XPost: sci.electronics.basics   
   From: kaz@kylheku.com   
      
   On 2011-11-19, Don Pearce wrote:   
   > On Fri, 18 Nov 2011 13:51:37 -0800 (PST), RichD   
   > wrote:   
   >   
   >>What's the difference between reverb, echo, and feedback?   
   >   
   > Echo is a single reflection of a sound - the kind you hear when you   
   > shout "Hello" near a cliff.   
   >   
   > If you put together many echoes, arriving from different distances   
   > into a jumble that you can't distinguish - that is reverb. You get   
   > that in, say, a large church.   
   >   
   > Feedback is a situation you only get when you have an amplifier and a   
   > speaker. The sound arriving from the speaker is a little louder than   
   > the one that originally hit the microphone, so that comes out of the   
   > speaker a little louder still. This loop will build until the system   
   > howls.   
      
   That is positive feedback. Feedback is not always positive.   
      
   If there is a lot of distance between the speaker and microphone,   
   you can hear echos rather than howls.   
      
   Feedback can be (but does not have to be) used for generating echos and   
   reverbs in effect units or software.   
      
   In digital effects units, feedback is used to make it look like there is more   
   RAM. Rather than computing all echoes from the original input function (which   
   requires enough RAM to store a window of sound representing the longest echo   
   time) the echos are faked by taking the output and feeding a fraction of it   
   back to the input. The same is done for faking long reverbs.   
      
   I have an old Yamaha unit here from 1989 which has only a 700 ms delay,   
   but the reverb can be cranked to 40 seconds, haha. Any sample you hear beyond   
   700 ms, related to the original signal, has already been through the digital   
   mill and is reappearing via feedback. Accordingly, the reverb starts to sound   
   like crap beyond 2 seconds.   
      
   A honestly modeled 40 second reverb would actually have an impulse response   
   sample of 40 seconds from a nice sounding hall, and use a 40 second window of   
   the input to do the convolution.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)   
|