Automatic Gain Control (AGC)

Does the shruthi-1 have it? If so, is it implemented on the analog or digital side? If on the analog side, do all of the boards have it or only some? If so which ones?

I ask because I just finished building my shruthi with the SVF filter and it doesn’t seem to have it (it doesn’t keep the level the same when I turn down the resonance). I don’t have the expansion header connected yet so maybe that’s causing things not to work correctly…

First, I have never heard of such a thing as “automatic gain control” in the context of synths. I have heard about it in the context of recording equipment, where it works like a slow compressor/limiter, and it wouldn’t make a lot of sense to put similar things in a synth (how would the AGC circuit know that the drop in volume it “sees” has to be compensated and is not something programmed in the patch by the user?)

The SVF board needs the expansion connector installed, and the svf filter type selected in the filter board menu. If you don’t do that, you will hear garbage!

The classic 4-poles topology has a level drop as resonance increases “built in” into its transfer function. Not very pleasant. The SSM2044 filter board does not compensate for this; loudness decreases as resonance increases. That’s a “feature” of all SSM2044-based synths and it would have been strange to fix it. What’s the point of using a “vintage” chip and then mess with the circuit to make it sound differently from the original circuits? The CEM3379, IR3109, SMR4 boards have resonance loudness compensation. The dual SVF doesn’t need it - it sounds like any other synth with a SVF!

Sorry, I’m a total newb so bear with me. I got the AGC term from reading this thread:

http://hackmeopen.com/?page_id=193&vasthtmlaction=viewtopic&t=28.0

and assumed they were talking about what you refer to as “resonance loudness compensation”. The rockit creator also mentions this as a reference for an AGC curcuit:

Filter 1 on the SVF seems to work normally in lpf mode without the expansion connector, I just can’t switch modes. Filter 2 kinda works but the resonance doesn’t track the keyboard and again, can’t switch modes. Just FYI.

What about polivoks?

What are some other synths that have an dual SVF?

> What are some other synths that have an dual SVF?

I don’t know of any synth with two SVFs, but the fact that there are “two” of them is not a problem (after all, the default configuration on the dual SVF board, ie that of all the presets, is to have the second filter wide open - except on your board where the default routing of the filter is “random” [depends on the state of the 74hc595 at boot]).

If you look at SVF circuits (eg the oberheim SEM, the Polivoks circuit, the PAiA Fatman), all of them work like the circuit in the Shruthi-1 dual SVF board, and the transfer function is indeed the same. Cranking the resonance up doesn’t change the volume of the partials far away from the cutoff frequency, but it boosts all the partials in the neighborhood of the cutoff frequency. Think of the resonance as the (positive) level of a parametric EQ with a high Q, boosting frequencies near the cutoff. So yes, it will increase the overall signal level. But this is an absolutely normal behavior, that’s how all synths have behaved and are expected to.

I have no idea why the HackMe has this loudness drop problem, and why its author is looking into AGC, but everything about this synth is so strange…

As far as i understand Analog Filter Design (which is only about 2,4E-5481684 times what pichenettes understands) a slight drop of overall Level when turning up resonance is normal. I can confirm this for all analog/hybrid Gear i own (won’t post here because if my wife reads shell be unhappy about its latest addition…). If your Filterdesign changes Level to much while manipulating Resonance you’d rather get your Filter Design fixed than try to compensate with some kind of AGCthingy.

Thank you for clarifying. For the record I’m totally OK with it working that way, I just wanted to be sure everything was working as expected. I didn’t mean to imply there was a problem with the SVF (other than me not having the expansion connectors connected).

I agree that messing with the circuit to try to change that behavior would be strange. The level drop as resonance increases can be very cool sounding though, especially when you crank the resonance back to 0 and the bass comes back with a vengence. I think both designs have their place in a studio. I am interested in having another shruthi with the opposite behavior so I can have both :slight_smile:

So… which way will the MS-20 behave and when will you be done with it? :smiley:

Get a Polivoks in the meantime… its so red!

if it makes sense in a patch, you can build “AGC” in the modulation matrix by the way.

for example: lfo1->res & lfo1->vca

> a slight drop of overall Level when turning up resonance is normal

On a 4-pole with the “4 cascaded 1-pole LPFs with a variable path from the output back to the input” topology (pretty much all the 4-poles, at the exception of the Jupiter-6), yes.

On a 2-pole there is no such thing as a level drop ; so the MS-20-like filter will behave like the SVF when it comes to levels.

It’s all in the transfer function.

Thank you, everyone, for the tips and education!

As if i could understand that function…

So if the Jupiter-6 hasn’t this drawback - when can we expect J6 Demi Kits? :wink:

I’m sorry, I just realized I’m confused by that last statement pinchenettes. I’ll try to clarify:

There seem to be 2 types of filters:

Type 1: The resonance only boosts the level at the cutoff
Type 2: The resonance affects the overall level (level drop)

Are you saying that pretty much all 4-pole filters will be Type 2 and pretty much all 2-pole filters will be Type 1? So we have:

Type 2:
SMR-4
SSM2044
IR3109

Type 1:
dual SVF
MS-20

Unknown (are these 2 or 4 poles? their pages don’t say…):
Polivoks
CEM3379

Type 2:
CEM3379

Type 1:
Polivoks (when you do the theoretical analysis, the polivoks is a SVF, but it’s done in a very dirty way that brings a lot of distortion)