[linux-audio-dev] Re: [linux-audio-user] Re: New LADSPA Hints - clarification of definitions

Allan Klinbail allank at labyrinth.net.au
Mon Mar 3 01:44:01 UTC 2003


Hi All

Can I clarify a few definitions here based on real hardware usage.   
>  
> > > I think there was one more too, but I can't think what it was...
> > 
> > I remember one about sidechains or something. I don't know what a sidechain is,
> > though, I just recall seeing it somewhere :)
> 
> I dont think that one is practical.

A sidechain is  a send/return insert for a single device.. e.g. if you
don't have mutiband compressors (but had three standard compressors) 
but wanted to compress specific frequencies you would plug an eq rack
into your sidecahin in/out plug of your compressor..then the signal
output of the compressor is what would be compressed. Say by boosting
the bass frequencies, the bass frequencies would be compressed more than
the rest of the signal.  ..  A simpler example would be wanting to
distort a signal before compressing it. The distortion unit would be
plugged into the sidechain. I suggest going to a hardware manufacturers
site, say DBX and downloading a pdf of a compressor manual and
investigating the block diagram. 

 The reason for using a sidechain (in the hardware world)  as opposed to
plugging everything inline.. is simplification of patching and
minimising the number of cables being used.  

In software there would be no reason other than aestethics to include a
sidechain as it is possible to plug modules into each other in series
without getting ground noise from excessive amounts of cabling. So Steve
you are correct it isn't really useful.  

> For example the distinction between the control output of an envelope
> follower and the gain reduction output of a compressor, the follower is
> inteded to drive eg. a filters cutoff, wheres the gain reduction is just
> inteded to be viewed by the user as a clue as to what is going on.

Correct about control output. However gain reduction on a compressor
actually controls the output level or gain from the unit, it is not
simply for "the user as a clue as to what is going on".. Often when
compressing the resultant output may be larger than desired... When
compressing a signal,  input level is often boosted to get more of an
effect from the compressor (usually this is controlled with the input
gain on the desk) which is why the output level would often be very
high.. gain reduction is a way of counteracting this.

Here is a simplified diagram of a system using desk inserts for inline
compression within one channel of an analogue mixing desk. 


    0 --->0------>    0>  -------> 0 ------>0 ------------> BUS  
 input   input   compressor       eq      level fader  
         gain       via         section
                   insert 
                   		
So using an insert means the input gain effects the input to the
compressor and not directly to the desk channel. Without gain reduction
the compressor is likely to send a signal that is is either hot (too
high and hence distorting within the channel) or in some cases too low,
making it impossible to give it effective signal level within the entire
mix. This would be useful in a software compressor as well.

As basic audio engineering tells us, the signal must be high enough to
be clear of the ground noise and low enough not to distort within the
mix.     

I hope this helps.. 

cheers

Allan 





More information about the Linux-audio-dev mailing list