librosa VS pydub

Hello! welcome once again to the continuation of the last blog post about audio analysis using the Librosa python library, if you missed this article don't worry here you can enjoy audio analysis techniques with Librosa.

Changing pitch of the audio:- in this technique python package for audio analysis like Librosa is the best tool to go with, by adding effect on the audio pitch to create new audio data.

You might try "librosa" for tempo/beat detection:

https://github.com/librosa/librosa/blob/main/librosa/beat.py

       Track beats using time series input

https://github.com/librosa/librosa (According to their own page, though: "librosa uses soundfile and audioread to load audio files. Note that soundfile does not currently support MP3, which will cause librosa to fall back on the audioread library". This basically means that in-buffer reading isn't supported for MP3 files. They even recommend you to install ffmpeg or GStreamer so that audioread can work for MP3, too. I'm not exactly sure what they do internally but I assume it's something like: Convert MP3 to WAV using ffmpeg, then using audioread, but this is only an assumption)

Hey OP, I don't have a ton of experience with speech synthesis, but I believe this is usually accomplished with a library like librosa to separate background noises (which your robots voice and frequency would be a part of) from other vocals.

Once python is installed, you will also need to install the "pydub" package for this script to work. If you're on a Windows computer, you can do this from the command line (run the "cmd") program. If you're on mac, you can do this from the terminal. Basically, the way that you do this is using "pip" -- a "helper" program that comes with python. Once you launch the command line, just run the command python -m pip install pydub --upgrade and you should see a message showing that it successfully installed. If you're struggling with this step, just google how to "pip install python packages" and you can find a lot of beginner guides.

That is kinda what serverless functions are built for. Looks like python has some good libraries for this: https://github.com/jiaaro/pydub.

The files are big, so it's not feasible to load one in all at once. They have to be streamed/chunked somehow. (sadly, pydub doesn't support this...)

You can overlay audio files/segments onto each other using pydub.

Pydub has audio playback functionalities, though iirc pydub itself depends on 3rd party libraries to do this. Maybe check its documentation.

Seems to me a good fit for your use case would be Pydub: https://github.com/jiaaro/pydub

If there isn't sound during the black screen you can use the pydub module's pysilence.detect_silence() method to detect the start and end of video silence. And use a video editing module such as MoviePy to remove those sections.

At the moment Pydub looks like the closest match, but if it requires an ffmpeg install it's going to be a problem for pyinstaller building .exes/.apps. Same for some PortAudio wrappers I've found.

Oh, there are plenty of people like me with elaborate systems to do manipulation of audio as numpy arrays, but in terms of modules that people use, the top module for audio processing is probably https://github.com/jiaaro/pydub, which has its own interface and then there are a couple of wrappers for ffmpeg and libav, also not numpy.