Why Multiplication in Time Becomes Convolution in Frequency
In signal processing, there is a powerful dual relationship,
Multiplication in time domain becomes convolution in frequency domain
This is the “other side” of the convolution theorem.
The Key Relationship
Theorem
Meaning
- Time domain → multiplication
- Frequency domain → convolution
Intuition
“Multiplying signals mixes their frequencies”
Why Does This Happen?
Fourier Transform represents signals as sum of sine waves (frequencies)
When you multiply two signals, you are mixing their sine waves
Result
Frequencies spread and combine
Key Insight
Multiplication in time causes convolution in frequency
Simple Example
Multiplication of two Sine waves, let fm = 10Hz, fc = 100Hz
due to trigonometric identities
Result in Frequency domain
Spectrum shift and spreading (convolution in the frequency domain):
- fc - fm = 90Hz
- fc + fm = 110Hz, which corresponds to Double SideBand Suppressed Carrier (DSB-SC) modulation
Sin(10Hz) and Sin(100Hz) signals
Multiplication in the time domain: Sin(10Hz) x Sin(100Hz) → modulated signal
Convolution in the frequency domain: FFT of Sin(10Hz) * FFT of Sin(100Hz) → 90Hz(100Hz - 10Hz) and 110Hz(100Hz + 10Hz)
Interpretation
The product of two sinusoids results in sum and difference frequencies (i.e., amplitude modulation-like behavior)
Practical Meaning
Communication (Modulation)
- Signal x Carrier
- Moves signal to higher frequency
Audio Effects
- Signal × Envelope
- an operation that controls the amplitude (dynamics) of an audio signal over time
Signal Gating
- Multiply by window
- Alters frequency components
Windowing Effect
Multiplying by a window cuts signal in time
Result
Finite data length in time domain → Sinc-like, side lobe in frequency domain
How does signal truncation affect FFT results, particularly in terms of spectral leakage?
Key Insight
Multiplying a signal by a window localizes it in time, which results in a convolution in the frequency domain, causing spectral spreading (smearing, spectral leakage). This reflects the fundamental time–frequency trade-off: stronger time localization leads to broader frequency content.
Comparison with Convolution Theorem
| Convolution in time domain | Multiplication in frequency domain |
| Multiplication in time domain | Convolution in frequency domain |
Key Idea
Dual relationship
Key Takeaways
- Time multiplication → frequency convolution
- Causes spreading of frequency
- Basis of modulation and windowing
- Essential for understanding spectral behavior
Conclusions
Multiplication in the time domain leads to convolution in the frequency domain, revealing a fundamental dual relationship in signal processing.
- This operation causes frequency spreading, meaning spectral components are mixed and redistributed.
- It explains key phenomena such as modulation, windowing, and spectral shaping, where signals are shifted or altered in frequency.
- While useful, it also introduces effects like spectral broadening, especially when signals are time-limited.
In summary,
time-domain multiplication transforms and spreads frequency content, making it a critical concept for understanding modulation, filtering effects, and real-world signal behavior.
Suggested Further Reading
You may also be interested in these topics:
Why Multiplication in Time Becomes Convolution in Frequency
In signal processing, there is a powerful dual relationship,
Multiplication in time domain becomes convolution in frequency domain
This is the “other side” of the convolution theorem.
The Key Relationship
Theorem
Meaning
Intuition
“Multiplying signals mixes their frequencies”
Why Does This Happen?
Fourier Transform represents signals as sum of sine waves (frequencies)
When you multiply two signals, you are mixing their sine waves
Result
Frequencies spread and combine
Key Insight
Multiplication in time causes convolution in frequency
Simple Example
Multiplication of two Sine waves, let fm = 10Hz, fc = 100Hz
Result in Frequency domain
Spectrum shift and spreading (convolution in the frequency domain):
Sin(10Hz) and Sin(100Hz) signals
Multiplication in the time domain: Sin(10Hz) x Sin(100Hz) → modulated signal
Convolution in the frequency domain: FFT of Sin(10Hz) * FFT of Sin(100Hz) → 90Hz(100Hz - 10Hz) and 110Hz(100Hz + 10Hz)
Interpretation
The product of two sinusoids results in sum and difference frequencies (i.e., amplitude modulation-like behavior)
Practical Meaning
Communication (Modulation)
Audio Effects
Signal Gating
Windowing Effect
Multiplying by a window cuts signal in time
Result
Finite data length in time domain → Sinc-like, side lobe in frequency domain
How does signal truncation affect FFT results, particularly in terms of spectral leakage?
Key Insight
Multiplying a signal by a window localizes it in time, which results in a convolution in the frequency domain, causing spectral spreading (smearing, spectral leakage). This reflects the fundamental time–frequency trade-off: stronger time localization leads to broader frequency content.
Comparison with Convolution Theorem
Key Idea
Dual relationship
Key Takeaways
Conclusions
Multiplication in the time domain leads to convolution in the frequency domain, revealing a fundamental dual relationship in signal processing.
In summary,
time-domain multiplication transforms and spreads frequency content, making it a critical concept for understanding modulation, filtering effects, and real-world signal behavior.
Suggested Further Reading
You may also be interested in these topics: