|
Post by feijai on Oct 8, 2023 9:57:56 GMT
I apologize if this has been answered elsewhere: I couldn't find it. I don't have an oscilloscope handy so I can't determine the answers.
Given two CV inputs A and B, and a Round(x) function which rounds to either 5V or 0V
- I assume AND just does (A / 5v * B / 5v) * 5v ? So it can't do true ring modulation on audio because there's no negative voltage? That is a pity: it'd have been nice to have an Audio/DC switch, where Audio would subtract 2.5v, then do the multiply, then add 2.5v again. Or does AND actually do Round(A / 5v * B / 5v) * 5v, that is, pushing to 5V or 0V? That'd be unfortunate.
- I don't see any way to do OR with combinations of INV/NOT, AND, NAND, or XOR. Is there a way to do it (shy of buying the OR module).
- Is INV doing 5V - signal, or is it doing a true NOT, as in Round(5V - signal)? I probably could figure this out from an LFO but I thought I'd be lazy and ask here.
|
|
|
Post by rockysmalls on Oct 8, 2023 11:12:20 GMT
I apologize if this has been answered elsewhere: I couldn't find it. I don't have an oscilloscope handy so I can't determine the answers.
Given two CV inputs A and B, and a Round(x) function which rounds to either 5V or 0V
- I assume AND just does (A / 5v * B / 5v) * 5v ? So it can't do true ring modulation on audio because there's no negative voltage? That is a pity: it'd have been nice to have an Audio/DC switch, where Audio would subtract 2.5v, then do the multiply, then add 2.5v again. Or does AND actually do Round(A / 5v * B / 5v) * 5v, that is, pushing to 5V or 0V? That'd be unfortunate.
- I don't see any way to do OR with combinations of INV/NOT, AND, NAND, or XOR. Is there a way to do it (shy of buying the OR module).
- Is INV doing 5V - signal, or is it doing a true NOT, as in Round(5V - signal)? I probably could figure this out from an LFO but I thought I'd be lazy and ask here.
all interesting questions ... i look forward to hearing the answers also
|
|
|
Post by robertlanger on Oct 8, 2023 11:27:50 GMT
The LOGIC module is a bare digital, boolean logic module, your example of AND is correct. You can use it for audio; I personally prefer the XOR (exclusive or) for very gritty stuff. But yes, it's not a ring modulator which works analog / with analog values. You can create a logic OR with the LOGIC module by using so-called NAND-Logic; Wikipedia has a good explanation: en.wikipedia.org/wiki/NAND_logic What you do is inverting both signals by using IN3 / IN4, feed INV3 / INV4 into the inputs 1A and 1B; the output NAND 1 gives you the OR then. ...boolean logic is amazing, isn't it? Basically, with NAND gates, you can build EVERYTHING digital, even a complex processor! Same goes for NOR gates; e.g. the Apollo guidance computer that controlled Apollo 11 at the moon landing was built entirely with NOR gates (with 3 inputs each, and 4100 of them) and ran real programs stored in a ROM memory. See en.wikipedia.org/wiki/Apollo_Guidance_Computer
|
|
|
Post by feijai on Oct 8, 2023 14:28:57 GMT
The LOGIC module is a bare digital, boolean logic module, your example of AND is correct. I've been doing some testing and it appears to me that AND is rounding. That is, it is doing Round(A / 5v * B / 5v) * 5v, where Round(x) = 5V if x >= 2.5V and 0V if x < 2.5V. That's fine for digital signals but makes it so it can't really do even half-ring modulation. Is this correct? :-( It also appears that INV is doing rounding as well, that is, it's Round(5V - signal), so you can't use it as an inverter as might be used in an attenuverter -- for example to change an LFO's ramp to a saw. Is this correct? :-( I thought I had worked through all the gate options, even with NAND, and assumed that you needed three NANDs to do an OR, but no, I forgot that you've got two NOTs (INV) there as well, so you can just do OR(A,B) = NAND(INV(A), INV(B)). I'm a computer science professor, so it's a bit embarassing I'd not worked it out right.
|
|
|
Post by pt3r on Oct 8, 2023 15:01:22 GMT
I’m sure the inverter on the ATT module does not round i.e. It does invert the signal, That’s why it can be used for side chaining and subtracting voltages.
|
|
|
Post by tIB on Oct 8, 2023 16:28:56 GMT
The LOGIC module is a bare digital, boolean logic module, your example of AND is correct. I've been doing some testing and it appears to me that AND is rounding. That is, it is doing Round(A / 5v * B / 5v) * 5v, where Round(x) = 5V if x >= 2.5V and 0V if x < 2.5V. That's fine for digital signals but makes it so it can't really do even half-ring modulation. Is this correct? :-( It also appears that INV is doing rounding as well, that is, it's Round(5V - signal), so you can't use it as an inverter as might be used in an attenuverter -- for example to change an LFO's ramp to a saw. Is this correct? :-( I thought I had worked through all the gate options, even with NAND, and assumed that you needed three NANDs to do an OR, but no, I forgot that you've got two NOTs (INV) there as well, so you can just do OR(A,B) = NAND(INV(A), INV(B)). I'm a computer science professor, so it's a bit embarassing I'd not worked it out right. Not sure I'm following correctly but it seems you might be confusing the inversion here - as I have it it's a digital inversion, meaning it's 1 or 0. 1 when your input signal is 0 and vice versa. I believe NAND is your friend for logic based ring modulation.
|
|
|
Post by feijai on Oct 8, 2023 22:30:21 GMT
I’m sure the inverter on the ATT module does not round i.e. It does invert the signal, That’s why it can be used for side chaining and subtracting voltages. Lacking an oscilloscope I can't verify precisely. But on my logic module, if I route a sawtooth wave into the inverter, and then take the result out to CV a filter, it sounds very much like the filter bounces back and forth between two positions. But if I put the saw directly into the filter, the result is smooth interpolation. From this I figured that the inverter takes any value X between 0 and 5v, but instead of changing it to 5-X, it changes it to ROUND(5-X), where Round(u) is 0 if u < 2.5, else 5. It looks like AND does as well. This would make sense for digital logic but is unfortunate for analog. You're saying that this is not the case? I would like to be wrong!
|
|
|
Post by pt3r on Oct 9, 2023 6:00:41 GMT
I say the inv on the ATT/2V module is a true inverter that does not simply swap between 0 and 5 volt. Feed an lfo into it and the inv output will produce the inverted lfo, same thing goes for other cv sources. I am absolutely sure about that one since I checked it on my meter module. Whereas the inv on the logic modules is just more like a not gate that produces 0v or 5v depending on the voltage you feed into it.
|
|
|
Post by feijai on Oct 9, 2023 11:24:18 GMT
I say the inv on the ATT/2V module is a true inverter that does not simply swap between 0 and 5 volt. Feed an lfo into it and the inv output will produce the inverted lfo, same thing goes for other cv sources. I am absolutely sure about that one since I checked it on my meter module. Whereas the inv on the logic modules is just more like a not gate that produces 0v or 5v depending on the voltage you feed into it. Thanks. I just did another verification. NAND, AND, XOR, and INV are all rounding on the LOGIC module -- that is, their outputs are always fixed high or low (probably 0 or 5v). So perhaps INV would be better called NOT. Useful for certain cases but not in my situation: I gotta go get an attenuverter!
|
|
|
Post by tIB on Oct 9, 2023 11:30:22 GMT
I say the inv on the ATT/2V module is a true inverter that does not simply swap between 0 and 5 volt. Feed an lfo into it and the inv output will produce the inverted lfo, same thing goes for other cv sources. I am absolutely sure about that one since I checked it on my meter module. Whereas the inv on the logic modules is just more like a not gate that produces 0v or 5v depending on the voltage you feed into it. Thanks. I just did another verification. NAND, AND, XOR, and INV are all rounding on the LOGIC module -- that is, their outputs are always fixed high or low (probably 0 or 5v). So perhaps INV would be better called NOT. Useful for certain cases but not in my situation: I gotta go get an attenuverter! I think inversion labelled as such on a digital logic module is appropriate tbh - it's a logic inversion. Serge does it the same fwiw. The wonkystuff 4vca is somewhat useful for inversions too fyi- I often use it with the 4att/mix to invert signals rather than as a vca.
|
|
namke
wonkystuff
electronics and sound, what's not to like?!
Posts: 686
|
Post by namke on Oct 9, 2023 11:32:48 GMT
I suppose that it depends on your experience, but I never considered the LOGIC blocks as anything other than digital, Boolean logic — inverter is a common term when describing such gates (even though some implementations like the CD4069 actually have a small linear region!)
Interesting to consider the binary operation as ‘rounding’ too; I’m wondering whether this train of thought might be interesting in some cases 🤔🤔🤔
|
|
|
Post by rockysmalls on Oct 9, 2023 11:47:29 GMT
I say the inv on the ATT/2V module is a true inverter that does not simply swap between 0 and 5 volt. Feed an lfo into it and the inv output will produce the inverted lfo, same thing goes for other cv sources. I am absolutely sure about that one since I checked it on my meter module. Whereas the inv on the logic modules is just more like a not gate that produces 0v or 5v depending on the voltage you feed into it. Thanks. I just did another verification. NAND, AND, XOR, and INV are all rounding on the LOGIC module -- that is, their outputs are always fixed high or low (probably 0 or 5v). So perhaps INV would be better called NOT. Useful for certain cases but not in my situation: I gotta go get an attenuverter! is it the 1u Logic or the tiny uLogic? .. or are they exactly the same circuit/outputs?
|
|
|
Post by keurslagerkurt on Oct 9, 2023 13:18:35 GMT
I suppose that it depends on your experience, but I never considered the LOGIC blocks as anything other than digital, Boolean logic — inverter is a common term when describing such gates (even though some implementations like the CD4069 actually have a small linear region!) Interesting to consider the binary operation as ‘rounding’ too; I’m wondering whether this train of thought might be interesting in some cases 🤔🤔🤔 I never considered it as rounding either, I supposed there was just a bunch of CMOS CD40xx chips in there for the different operations, so I only supposed there might have been some hysteresis involved? Still got to experiment with the CD4069 further.. I've got a bunch of them on the breadboard and tons of circuits online which use it to do filters, weird VCA's, even someone who made a 'big muff' fuzz pedal with almost exclusively CD's! Crazy chip.
|
|
|
Post by feijai on Oct 9, 2023 19:41:13 GMT
is it the 1u Logic or the tiny uLogic? 1U. Don't know about uLogic, though I presume it'd be the same.
|
|
|
Post by funbun on Oct 10, 2023 22:25:02 GMT
Wait, I can get ring modulator-like results with NAND?
|
|
|
Post by tIB on Oct 11, 2023 5:32:04 GMT
Wait, I can get ring modulator-like results with NAND? Don't think it's a true ring mod, though will obtain similar results. I believe aro used and gates for the ring mod in a few classics (2600 and odyssey to name two). So yeah - feed it square waves!
|
|
|
Post by feijai on Oct 11, 2023 7:20:10 GMT
Wait, I can get ring modulator-like results with NAND? It depends on what "like" means.
Ring modulation is very simple: you are taking two (signed) waves and multiplying them. For AE Modular, this would mean that if you have wave A from 0..5V and wave B from 0...5V, you first change them to -1V...+1V respectively, then multiply them, then change the result back to 0..5V. As in: ((A/2.5 - 2.5) x (B/2.5 - 2.5)) x 2.5 + 2.5. It's called ring modulation because it's classically done in analog with a ring of four diodes.
In binary, AND is just a multiply, and so if AE Modular were signed, like -5V to +5V, AND could be implemented in such a way as to do ring modulation. But it's not -5V to +5V. Furthermore, LOGIC's AND is rounding -- after multiplying two results it's forcing them to go to either 0 or 5V at the output. This means that the result is basically a pulse wave of sorts. You can get a ring-modulated output only if you fed two *pulse* waves as input.
I would not call AND or NAND, as implemented in LOGIC, as ring-modulation-like. But they can certainly do interesting audio-rate combinations. It's be neat to see a ring modulator (maybe throw in AM and various kinds of soft/reversing sync) as a module.
|
|
|
Post by MikMo on Oct 11, 2023 11:34:39 GMT
Wait, I can get ring modulator-like results with NAND? I think that if you stay in the logic / SQR wave domain XOR is what gives you "ringmod". Like the MS20 ringmod, which as i remember it, is made up of logic gates that in the end is equivalent to XOR. Try to run 2 sqr waves through the XOR in the logic module, then take the XOR output into the other XOR input together with a third sqr wave. Now carefully adjust the frequencies of the 3 sqr waves. You can get some crazy flanger like sounds, and a lot of digital hell. Very raw and noisy, but not the metallic bell like sounds one might expect from a real ringmod.
|
|
|
Post by leethargo on Oct 12, 2023 5:48:05 GMT
Since AM (amplitude modulation) was mentioned, this can actually be done with a VCA, by putting another audio signal in the CV in, right? I guess this is quite close to ring mod.
|
|
|
Post by tIB on Oct 12, 2023 6:06:14 GMT
Since AM (amplitude modulation) was mentioned, this can actually be done with a VCA, by putting another audio signal in the CV in, right? I guess this is quite close to ring mod. Somewhat similar results but different in that the carrier frequency is preserved along with the sidebands. In RM that isn't the case (I think!).
|
|
|
Post by tIB on Oct 12, 2023 6:11:56 GMT
I find bitcrushing is another that generates results in a similar ballpark at less crazy settings. Maybe worth mentioning you can do that with an S&H at audio rates.
|
|
|
Post by feijai on Oct 12, 2023 7:34:45 GMT
Since AM (amplitude modulation) was mentioned, this can actually be done with a VCA, by putting another audio signal in the CV in, right? I guess this is quite close to ring mod. Somewhat similar results but different in that the carrier frequency is preserved along with the sidebands. In RM that isn't the case (I think!). That is correct. AM is the same thing as RM with the original carrier mixed in. Fun fact: the Kawai K1 had RM. The Kawai K4 had AM, which the Kawai engineers thought was an improvement. But it wasn't. RM is far more desirable than AM because in AM you can't get rid of the original sound. As a result the K1 has a cult following just for the ringmod, while the (superior) K4 is largely forgotten.
|
|
|
Post by tIB on Oct 12, 2023 11:40:31 GMT
Somewhat similar results but different in that the carrier frequency is preserved along with the sidebands. In RM that isn't the case (I think!). That is correct. AM is the same thing as RM with the original carrier mixed in. Fun fact: the Kawai K1 had RM. The Kawai K4 had AM, which the Kawai engineers thought was an improvement. But it wasn't. RM is far more desirable than AM because in AM you can't get rid of the original sound. As a result the K1 has a cult following just for the ringmod, while the (superior) K4 is largely forgotten. I've had several of the K's - the k1 which has a really lively digital sound, albeit without filters unfortunately, the k3m which was lovely - really nice SSM filters but limited in terms of range of parameters (0-31 was it?! Who knows why) - and the k5m which somehow managed to do bass pads due to its thinness, aliasing and additive engine - I used to love sending it really low chords and hearing it unfold.
|
|
|
Post by feijai on Oct 12, 2023 12:31:58 GMT
I have owned a K1r, K4 (my first synth, still own it), and K5m (and wrote patch editors for all of them), and have a close friend with a K5000s.
The K4 is a fine rompler with a first-rate filter, really very good. But it has AM. :-( Ring Mod is the K1's killer feature.
The K5 is, in my opinion, a terrible sounding synth, noisy even with custom modifications, and thin. But the K5000 series, which was Kawai's apology for the K5, are *awesome* synthesizers, highly desirable even now.
I would love to get a K3. It's a great hybrid synth with some of the best sounding bass of any synth ever.
|
|
|
Post by tIB on Oct 12, 2023 16:45:28 GMT
I have owned a K1r, K4 (my first synth, still own it), and K5m (and wrote patch editors for all of them), and have a close friend with a K5000s.
The K4 is a fine rompler with a first-rate filter, really very good. But it has AM. :-( Ring Mod is the K1's killer feature.
The K5 is, in my opinion, a terrible sounding synth, noisy even with custom modifications, and thin. But the K5000 series, which was Kawai's apology for the K5, are *awesome* synthesizers, highly desirable even now.
I would love to get a K3. It's a great hybrid synth with some of the best sounding bass of any synth ever.
The thinness is what I loved about the 5 - it was dreadful at most things but so thin sounding you could stack five notes in the lowest octave and get paid like weirdness from it. Odd! The 3 is certainly the one I miss - and you're right in mono mode it was an absolute bone shaker of a bass monster!
|
|