petewarden / extract_loudest_section Goto Github PK

The usage of the tool in Ubuntu is a little confusing for beginners. Please append in the readme file where the two command line arguments i.e. input and output directory are to be entered.

When I run this on Colab, since my data is too many, so it causes Colab disconnect my runtime. The only way I can think about is running this on my own PC.

line 71 in main.cc.
Why "fabsf(input_value * input_value)"?
Not "fabsf(input_value)"

While line 77 is "fabsf(trailing_value)"

Hello,

All my ogg -> wav files are being reported as "skipped too quiet" yet I can hear the files using the headset/microphone (without volume amplification) that was used to capture the audio.

Is this an operator/user error? Apologies for posting it as an issue. Is there a tool that you recommend to convert the wav files to a higher volume setting? (Again, my ignorance about wav files is on full display).

Kind regards.

Would it be foolish of me to assume it is as trivial as changing desired_length_ms = 1000 to desired_length_ms=10000?

I understand you mentioned in your README that this tool is not designed to be flexible, but I'm wondering if the answer here is simple.

Otherwise I will see what I can do with some digging and learning.

Thanks!

Hi Pete Warden,

Thank you so much for providing this useful tools and command dataset for us. When I clean the dataset, I notice that hundreds of audio are not complete most of which are at end of an audio. Thus, I was wondering if the segmentation does not work good or original audio does not have complete commands.