Hello there,
This is my first post so I make the standard and obligatory apology in the event that I am covering old ground but here goes.
I have a work-flow/technical question regarding the correct way to interpret a frequency spectrum plot and then use it to decide where (ie which frequencies) and by how much I should apply EQ.
I will take a raw track and run a high- or low-pass filter (depending upon the frequency range of the instrument) over it to remove unwanted frequencies.
I will then run a frequency spectrum plot after listening to the track and use a combination of both my ears and the plot to determine if I need to apply some EQ.
Now let’s say my ears determine that I need the bass to be filled out a little between 50 Hz and 200 Hz and I note from the frequency spectrum that the dB level in this region ranges from (say -48dB to -36dB). So I decide to give a uniform dB boost of 3dB (for starters).
I tend to find that it then produces an excessive amount of clipping that I have to then clean up.
My broad question is this: how do achieve the bass boost but without generating all of that clipping?
Do I take a corresponding 3dB from some of the higher frequencies?
Do I apply a small amount of compression initially to give myself some head-room?
Could I apply some attenuation via the “Amplify” effect?
Could I reduce peaks by applying the “Normalise” effect?
A combination of the above?
Something else in addition to or instead of those that I have suggested?
Thanks in advance for any advice.
Grant
Your working method looks good other than the clipping issue.
There are several ways to resolve the clipping issue. Here are some of my preferred methods :
-
Use the Amplify or Normalize effects before using the Equalization effect so that the peak level is well below 0 dB.
This allows headroom if boosting frequencies is required. The advantage of this method is that if you use “Preview” in the Equalization effect (very useful) the preview will not be clipped. Amplification or Normalization can (and probably should) be used again after equalizing and before exporting to achieve the desired peak level (which should always be less than 0dB). -
Always use 32 bit float format.
This is the default for Audacity when recording or using native import. When importing files with FFMpeg (there may be other cases too) the tracks will be 16 bit, but can be very quickly converted to 32 bit float from “Tracks drop down menu > Set Sample Format” Audacity Manual
32 bit float supports audio above 0dB, so even though the sound after equalization may be clipped by the sound card, the audio data in the track is not itself damaged. This means that you can then use Amplify or Normalize to bring the peak level back down after equalizing and no harm will have been done. -
Instead of boosting some frequencies by 3 dB, if you cut all other frequencies by -3 dB the overall effect will be the same, but without clipping.
Thank you Steve for this response. I am going to apply your advice to my latest efforts. Cheers. Grant