Hello,
From the information shared, I could infer that you have queries related to the modulation schemes of "Binary Frequency Shift Keying (BFSK)" and "8-Phase Shift Keying (8-PSK)". "BFSK" modulates the data by varying the frequency of the carrier wave. Specifically, it uses two distinct frequencies to represent binary '0' and '1'. "8-PSK" modulates the data by changing the phase of the carrier wave. It divides the circle into 8 equal parts (each 45 degrees apart), and each phase represents a unique 3-bit binary number.
When the frequency separation in "BFSK" is relatively large (e.g., 8 times the bit rate), the scatter plot might superficially resemble that of "8-PSK" because:
- Distinct Points: For "8-PSK", the multiple distinct points are the 8 different phases equally spaced around a circle. For "BFSK" with large frequency separation, the rapid changes in frequency can create a pattern that, when sampled, might appear to have multiple clusters, especially if visualized in a certain way (e.g., using a specific type of visualization or demodulation technique that doesn't perfectly align with "BFSK's" nature).
- Circular Pattern: With a high frequency separation, the "BFSK" signal's representation might cycle through different phases as the frequency changes, potentially creating a pattern that looks circular or has multiple clusters, somewhat resembling the 8 points of "8-PSK".
In "PSK", noise primarily affects the phase of the signal, which can cause the points in the scatter plot to spread out, especially at lower "Signal-to-Noise Ratios (SNR)". This spreading occurs because the phase is very sensitive to noise, and even a small amount can significantly alter the perceived phase of a signal. "BFSK", on the other hand, is generally more robust to noise at lower SNRs. This is because:
- Frequency Separation: The key distinguishing factor between the symbols is their frequency, not their phase or amplitude. Frequency is inherently more resistant to noise, especially when the frequencies are sufficiently separated. Thus, the points in a "BFSK" scatter plot do not spread as much with lower "SNR", because the noise is less likely to cause a frequency to be misinterpreted as the other.
- Energy Distribution: "BFSK" signals tend to have their energy more spread out in the frequency domain, making them less susceptible to being disrupted by noise, which often affects a narrower band of frequencies at any given time.
In summary, the apparent similarity of "BFSK" (with large frequency separation) to "8-PSK" in scatter plots is mostly superficial and depends on the visualization method. The fundamental difference in how "BFSK" and "PSK" modulations encode data results in their differing responses to noise, especially at lower SNRs.
For a better understanding, here is a helpful article that compares various modulation techniques: https://www.gaussianwaves.com/2010/04/performance-comparison-of-digital-modulation-techniques-2/
I hope this helps!
