iranroman / musicinformationretrieval.com Goto Github PK

Hello,
Your codes are actually written in python2. But in the readme, you told user to download anaconda python3.x. This causes SyntaxError in every print statement, and it's really annoying...

After successfully installing conda install -c conda-forge librosa ffmpeg as directed here:
https://github.com/stevetjoa/musicinformationretrieval.com/blob/gh-pages/README.md

I am able to import the following in a Jupyter notebook, without error.
import scipy, sklearn, pandas, librosa

However, I believe this line is incorrect:
"Try executing the content from https://musicinformationretrieval.com inside this blank notebook."
because you cannot execute one notebook from inside of another.

When I attempted to load ipython_audio.ipynb, this line results in an ModuleNotFoundError exception

import stanford_mir; stanford_mir.init()

The OS is Windows 7 Professional, 64-Bit
Anaconda 4.5.11, Python 3.6.5

I noticed that in this part the link to librosa's documentation of librosa.feature.mfcc is broken

I was following the tutorials step by step doing nothing but copy-pasting then I encountered a weird situation. When I was on basic feature extraction, my visual outputs of zero crossing rate and spectral centroid were different than it's supposed to be in tutorial. You can see the differances below
(my output - tutorial output in order)

I'm sure I did nothing different and restarted&did it all again three times just in case, the result didn't change. Also my input hasn't been touched and I can confirm that they match in the first visuals (plt.figure part).

what may be causing this differance? thanks in advance

Thanks for this resource.

https://musicinformationretrieval.com/evaluation_chord.html seems broken, referenced on https://musicinformationretrieval.com/index.html.

Hi! This is a really fantastic effort. Thanks for maintaining it!

I noticed the RiffStation website in the "Why MIR" section seems to no longer be active: https://www.riffstation.com/ -- thought I'd drop an issue to notify.

Location: https://musicinformationretrieval.com/why_mir.html#Commercial-Applications

In 2. Cross Validation of Machine Learning section, currently acc = sklearn.cross_validation.cross_val_score(model, training_features, training_labels, cv=5) is being used, but in scikit-learn version starting at 0.18, this is changed to sklearn.model_selection.cross_val_score().
Perhaps add if/else statement for importing the correct cross_val_score() function depending on these different version?

Thanks for the nice tutorial :)

Hi, Steve. I don't know if there is any function in librosa or some python library somewhere else to find chord progression or chord recognition. Can you give me some advice? Thank you!

Can I use stanford-mir for solving this question :
https://dsp.stackexchange.com/questions/45223/how-to-detect-and-remove-ringback-tone-and-ivrs-voice-etc-from-the-beginning-of

It would be much more accessible.

1. Create a notebook in Google Collaboratory which:

• Clone this repository
• Save the cloned files in the user Google Drive (see this tutorial)

2. Update the Readme.md with the instructions to import and run the notebook in Google Collaboratory

The notebook should be shared for other users copy to their own Google Drive.
So, there is no need to a person interested to run the notebooks install Anaconda in his computer. All notebook could be executed online.

Hey,

Great job on the examples, really helpful. I am trying to align audio sequences of the same sentence spoken by different speakers (different dialects and speed of speaking). I'm following your DTW example with different audio speeds.

As I am really new to audio processing, can you please guide me as which features would work better in this case - MFCC or Chroma or STFT?

I want to warp/align my target audio to the reference audio after computing the DTW path. How do I use the path to speed up my slow audio to the fast one or vice versa? Can this be done using librosa as well? Any example on how to save the final aligned audio?

plt.figure(figsize=(12, 4)) librosa.display.waveplot(x, sr=sr)

AttributeError Traceback (most recent call last)
in ()
1 plt.figure(figsize=(12, 4))
----> 2 librosa.display.waveplot(x, sr=sr)

/Library/Python/2.7/site-packages/librosa/display.pyc in waveplot(y, sr, max_points, x_axis, offset, max_sr, **kwargs)
438 axes = plt.gca()
439
--> 440 kwargs.setdefault('color', next(axes._get_lines.prop_cycler)['color'])
441
442 locs = offset + core.frames_to_time(np.arange(len(y_top)),

AttributeError: _process_plot_var_args instance has no attribute 'prop_cycler'

On https://musicinformationretrieval.com/python_basics.html under Package Installation section, it still recommends downloading Anaconda for Python 2.7 but not Python 3.

The new website looks great. Please note that the NMF Audio Mosaicing link is broken on the front page.

Is there any interest to put this content available on google collaboratory for easy run?

I am using Python 3.7.3 and Jupyter Notebook 6.0.0, and have installed dependencies - numpy, scipy, pandas and librosa:

import stanford_mir
ModuleNotFoundError: No module named 'stanford_mir'