Open giampierosalvi opened 5 years ago
Thank you for noticing this.
Yes, you are correct that line lines were mistakes. However, due to the nature of the regression algorithm used, this did not have a large impact on the resulting predicted brightness.
I've changed these lines as you suggested.
Andy
Hi, if I read the code in Timbral_Brightness.py correctly, line 125 should say: all_hp_centroid_tpower.append(hp_centroid_tpower) and line 132 should say: weighted_mean_hp_centroid = np.average(all_hp_centroid, weights=all_hp_centroid_tpower)
Otherwise the hp_centroid array would contain a mixture of centroid frequencies and powers, and the average in line 129 and 132 would not make too much sense.
Is this correct? Can you fix it? Thank you! Giampiero and Jérôme