Too reiterate, try translating the values you put into your patch between dB and RMS. dB is logarithmic while RMS is linear, so your absolute gain increase of 4 times the original signal translates to 112 dB. 100 dB is the "clipping point", or a Pd signal level of "1", thus 4x the original signal is 12 dB over clipping. And your limiting threshold of 90 dB translates to a scaling factor of 0.316 times the original signal.
Confused? Want a picture?
http://www.phys.unsw.edu.au/jw/graphics/LI.GIF
(Source: http://www.phys.unsw.edu.au/jw/dB.html)
Where the author has written "power", think RMS. So twice the RMS value is an increase of +3 dB. Likewise, half the RMS values is -3 dB. Because you were calculating your limiter entirely in dB, the RMS values needed for correct limiting were all wrong (far too little) so it didn't appear that your limiter was limiting at all!
Best! Derek
Derek Holzer wrote:
But this one does. Working in RMS values appears to be the solution to your problem.
D.
Mirko Maier wrote:
Thank you, but it definitely doesn't limit the signal!
-------- Original-Nachricht --------
Datum: Wed, 26 Aug 2009 12:16:17 +0200 Von: Derek Holzer derek@umatic.nl An: Mirko Maier mirkom@gmx.de CC: pd-list@iem.at Betreff: Re: [PD] limiter questions
You may want to investigate the relationship between dB and RMS further. Attached is a working patch, however very crudely. Have a look at the numbers given by the [dbtorms] and [rmstodb] objects. Note use of [line~] as well.