Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/machow/scrampy

rearrange audio like it's no one's business.
https://github.com/machow/scrampy

Last synced: about 1 month ago
JSON representation

rearrange audio like it's no one's business.

Awesome Lists containing this project

README 

        
scrampy

=======

Scrambling audio like it aint nobody's business.



Examples

--------

scrampy uses something called blueprints.

See "How it Works" for all the juicy details on how these work.



**Ipython Notebook Examples**



- [create a blueprint for interleaved audio](examples/example_create_interl_bp.ipynb)

- [generate audio from blueprint](examples/example_use_simple_bp.ipynb)



All examples below may be run from the shell (and untested on Windows :O).



### Scrambline and Unscrambling



Below, we use the blueprint found in `examples/simple/interl_blueprint.csv`:

                  name      start      end      length      order                  0          NTF           0        8921        8921       0              1       pieman       14000       26970       12970       0              2          NTF        8921       19694       10773       1              3       pieman       26970       41620       14650       1      



This blueprint interleaves two audio narratives, to generate the interleaved audio.

The narratives are in `aud/{narrative_folder}`.



    cd examples

    scrampy make-audio simple/interl_blueprint.csv \

        "{pieman: aud/pieman/pieman.mp3, NTF: aud/NTF/NotTheFall.mp3}"



This produces interleaved audio as `default.wav` in the current directory.

Notice that there is a new blueprint titled `default.csv` produced as well.

This blueprint contains start and stop times with respect to the audio file just produced..

                  name      start      end      length      order      old_name                  0       default.wav           0        8921        8921       0          NTF              1       default.wav        8921       21891       12970       0       pieman              2       default.wav       21891       32664       10773       1          NTF              3       default.wav       32664       47314       14650       1       pieman      



scrampy can use this to pull out segments corresponding to just one story. 

For example, to align it with `aud/pieman/pieman_blueprint.csv`, which describes just one of the audio sources:

                  name      order                  0       pieman       0              1       pieman       1      



to align, we would just do..



    scrampy align default.csv aud/pieman/pieman_blueprint.csv



Note that all you need is name and order columns in the second file.

By default, the output is saved as aligned.csv, so we can use this file to reverse the process:



    scrampy make-audio aligned.csv "{default.wav: default.wav}" \

        --audout pieman.wav --bpout pieman_blueprint.csv



### Inserting gaps between each segment

Gaps can be inserted after each row using the `insert-gaps` command.

For example, if your current directory is `examples`



    scrampy insert-gaps simple/interl_blueprint.csv                 # print new blueprint

    scrampy insert-gaps simple/interl_blueprint.csv -o new_bp.csv  # save as new_bp.csv



Now, we just need to specify where the gaps file is, in addition to other audio files.



    scrampy make-audio new_bp.csv \

        "{pieman: aud/pieman/pieman.mp3, NTF: aud/NTF/NotTheFall.mp3, gap: aud/silence.mp3}" \

        --audout gaps.wav --bpout gaps.csv



Notice that when we generated audio from this blueprint, each segment+gap in `gaps.csv` ends on a 1500ms TR:

                  name      start      end      length      order      old_name                  0       gaps.wav           0        8921        8921       0          NTF              1       gaps.wav        8921       10500        1579      -1          gap              2       gaps.wav       10500       23470       12970       0       pieman              3       gaps.wav       23470       25500        2030      -1          gap              4       gaps.wav       25500       36273       10773       1          NTF              5       gaps.wav       36273       37500        1227      -1          gap              6       gaps.wav       37500       52150       14650       1       pieman              7       gaps.wav       52150       54000        1850      -1          gap      



Order for gaps is set to -1 for all by default, since it doesn't really matter.



How it Works

------------

### Blueprints

scrampy uses spreadsheets saved in csv format to construct audio from different sources.

The spreadsheets are refered to as blueprints.

For example, the blueprint in `examples/simple/interl_blueprint.csv` looks like this:



                  name      start      end      length      order                  0          NTF           0        8921        8921       0              1       pieman       14000       26970       12970       0              2          NTF        8921       19694       10773       1              3       pieman       26970       41620       14650       1      



Here is what each column describes:



* name: name describing source of audio

* start, end: interval in milliseconds where segment is located in audio corresponding to {name}

* length: length of that segment in milliseconds

* order: order of segment with respect to source in name column



Basically, each segment should have a unique combination of **name** and **order** (unless it's being repeated).



### Creating blueprints

As long as it has the columns listed above, and uses milliseconds, it's a blueprint!

There is also a helper command, `parse-splits`, that can generate a basic blueprints

from a csv that just gives the name and break points--e.g.:



                  name      break                  0       NTF       0:00.000              1       NTF       0:08.921              2       NTF       0:19.694      



Notice that here, **you can give breaks in either milliseconds or clock time**.

    

Help

----



`scrampy -h` shows possible commands



`scrampy {CMD_NAME}` shows help for a command (e.g. `scrampy align -h`)