import stft
import scipy.io.wavfile as wav

fs, audio = wav.read('input.wav')
specgram = stft.spectrogram(audio)
output = stft.ispectrogram(specgram)
wav.write('input-anchor.wav', fs, output)

'input-anchor.wav' sounds not like 'input.wav' at all.

Hi! I'm using this package indirectly through the mdct package. I get FutureWarnings that I see are already fixed in the repo, but are not released on PyPI. Could you make a new release? Thanks :)

I successfully ran the example. However, the reconstruction is not the same as the original. Perhaps due to the lack of the phase information?

i think line 39 until 47 from utils.py should be removed (the incorrect definition is shadowed now so it shouldn't affect the functions that call center_unpad)

I'm experimenting with an audio of 35751 samples (sampled at 25kHz sampling rate) using a framelength of 5000 and overlap of 16 which is 312 hopsize. It seems that there's a dimension mismatch at the line
data = data * window Is there any workaround for this issue? Thanks!

Problem

I'm not able to call the function stft.spectrogram, after following the tutorials here. Here is my console output:

makeing spectrogram for file: data/055 - Angels In Amplifiers - I'm Alright.wav
/usr/lib/python3/dist-packages/scipy/io/wavfile.py:273: WavFileWarning: Chunk (non-data) not understood, skipping it.
  WavFileWarning)
Traceback (most recent call last):
  File "/mnt/c/Users/tstan/Documents/Class spring 21/ML/Project/stft.py", line 1, in <module>
    import stft 
  File "/mnt/c/Users/tstan/Documents/Class spring 21/ML/Project/stft.py", line 18, in <module>
    specgram = stft.spectrogram(audio)
AttributeError: module 'stft' has no attribute 'spectrogram'

About my system

I've ran into this on both Windows and Linux (runnign wsl--yes it runs a python executeable on the linux side)
I've tried fixing this with running both python 2.7 and 3.7.

Any help on how to fix this?
Thanks,
-T

although this is a minor, spectrograms are usually shaped (frame, frequencybin)