On Tuesday, 8 July 2025 at 18:11:27 UTC, Matthew wrote:
Hi,
What do the 4096 resulting complex numbers represent?
How should I use the result to check whether the 1209Hz,
1336Hz, 1477Hz, or 1633Hz tones are present in that part of the
sound?
Thanks,
Matthew
The result of FFT should be the same as in NumPy[1] I suppose.
So D's module doesn't have the functionality that you want, so
you need to write it by yourself.
chatGPT showed this python code, that you can port to D:
```python
n = len(data)
fft_result = np.fft.fft(data)
frequencies = np.fft.fftfreq(n, d=1/fs)
magnitude = np.abs(fft_result)
# Frequencies to detect (rounded to avoid precision issues)
dtmf_freqs = [697, 770, 852, 941, 1209, 1336, 1477]
# Frequency resolution
resolution = 5 # ±5 Hz tolerance
# Find peaks near DTMF frequencies
detected_freqs = []
for target in dtmf_freqs:
idx = np.where((frequencies > target - resolution) &
(frequencies < target + resolution))
if np.any(magnitude[idx] > np.max(magnitude) * 0.1): #
adjust threshold if needed
detected_freqs.append(target)
print("Detected DTMF frequencies:", detected_freqs)
```
There is no command for `fftfreq` in D, but you can write your
own implementation, based on the NumPy documentation[2].
Also you can be interested in some other packages and examples:
- audio-formats[3] for reading and decoding different formats
- NumPy-similar package example, which is similar to NumPy
functionality[4]
From my perspective - solve it in NumPy will be safer approach,
but it should be doable in D as well.
References:
[1]
https://numpy.org/doc/2.1/reference/generated/numpy.fft.fft.html
[2]
https://numpy.org/doc/2.1/reference/generated/numpy.fft.fftfreq.html
[3] https://github.com/AuburnSounds/audio-formats
[4]
https://github.com/libmir/numir/blob/master/example/audio_separation/source/app.d
