deepseek-r1 on Jetson Orin Nano Super FFT Demo 1

deepseek-r1 / 7.6B/ Jetson Orin Nano Super

Prompt

fast fourier transformation (FFT)에 대해 이해하고 싶습니다. 진폭과 주파수가 다른 sine과 cosine 웨이브들을 각각 5개씩 더한 입력 데이터를 만들고, 이 데이터를 FFT하여 결과에서 입력한 진폭과 주파수를 확인하는 파이썬 예제를 부탁합니다. 그래프는 plotly로 해주세요.

Response

Jetson Orin Nano Super에 보관되어 있던 대화가 사라졌다.

코드

import numpy as np
import plotly.graph_objects as go

# Parameters
sample_rate = 1000  # Sampling rate in Hz
duration = 2  # Duration in seconds
n_samples = sample_rate * duration

# Time array from 0 to 2*pi
t = np.linspace(0, 2*np.pi, n_samples)

# Generate signals with different amplitudes and frequencies for sine and cosine
np.random.seed(42)  # For reproducibility

def generate_signal(frequency):
    amplitude = np.random.uniform(0.1, 1)
    return amplitude * np.sin(t * frequency), amplitude

def generate_cosine_signal(frequency):
    amplitude = np.random.uniform(0.1, 1)
    return amplitude * np.cos(t * frequency), amplitude

# Generate 5 sine and 5 cosine signals with different frequencies
input_signals = []
sine_amps = []
cosine_amps = []

for i in range(1, 6):
    sine_wave, amp = generate_signal(i)
    cosine_wave, cos_amp = generate_cosine_signal(i)

    input_signals.append(sine_wave + cosine_wave)
    sine_amps.append((i, amp))
    cosine_amps.append((i, cos_amp))

# Combine all signals
input_signal = np.sum(input_signals, axis=0)

# Compute FFT
fft = np.fft.fft(input_signal)
magnitude = np.abs(fft)
frequency = np.fft.fftfreq(n_samples) * sample_rate

# Create time domain figure
time_trace = go.Scatter(x=t, y=input_signal,
                         name='Time Domain Signal',
                         mode='lines')

# Create frequency domain figure
freq_trace = go.Scatter(x=frequency[:n_samples//2], 
                        y=magnitude[:n_samples//2],
                        name='Frequency Spectrum',
                        mode='lines')

sine_traces = []
cosine_traces = []

for freq, amp in sine_amps:
    x = np.fft.fftfreq(n_samples) * sample_rate
    y = amp * np.sqrt(2)
    trace = go.Scatter(x=x[x > 0], y=y[:n_samples//2], 
                       name=f'Sine Component (f={freq}, A={amp})',
                       line=dict(color='red'))
    sine_traces.append(trace)

for freq, cos_amp in cosine_amps:
    x = np.fft.fftfreq(n_samples) * sample_rate
    y = cos_amp * np.sqrt(2)
    trace = go.Scatter(x=x[x > 0], y=y[:n_samples//2], 
                       name=f'Cosine Component (f={freq}, A={cos_amp})',
                       line=dict(color='blue'))
    cosine_traces.append(trace)

# Create layout
fig_time = go.Layout(
    title='Time Domain Signal',
    xaxis_title='Time',
    yaxis_title='Amplitude',
    showlegend=True,
    hovermode='x unified'
)
fig_freq = go.Layout(
    title='Frequency Spectrum',
    xaxis_title='Frequency (Hz)',
    yaxis_title='Magnitude',
    hovermode='x unified',
)

# Create time domain figure with traces
time_fig = go.Figure(time_trace, layout=fig_time)

# Add sine and cosine components to frequency figure
freq_fig = go.Figure()
for trace in freq_traces:
    freq_fig.add_trace(trace)

# Update frequency figure
freq_fig.update_layout(hovermode='x unified')
freq_fig.update_xaxes(tickangle=-45, tickformat='.1f')
freq_fig.update_yaxis(type='log')

# Show plots
time_fig.show()
freq_fig.show()

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
Cell In[2], line 61
     59     x = np.fft.fftfreq(n_samples) * sample_rate
     60     y = amp * np.sqrt(2)
---> 61     trace = go.Scatter(x=x[x > 0], y=y[:n_samples//2], 
     62                        name=f'Sine Component (f={freq}, A={amp})',
     63                        line=dict(color='red'))
     64     sine_traces.append(trace)
     66 for freq, cos_amp in cosine_amps:

IndexError: invalid index to scalar variable.

커멘트

• 디버깅하지 않았다.