"Spectrum follower" plug-in ?

Would it be possible to create an Audacity plug-in which would automatically modify the equalization of a track so its frequency response closely matched that of another track.

e.g. export spectrum text, (“spectrum.txt” file ), from each track, calculate the differences, then somehow import the difference data to the graphic equalizer and apply it to the track. [or automatically apply the “difference” equalization directly without the “import to graphic equaliser” step].

I was thinking such a plug-in would be useful for continuity purposes.

This idea seems somewhat similar to the “78rpm EQ Curve Generator” project and could be used for that purpose, generically : you wouldn’t need to know the different equalisation settings used by the different record labels. If you have a standard 78 recording and a recording from another 78, the hypothetical “spectrum follower” plug-in could alter the equalization of the latter to match the former.


Crinkle newspapers in front of a very high quality microphone and then crinkle newspapers in front of a crappy microphone – same environment. Apply the difference to an equilization of the crappy microphone.


There’s a proposal page for just such as thing on the wiki: http://wiki.audacityteam.org/wiki/Proposal_Plot_Spectrum_Enhancements

Koz, want to add yourself under “Developer/QA backing”?

– Bill

i can think of other uses where this would save a lot of work

assuming you have a “standard” 78 record to test with

i think i would just tweak a 78 by ear using tone controls

<<<Koz, want to add yourself under “Developer/QA backing”?>>>

I have an account. How do I do this:

==Developer/QA Backing==


Did it for you.

– Bill

This hypothetical “spectrum follower” feature could be applied in a similar fashion to noise reduction: in 2 steps …

  1. Highlight a section of one track, then press a button, lets call it “save spectrum”.

  2. Highlight another section, (typically on another track), then press an “apply saved spectrum” button and that portion of the second track is equalised so its spectrum matches the first sample.

Personally I’d be prepared to copy and paste the spectrum.txt data into nyquist code, like this

but the method I have described above would be more widely acceptable.

I’m not sure that the (limited ?) demand for such a feature would justify the effort required to create it.