pyAudioAnalysis-audioFeatureExtraction 錯誤糾正

1. TypeError: mfccInitFilterBanks() takes 2 positional arguments but 7 were given

The issue

---

In the function stFeatureSpeed()from audioFeatureExtraction.py make a call: [fbank, freqs] = mfccInitFilterBanks(Fs, nfft, lowfreq, linsc, logsc, nlinfil, nlogfil), but this function is only able to receive two arguments. It has the following signature: def mfccInitFilterBanks(fs, nfft)

The "extra" arguments are: lowfreq, linsc, logsc, nlinfil, nlogfil

These are all defined in the mfccInitFilterBanks function itself wit exactly the same values.

Note that there is a comment in the source code (audioFeatureExtraction.py) that says that the stFeatureSpeed() function is work in progress. so there could be many (non obvious) bugs. apart from this one.

Solutions

---

Simply remove the arguments from the call in the stFeatureSpeed function.

2. TypeError: 'float' object cannot be interpreted as an integer

---

A quick fix might be to replacenfft = Win / 2 on line 680 of audioFeatureExtraction.py
with nfft = int(Win / 2) ,and modify this line of code x = signal[cur_p:cur_p + win]-> x = signal[int(cur_p):int(cur_p + win)]in function stFeatureSpeed.

3. TypeError: 'numpy.float64' object cannot be interpreted as an index

---

add astype(numpy.int) at the end of the line 125 :
M = numpy.round(0.016 * fs).astype(numpy.int) - 1

參考資料

GitHub
音頻特徵提取——pyAudioAnalysis工具包