4 thoughts on “Audacity Forum”

  1. Hello,

    I’m a REAL noob when it comes to Audacity and more specifically to Nyquist programming, but I’m certain a plug-in would provide the audio result I’m looking for. Much of the discussions about audio dynamic range revolve around compression of dynamics (peak limiting and expansion). What I am looking for is something to uncompress audio since in my opinion most recordings are way to compressed to begin with. This problem was actually addressed way back in 1974 in the Phase Linear 4000 which offered a peak unlimiter / downward expander. Not sure as to threshold set points but would imagine a progressive maximum peak unlimit of +3.0 db and maximum downward expansion of -6.0 db would probably work nicely to revitalise the “squashed” sound of so many recordings. I hope this makes sense. Thanks for listening.

    1. Hi Ron,

      Yes it can be done (sort of). The effect you describe is called an “Expander”.

      The theory of a “dynamic expander” is pretty straightforward, and it no more difficult to make as a plug-in than a dynamic compressor. The difficult part is achieving “musical” results.

      With dynamic compressors (and limiters) there are many parameters (the user may not have access to all of the parameters but nevertheless they are still present or implied within the algorithm). Typically a full featured compressor will have controls for Threshold, Knee, Attack, Hold, Release, Lookahead, Channel link, Peak/RMS, and Ratio. The difficulty when trying to “de-compress” is that none of these settings are known.

      We can consider an extreme example to illustrate the problem:

      “Hard Clipping”
      This can be described as a compressor (or limiter) effect that has an infinite compression ratio above a specified threshold, with instant (zero) attack, hold, release and lookahead.
      Alternatively it can be described as “chopping the peaks off”.
      “Un-compressing” audio that has been clipped in this manner is not possible because neither we, nor the computer, know what the waveform above the threshold should be like (because it is absent). The best that we can hope for is an approximation (a “guess”) of what the peaks “may” have been like prior to clipping.

      A similar problem exists with less extreme forms of dynamic compression. We don’t know precisely how the waveform has been deformed by the compression, so we cannot calculate the required gain changes to reverse the compression.

      In most cases, applying dynamic expansion to previously compressed audio provides disappointing results, precisely for the reason that we are not able to match the (unknown) parameters that were used when compressing. (Often when CDs are mastered, the compression is frequency dependent, which makes the situation even more impossible).

      Having said that, I have written an Expander plug-in which generally produces “musical” results, so I’ll post that to this blog within the next couple of days.


      1. I’d say it’s the same in this particular case. Although I also use difefrent busses when doing parallel compression. Mostly because sometimes you don’t want to send the entire drums to the compressed buss, maybe KD, SD & OH only. To do that, you would need difefrent busses, so I think it’s easier to start with two busses form the beginning. Cheers!

        1. Interesting Shahzad. How do you do that in Audacity or Nyquist? (in case you didn’t realise, that is the subject of this blog), or are you just trolling?

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.