Enhance EMAFilter for Audio Signals: Optimized, Memory-Efficient, and Stream-Friendly

[Manas-Dikshit]

This PR enhances the EMAFilter, a tool used to smooth audio signals using an Exponential Moving Average (EMA). The main goal of these changes is to make the filter faster, more memory-efficient, and enterprise-ready, all while keeping its original functionality intact.

1. Smarter Memory Usage
   We added a new method, applyInPlace(double[] samples), which allows the filter to process audio samples directly in the original array. This means we don’t need to create a new array for every batch of audio data, saving memory—especially helpful when dealing with large audio files.

2. Faster and More Efficient Processing
   The filter loop has been optimized to reduce unnecessary calculations. Additionally, we introduced the next(double sample) method, which lets you process streaming data in real-time. This makes the filter suitable not just for offline processing, but also for live audio applications.

3. Enterprise-Ready Improvements
   The class is now final and the smoothing factor (alpha) is immutable, ensuring safe usage across multiple threads or systems.
   Robust input validation has been added to prevent errors when passing null or invalid values.
   Clear and detailed Javadocs explain exactly how each method works, making the code easier to maintain.
   A toString() method has been included for easy debugging and monitoring of the current EMA state.

4. Better Usability
   apply(double[] samples) returns a new array with smoothed values, leaving the original untouched, giving developers flexibility depending on their needs.
   next(double sample) is perfect for real-time, sample-by-sample processing.
   getLastEma() lets you quickly access the most recent EMA value without recalculating everything.

5. Clear Initialization Handling
   The filter uses Double.NaN to represent the initial state of lastEma, making it obvious when the filter hasn’t processed any data yet.

Why this matters:
Audio processing often involves large datasets or real-time streams, so efficiency, memory savings, and usability are critical. With these improvements, the EMAFilter is now better suited for modern Java applications, whether you’re handling offline batches or live audio streams.