Skip to content

Given a source and a target audio file, this script computes the impulse response (IR) of of a filter so that the frequency spectrum is matched as close as possible

License

Notifications You must be signed in to change notification settings

iv461/cloning_equalizer

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 

Repository files navigation

cloning_equalizer

What is this ?

This a MATLAB script that given a source audio file computes a filter that matches the frequency spectrum of another target audio file.

img

This is also called cloning equalizer or matching equalizer.

Requires Signal Processing Toolbox.

For what do I need this ?

  • If you play guitar using software modeling amps, trying to achieve a particular sound by manually adjusting an equalizer can be quite difficult
  • Luckily, most guitar amplifier software has a way to create a custom filters by using an impulse response
  • This script computes the this impulse response

Advantages

  • Works with every guitar amplifier software that supports impulse responses
  • Does not require a ton of training data, a couple of seconds are enough

How to use

First, try to get as close as possible to the target sound manually.

Then, record how your amp sounds currently (source file) and how it should sound like (target file), five seconds are enough.

Change at the beginning of the cloning_equalizer.m script the source_audio_filename and target_audio_filename variables. Both files should be mono files in WAVE format.

Then start the script in MATLAB.

The filter gets computed and the impulse response is written to impulse_response.wav.

The source with the filter applied is also saved to source_filtered_firls.wav, you can have a listen how it sounds like.

Limitations

Impulse responses describe linear, deterministic filters, they cannot describe noise generators. This means if you need a noisy target sound you would have to tune your amp approximately correct.

In other words, amplifier modeling is difficult and you cannot create a completely different sound with using only an impulse response, the source audio file must be somewhat close to the target audio.

Other related Software

You may be also interested in neural network based amplifier modeling tools like:

License

This code is licensed under MIT.

About

Given a source and a target audio file, this script computes the impulse response (IR) of of a filter so that the frequency spectrum is matched as close as possible

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages