Harmonize 2 audio sounds

Hello all,

I’m using Audacity 2.0.6.

Context :

  • I have downloaded two versions of a song (.mp3 format)
  • I would like to “harmonize” frequencies of the second version, in order to match the frequencies of the first one. (Sorry if I don’t use the appropriate words, but I am a novice in the audio world).
    To illustrate this, as an example, I clearly hear that in the second version, there is more “bass” than in the first version, and I would like to change the second version to have to same “level” in order to mix them without hearing a change.

Based on my first researchs, I understand that :

Thank you in advance for all of your advices !

The Equalization effect is exactly what you need.
There is some information in the manual that you may find useful: Audacity Manual

The usual way to adjust the equalization, is “by ear”.
You may also find it helpful to compare the frequency spectra of the two tracks using the “Plot Spectrum” tool: Audacity Manual

Hello Steve,

Thank you for your prompt answer !

Based on your recommendation,

  1. I used Audacity to trace the Spectrum for each version of the song (here and here) and I exported them (with a size of 65536).

  2. Within an Excel spreadsheet, I have compared both with a diff function, and I traced the result (here)

I tried to use this to build my Equalizer to be applied on the second version, but I’m facing the 2 following issues in Audacity (see here) :

  • I can’t go further than 20kHz (but I think this is related to “audible” frequencies)
  • I can’t add more than -20dB on the 3 last frequencies, whereas I have different values to add : -52db for 12,5kHz, -40dB for 16kHz, -25db for 20kHz

How can I proceed with this stuff to add more ?

Thank you again for your expertise !

I like the approach. To a degree that approach will work, but you are running into physical limitations of the material that you are dealing with.

Looking at “audio version 1”
480134Audioversion1.png
From about 12000 Hz upward, the sound level drops very rapidly to almost nothing.
When the level drops as low as that, it is reasonable to assume that there is no meaningful audio above that frequency. That low level signal showing between 12 - 16 kHz will almost certainly be nothing more than a very faint “hiss”. You will not be able to restore any meaningful audio above 12 kHz because there isn’t any.

Steve did produce a plugin which could be used to match the frequency-response of one track to that of another in Audacity , but using it is not a simple matter * , particularly if you are a “novice”, [ trial and error with the equalizer may be easier ].

Tip: If you are going to use Steve’s “SpectrumToEqCurve.ny” plugin don’t choose more than 1024 points : it may crash Audacity. ( An equalization-curve with 1024 points may take a few minutes to calculate ).

[ * if you want track X to sound like track Y …
Step#1 obtain a equalization-curve for track X using Steve’s “SpectrumToEqCurve.ny” plugin,
invert that equalization-curve in the equalizer, then apply that inverted equalization-curve to track X so it then has a flat frequency-response.
Step#2 obtain a equalization curve for track Y , and apply it to the flattened version of track X ].