Open jfsantos opened 10 years ago
Hello, Trying to get it merged with DSP.jl was actually my plan, eventually. Some help would be appreciated with testing and documentation, I have basically been going through some of matlab's signal processing toolbox and independently implementing the functions. I am fairly new to julia, so any feedback would also be appreciated.
What are the steps from here?
Hi,
you basically would have to clone DSP.jl, add the functions to a module (or separate modules, depending on their functionality), and submit a pull request. In the case of the functions you have in this directory, I would say parametric_modeling could be a module and linear_predictive_coding would fit better as a module in another package. There's some work already on code for extracting other speech features (such as davidavdav/MFCCs and davidavdav/SignalProcessing), and I think in the end we could merge everything speech-related in a Speech.jl package.
Regarding tests, we write a separate test file for each module (in test/modulename.jl); usually, we compare the output with a known output (which you can generate with MATLAB/Python, and then store it in the test/data directory as a text file). If there are known corner cases, we test them too.
Documentation is added on a per-module basis directly to RST files in the doc/ folder, as currently there is no way of adding docstrings to .jl files.
Hi James,
I found this repository through your blog posts on LPC implementations in Julia. We are working on a library for DSP in Julia (JuliaDSP/DSP.jl) and I think these functions would be an interesting addition. Would you be interested in merging these functions with DSP.jl? I can help on testing and documentation if you are interested.
Thanks!