:
+ audio_signal (List[float]): List of numerical values
+ representing the audio signal.
+
+ Returns:
+ np.ndarray: Array containing the smoothed audio signal.
+
+ Example:
+ >>> ema_filter = EMAFilter(0.2)
+ >>> np.allclose(ema_filter.apply([0.1, 0.5, 0.8, 0.6, 0.3, 0.9, 0.4]),
+ ... [0.1, 0.18, 0.304, 0.3632, 0.35056, 0.460448, 0.4483584])
+ True
+ """
+ ema_signal = []
+ for sample in audio_signal:
+ if self.ema_value is None:
+ # Initialize the EMA with the first sample
+ self.ema_value = sample
+ else:
+ # Calculate the EMA for the current sample
+ self.ema_value = self.alpha * sample + (1 - self.alpha) * self.ema_value
+ ema_signal.append(self.ema_value)
+ return np.array(ema_signal)
+
+if __name__ == "__main__":
+ import doctest
+ doctest.testmod()