Diva Vs. Real Analog

[Urs]

Does it cover the crossover distortion from the two amps and the feed-thru both signal and CV, forward via the feedback path?

I don't know what that means. What are the two amps and what signal is supposed to bleed through the feedback path? The feedback path is buffered... there is no audible bleeding.

We concentrated on the things you can hear and see. E.g. put the filter to self oscillation and see how the shape matches that of an MS-20 rev.1. Add a low frequent oscillator input to hear the cutoff modulation which matches that of a MS-20.

That's all it takes. From then on the audible differences are in the same ballpark as the differences between two MS-20s of the same revision.

[aciddose]

The feedback path isn't buffered in the MS-10.



You should see the signal passes forward just as easily as backward.

The only cancellation is due to the opamp controlling the output via negative feedback, although this isn't perfect due to the capacitor which has a non-zero resistance.

So you get a small amount of high-frequency leakage fed forward via the feedback path.

It's actually a two-way

Pass over-driven (maximum voltage, +/- 15, white noise is best) into the low-pass input and adjust the res anywhere, you'll get significant hiss with the cutoff at minimum.

https://soundcloud.com/aciddose-1/leakage
You get a lot of LFO leakage via the rails here too although this is likely an issue with the layout of my circuit (dangling PCB with wires not EMI-proof) but the noise is present in a proper circuit, even the MS-10.

Another property that I've found is amazingly difficult to get right is the way the input signal interacts with the filter self-oscillation. This is related to all the non-linearity and so of course if you can't get that exactly right this won't be produced the same way.

https://soundcloud.com/aciddose-1/analog-heartbeat

The difference though is so radical when you sweep the cutoff around below the input signal frequency it is really hard to say anyone wouldn't be able to tell. It's like a completely different drum-beat.

The cutoff is modulated by a square-wave, only adjusting the cutoff by a few cents.

[Urs]

Pass over-driven (maximum voltage, +/- 15, white noise is best) into the low-pass input and adjust the res anywhere, you'll get significant hiss with the cutoff at minimum.

I'll verify that once the MS-20s are set up in our new office. Can't say I recall that effect, but who knows?

[aciddose]

The MS-20s, especially a modern remake may not be the same circuit as the original MS-10 rev2.

You couldn't really say for certain without testing the original.

In any case this is entirely possible to model, just impractical to get perfect. I don't demand perfection although it would be nice to get something that I can use without constantly saying "this sounds completely wrong".

I've never heard an example that doesn't make me say that. I have little faith due to this.

I can do this myself if I'm going to burn a whole core on a single voice, and in that case why on earth wouldn't I just use the hardware with fifty other properties which are also hard to emulate? We're not even touching on the other aspects like the physical.

Anyway that's all way off-topic. All I wanted was a demonstration of a vaguely similar effect as the one I posted in my example. I'd have even been reasonably satisfied if someone posted something like "Sallen-key", which is software although not software I'm satisfied by, I'd at least be satisfied enough not to continue to push that there are "major" differences.

For me to be fully satisfied enough to give up on electronics however would require a damn near perfect model which I know for a fact simply isn't possible on current computers. It may very well become possible in the near future but it sure isn't right now.

[Urs]

The MS-20s, especially a modern remake may not be the same circuit as the original MS-10 rev2.

To add to our street cred, we have two original vintage MS-20s, a Rev1 and a Rev2. We also have a Monotron, but that's something entirely different, sound-wise.

(that said, we don't have any MS-10s... as much as I'd love to, I can't possibly stuff any more vintage synths into our office space)

[aciddose]

https://soundcloud.com/aciddose-1/interaction

This one demonstrates the crossover distortion I mentioned. The opamp I'm using is not an RC4558 but instead a discrete opamp.

(You may find this useful: http://www.electrosmash.com/jrc4558-analysis )

Notice the leakage goes to near zero while the oscillation stops.

The oscillation stops at very low frequency due to the high-pass in the feedback path, of course. This means the whole feed-forward path is zero due to the cross-over distortion.

The result is that the only signal you get is due to feed-forward + any PSU noise leakage due to various reasons, partially EMI, partially the opamp circuit, partially other circuits ahead of it (VCA, etc).

https://soundcloud.com/aciddose-1/discrete-opamp

I've normalized it myself so you don't need to download the flac and do it. This sound you should recognize immediately as the rectified 60hz, high-pass filtered + various sources of noise.

This becomes really important when you're compressing the output and the cutoff goes to zero. Real thing = zzzzzzt!, software = awkward silence.

[Teksonik]

What is lost? The fact you just said exactly the same thing I said in the text you quoted, then claimed it was lost somewhere?
The only way you could be attempting to argue with me by telling me exactly the same thing I just told you is if you are in some kind of deluded "us vs. them" state of mind.
I think that is what is lost with you people.
Just think about it for a moment please. It seems you're trying to argue with someone who just agreed with you, by agreeing with them!

Actually read my post.....I replaced hardware with software. Apparently that was lost on you..... just think about it for a moment please............

Here's what I really posted....bold to make sure you don't miss it this time.......

The point that is always lost in these arguments is that I would never be able to replace some of my softsynths because no hardware exists (nor does it seem likely to be practical any time soon) that could be considered even similar, let alone close enough that I couldn't tell the difference.

[Urs]

@aciddose I just tried to recreate your "interaction" example and I think it's possible to come quite close in Diva. But I could only reproduce it without any highpass filter before the lowpass. What kind of architecture are you using?

[Teksonik]

My point being that "hardware" or dedicated units are no longer the zenith of synthesis. "Softsynths" or computer based synths are now the gold standard.

The hardware fans can say "Ooooh software can't do "zzzzzzt" so it's inferior". But hardware can't even dream of creating the sounds of some modern softsynths.

I don't care about 40 year old zzzzzzt but then I don't restrict my definition of synthesis to a saw wave through a low pass filter...........

(Oh and as I said before I still own 6 hardware synths but they have been relegated to collecting dust)

[aciddose]

@aciddose I just tried to recreate your "interaction" example and I think it's possible to come quite close in Diva. But I could only reproduce it without any highpass filter before the lowpass. What kind of architecture are you using?

I'm using exactly the same circuit as the MS-10, only invert the transistors used for transconductance, add a high-pass capacitor at the input to the opamp, replace the opamp with a discrete version (as plain as possible), remove the active diode clamp, place a "normal" diode clamp in the feedback path after the res pot.

Okay not really "exactly the same" but I wouldn't claim I invented the circuit, it's only a minor modification of the KORG35.

https://soundcloud.com/aciddose-1/crossover-distortion

I've done a bit more recording of the crossover distortion effect. This is with the VCA completely off, the VCA CV input is grounded.

I'm playing with the res pot, cutoff and input level. Crossover distortion is sort of like the analog equivalent of a bit-crusher.

It's 60db gain to bring it up to audible levels. So in other words the noise and feed-thru effects are all below -60db. Not important until you really want a pure sine/etc (low-pass filter with cutoff really low) or you do something like use a compressor on the output.

[zerocrossing]

well, if you are happy driving a lamborghini computer simulation than driving the real thing, sure stick with virtual synths.

If virtual Lamborghinis would feel so good that many people would have trouble distinguishing between real and virtual one, I'm sure many more would enjoy Need for Speeds and whatnot.

The car analogy doesn't really hold water. A real car can actually be driven in the real world; the virtual car can't. By comparison, both real and virtual analog synths can be used to create music.

stfu dude, a virtual car ATTEMPTS to simulate the experience of driving, by showing you a simulated road, sky, trees, scenery on your monitor. i dont see the difference in that vs. interacting with a virtual analog on your monitor.

there is a feeling you can't get with a VA that you get interacting with a real synth sitting in front of you.
if you do, then thats the real schizo in a person taking over.

while i love my VA's Charlatan, Tal Elektro, Asynth and MinimoogV... i always
find myself going back to my AX80 just to feel that interaction with its buttons and switches and big data knob. VAs may have equalled real analogs sonically but not in many other ways.

Sorry, you're analogy is horrible. I drive a car, play racing games (I love Need For Speed Rivals) own actual analog synths and plug ins. No one (sane) has ever said, "I need to go to the market and pick up groceries for the week. I'll just go over to my XBOX." or "I feel like taking my Lambo out for a joyride in some underground race club, but I don't want to suffer any consequences for my actions." or even, "Should I buy Forenza Motor Sport or a Lotus?" Your analogy breaks down because virtual cars and real cars do not really provide the same answer to a question.

Now, I'm not saying that with significant simulation tech you couldn't get a car simulation to feel really good and perhaps it would even help you train for a car race, but who's got that tech? I know militaries use things like that for pilot training, but again, the real use for a plane, beyond training, is to go somewhere, just like the real use for a virtual instrument or hardware instrument is to produce a sound for expression. So, it is valid to ask, "Should I use Diva or an OB-x as my main instrument" because they both provide the same thing, even though how they do it is very different.

A real lambo is real... a virtual one is not.
A real Analog synth is real...a virtual one is not.

you can chose to use either one.
take your picK and STFU

OK, well then by what you're saying you should probably be able to guess which kind of synth, software or virtural a track was made with say... 70% of the time. Just like you could guess how I got to your house with a video game or with an actual car. They're right, you should join MENSA!

[wagtunes]

You know, I am so tempted to drag out my hardware synths and make a song with both hardware and software synths and see if anybody here can even come close to guessing what synths I'm using.

That will probably shut everybody up.

But sadly this tiresome debate is just going to go on and on and on until our sun goes nova.

Back to making music.

[chk071]

You know, I am so tempted to drag out my hardware synths and make a song with both hardware and software synths and see if anybody here can even come close to guessing what synths I'm using.

That will probably shut everybody up.

No way.

[wagtunes]

You know, I am so tempted to drag out my hardware synths and make a song with both hardware and software synths and see if anybody here can even come close to guessing what synths I'm using.

That will probably shut everybody up.

No way.

I am just so tired of this whole analog/digital/hardware/software blah, blah, blah.

When did we stop just making music?

[JerGoertz]

You're a little bit new round here wagtunes, but I think you're starting to catch on. KVR isn't a forum about making music, it's a place where people come to compare gear and debate about peripheral issues which have about as much direct relevance to actually making music as the color of the cupholders have to do with how to drive an automobile.

Welcome!