Tube Amps / Music Electronics
For current discussions, please visit Music Electronics Forum. New: view Recent Searches.
New: visit Schematic Hell!
The sunn still shines online!

Listen to great tunes streaming live right now!

ampage archive

Vintage threads from the first ten years

Search for:  Mode:  

Analog Reverb

9/14/2004 5:06 PM
Analog Reverb
Does anyone know of a way to model the mechanical properties of a reverberation spring with discrete analog components?
9/17/2004 3:47 PM
Mark Hammer

There have been a number of threads about this over the years, and a number of commercial devices that have tried to address that challenge.  
The thing about reverb vs echo is that true reverberation in a defined space has multiple asynchronous reflections. Sound bounces off the ceiling from one distance with one set of reflective properties, off the walls with another setof distances and reflective properties,and off the floor with yet another set of distances and reflective properties. And if that wasn't enough, after they've finished bouncing off the first surface, they head off to other surfaces.  
Needless to say, running a signal through a delay chip and repeatedly feeding it back to the start again will NOT mimic that very well.  
For a short while, Matsushita/Panasonic tried to tackle this one by producing several multi-tap delay chips, the MN3011 and MN3214, specifically. These are exactly the same as the chips people used for chorus pedals and analog delays, except for one critical feature. They each have multiple "taps" or outputs along the delay path such that they output repeats of the input at a variety of different and unrelated times; 6 in the case of the MN3011 and 5 in the case of the MN3214. I gather, the intent was to market them for reverb add-ons in karaoke machines (generally the biggest market for inexpensive ambience generation). In short order, though, cheap digital reverb chips became widely available, and the extra parts and cost associated with mimicking reverb in the analog domain resulted in both the aforementioned chips biting the dust. IF you can find them, they are both very expensive. MN3011's can easily go for $50@.  
Only a handful of musician-oriented products and project articles were ever produced with the MN3011 and I know of nothing that ever used the MN3214 (maybe there was something, but it has never been brought to anyone's attention here or on other forums).  
Having said all of that, I *will* note the following. Analog delay lines tend to be oriented towards achieving the widest bandwidth at all usable delays. While a bright, clear echo repeat might be useful for lots of things, real-world room reverberation doesn't sound like that. Moreover, in the real world, bandwidth deteriorates quickly as stuff bounces off furniture, carpet, people, ceiling tiles, etc. What I've found is that a bit of additional lowpass filtering of the delay signal (keep the original bright and full bandwidth), plus a bit of lowpass filtering of the regeneration signal, can make analog delays sound much more like reverb than they normally do. Not a substitue for either spring or other reverb, but certainly much less pathetic a simulation. Locate the input lug of the "repeat"/"regeneration" pot and run a small to medium-value cap (try 0.1uf for starters) from there to ground. That will take out a little more high end with each repeat such that by the 4th iteration the signal is noticeably muter. The nice thing about this is that you can dime the repeat pot and the sound never gets too cluttered. Having very different bandwidths for the dry and wet signals helps out there.  
As for mimicking the mechanical properties of springs (as opposed to simply producing reverb), that's a whole other ball of wax. Slam the springs with a hot signal and not only do you get a longer decay, but you also get changes to bandwidth, deformations of the springs, and a bunch of other stuff. I suppose the closest you might get to that with a simple analog delay might be if the amount of regeneration was tied to an envelope follower that tracked input signal level.
   Page 1 of 1