value error

[CreativeLucence]

I have tried running this on windows 10, python 3.8.2, but I got this error: line 125, in main() line 89, in main transposed_sounds = [pitchshift(sound, n) for n in tones] line 89, in transposed_sounds = [pitchshift(sound, n) for n in tones] line 48, in pitchshift stretched = stretch(snd_array, 1.0/factor, window_size, h) line 29, in stretch s1 = np.fft.fft(hanning_window * a1) ValueError: operands could not be broadcast together with shapes (8192,) (8192,2)

I am relatively new to Python, and completely new to pygame, scipy, and numpy, so some assistance would be nice.

[CreativeLucence]

I think I fixed the issue by using np.dot(), but in doing so I have found that this code is really not suited for newer versions of numpy/python. This sort of thing comes up multiple times.

[Jholjhal]

I am also getting this error, s1 = np.fft.fft(hanning_window * a1) ValueError: operands could not be broadcast together with shapes (8192,) (8192,2)

[yoyoberenguer]

This error is due to the fact that you are trying to broadcast and array shape (8192, 2) into a smaller array length 8192. This scripts is coded for single track (mono sound track) sound and using a stereo sound track will raised the below error message:
ValueError: operands could not be broadcast together with shapes (8192,) (8192,2)

To avoid this error, you will have to transform your sound into a mono track. I would strongly recommend to use Audacity to load your sound file and export it as mono soundtrack . (Audacity) : In the menu go to Tracks -> Mix -> Mix stereo Sound to Mono

When done, export your new sound as a WAV format and use it instead of Bowl.wav sound.

Keep in mind that the mixer has to be init accordingly e.g for mono sound : pygame.mixer.init(frequency, -16, 1, 2048) for stereo sound : pygame.mixer.init(frequency, -16, 2, 2048)

Kind Regards

[yoyoberenguer]

@spacether Good day to you, The above answer provide the explanation of this issue. The ValueError is raised when numpy is trying to broadcast two arrays with different sizes/dimensions. This happen when a user is loading a stereo sound with the mono version of pianoputer.

If the user is loading a mono track sound then it will works fine for both versions of pianoputer (single & stereo mode). The stereo version has a method that can check for the sound model (mono or stereo) and process the sound accordingly.

A permanent fix would be to implement a method that check the sound model and throw and explicit error message when user is loading a stereo sound and using the mono version of pianoputer.

Best wishes,

[spacether]

Thank you for that explanation. This issue should be fixed in the next version of pianoputer which handles mono and stereo in the same python file. That update is in a PR right now.

[spacether]

The above pitchshift code no longer exists in master branch so this issue should be resolved. Testing with the included mono and stereo wav files verifies that this is resolved.