What is unique about Analog Synths ?

[Sacred Synthesis]

The imperfections/variance thing can be disproven by demonstation, how is that then?

From one video you have watched!

Some people have limited ability to take on information, it proves nothing.

A closed mind sees nothing.

Not at all.  I didn't need Doty's demonstration for the one point I've been sticking to, but there it is.  My only point from it being, do those added imperfections make the sound in any way more appealing, in accord with the qualities that people call "analog" - lush, rich, warm, fat, creamy?  Perhaps you'd disagree, but I don't think they improve the sound on any level.  They only give it variations which I personally dislike.  Again, this is my main point.  But your previous post about oscillators made much more sense to me. 

If software digital synths are another creature, then I have no say about them because I've never used one.  I can't imagine being stuck in front of a computer screen while composing or performing music.  To me, it would be like walking in the woods with ear plugs and blinders on, or bringing along the most annoying person I knew; it would ruin the beauty of the experience.  So, I've only been referring to hardware digital synthesizers.  And I mentioned the Nords only as representative of generally respected VA instruments.   

If we have to go into software synthesizers, or else, $4,000+ individual hardware synthesizers, then I admittedly have to quit there.  But it does seem like an exorbitant amount of money to spend in order to emulate a little ARP Odyssey.  And even then, can I be absolutely guaranteed not to hear even the slightest digital aliasing in the upper registers?   Anyways, at that annoyance or starting price, I would probably get out of synthesizers altogether and invest in a home church organ.  But that that won't be necessary, as long as DSI is around.

[Sacred Synthesis]

The imperfections/variance thing can be disproven by demonstation, how is that then?

From one video you have watched!

Some people have limited ability to take on information, it proves nothing.

A closed mind sees nothing.

I want to apologise for my tone in this post, I was having a bad day!

No problem, Bob.   

[Paul Dither]

Not at all.  I didn't even need Doty's demonstration, but there it is if you'd like one.  My only point from it being, do those added imperfections make the sound in any way more appealing, in accord with the qualities that people call "analog" - lush, rich, warm, fat, creamy?  Perhaps you'd disagree, but I don't think they improve the sound on any level.  They only give it variations.

Well, harmonic content is certainly the other thing that's important. That in conjunction with variance explains at least attributes like "rich" and "fat."

[Sacred Synthesis]

Right, harmonic content is the key to the character.  I agree.

[chysn]

A piano sounds alive because of, among other things, sympathetic resonance. The strings that aren't being played are set in motion by vibration from strings that are. It's easy enough to imagine stronger (lower-numbered) harmonics causing vibrations of strings an octave up, then another fifth, etc., weaker vibrations as we move up the overtone series. But then the sympathetically-vibrating strings cause their own, much weaker, interactions with strings at their harmonics, and it goes on like this until the very weight of the strings themselves causes them to be immovable by further overtones' overtones. But also we can't forget other factors like air temperature and pressure, the piano's temperament (even the "perfect" octaves get stretched intentionally during tuning), the materials, the number of picture frames sitting on the piano, how far the piano is away from the wall, or the player. It all adds up to "just can't really sample a piano."

But of course you can sample a piano, and convincingly. The more stuff you mix it with, the less the nooks and crannies of sound contribute, and so we hear piano samples all the time.

I expect that analog filters and oscillators are the same way. We know them when we hear them not because of "imperfection," but because of countless interactions of circuitry that simply aren't completely modeled. Which one was the OB-8 and which one was the Diva? I have no damn clue. I totally can't tell by listening to WAV files. But get behind the same keyboard with them for five minutes each, and I'd expect any of us would have a pretty good shot.

Digital emulation of an analog synth is about as interesting to me as an analog synth emulation of my piano. I respect digital synthesis for the things that it brings to its own table. In that context, or any context for that matter, there's no mutual exclusivity. That why I absolutely refuse to post anything on this topic.

[Paul Dither]

But of course you can sample a piano, and convincingly. The more stuff you mix it with, the less the nooks and crannies of sound contribute, and so we hear piano samples all the time.

Sure, but a piano sample will always be a sample of one piano at a particular time in a particular place under particular physical conditions. Basically like a snapshot.

[chysn]

Sure, but a piano sample will always be a sample of one piano at a particular time in a particular place under particular physical conditions. Basically like a snapshot.

Which is fine* until you decide to play a chord with that snapshot. The maths of sympathetic vibration will work out a bit differently for chords, it's not just a simple sum-of-single-notes. You'll get a decent caricature, but when a sampled piano is alone it will always give itself away somehow.

* But seriously, even if you do play chords, it's still "fine." But I hope you know what I mean.

[Paul Dither]

Which is fine* until you decide to play a chord with that snapshot. The maths of sympathetic vibration will work out a bit differently for chords, it's not just a simple sum-of-single-notes. You'll get a decent caricature, but when a sampled piano is alone it will always give itself away somehow.

* But seriously, even if you do play chords, it's still "fine." But I hope you know what I mean.

Sure. I just meant on an analytical level that's what makes the difference.

[chysn]

...push down middle C, slowly, softly, so that it makes no sound. The damper pulls back from the strings, but the hammer just barely touches the strings, and there's no--or very little--sound. Keep it held down. Then, bang on the C below middle C as hard as you can, staccato, while keeping the silent middle C held.

Middle C's not so silent anymore, it rings aethereally for a good long time, or until you release the key. This sort of interaction is going on with every note and chord played.

A similar thing goes on inside analog electronics. There's simply no way it can be perfectly modeled, and your best scenario is modeling it sufficiently for a particular purpose.

[Strange Quark Star]

Even this sort of sympathetic vibration can be modeled quite convincingly, as it is in my hybrid Kawai upright.

Given enough algorithmic sophistication and processing power there's nothing you can't model or reproduce. It then comes down to the choices the modeler has made and the economy of painstakingly replicating minute details when you could just use the real thing instead, which oftentimes has much nicer physical characteristics, too.

[chysn]

Even this sort of sympathetic vibration can be modeled quite convincingly, as it is in my hybrid Kawai upright.

This was also the claim-to-fame of the Generalmusic Pro2 digital piano of the late 90s. It was actually pretty impressive.

Given enough algorithmic sophistication and processing power there's nothing you can't model or reproduce.

Everything you say is true; of course there's nothing that technically can't be modeled with enough power. Personally, I won't even rule out our entire universe being a simulation. My point is, that in 2016, partiality to analog synthesizers isn't just an irrational preference, but is based on perceptible phenomena.

[Strange Quark Star]

Given enough algorithmic sophistication and processing power there's nothing you can't model or reproduce.

Everything you say is true; of course there's nothing that technically can't be modeled with enough power. Personally, I won't even rule out our entire universe being a simulation. My point is, that in 2016, partiality to analog synthesizers isn't just an irrational preference, but is based on perceptible phenomena.

Agree completely. In theory a model could do everything, which is why I disagree with people who say "analog will always be better." The reverse—digital is just as good—is just as wrong, though, since it clearly isn't at this point. If not in modeling the exact behavior, then surely in interface design or at the very least the fuzzy feeling you get when you know you have the real deal in your hands.

[Sacred Synthesis]

Instrument emulation is only the beginning of the equation.  The next factor is the sound system, which requires the ability to perfectly reproduce all the nuances of an acoustic imitation, and can be quite expensive. 

I've played some fairly good digital pipe organs, as well as digital pianos.  They were very impressive as technical feats, but personally, I find the whole idea of imitation to be rather bland.  It's like a woman who seems beautiful...until it rains; then you see the truth in a pool.  I don't even like fake candles.  In the end, I would far prefer to play a small genuine pipe organ than a massive glorious digital imitation.   

I could use a sampled or digital pipe organ sound in my recordings, but I happily prefer not to and to use instead a different sound that has similar strengths and musical virtues, and yet, that is clearly a synthesizer, not an organ.  I simply don't want the silly experience of playing a gigantic pipe organ sound in my little music room.  Even when I design a choir, brass, or string patch, I intend the sound to be similar to the real thing as a starting point or category of tones, but by no means do I want or need to fool anyone.  What I strive for is an excellent synthesized choir, brass, and string patch, and doing so often sends me in very inauthentic directions.  Again, because the objective is not imitation.

There are elements of truth, reality, and nature that I prefer to abide by at all times.  For me, these are part and parcel of the beauty of making good music.  They are the objective part.

[Strange Quark Star]

Instrument emulation is only the beginning of the equation.  The next factor is the sound system, which requires the ability to perfectly reproduce all the nuances of an acoustic imitation. 

I've played some fairly good digital pipe organs, as well as digital pianos.  They were very impressive on one hand as technical feats, but personally, I find the whole idea of imitation to be rather bland.  It's like a woman who seems beautiful...until it rains; then you see the truth.  I don't even like fake candles.  In the end, I would far prefer to play a small genuine pipe organ than a massive glorious digital imitation.   

Oh sure, all we talked about here was comparing the raw signal output of electronic instruments, particularly synthesizers.

But it's interesting to think about the whole experience, too. I have a nice example every day when I play my acoustic piano at home or switch it into digital mode at night, which puts up a bar to prevent the hammers from striking the strings. The physical feeling is the exact same, aside from feeling the strings vibrating in my fingertips, but I only hear a (quite good) digital emulation in my headphones. When practicing a new piece this actually is "good enough" for me, but of course the clean sound of a sampled, perfectly tuned grand coming out of headphones is nothing in comparison to filling the whole room with the real acoustic sound when I actually play a piece.

Of course there are great advances in the sound system domain, too, as can be seen in Yamaha's most advanced digital grands on the piano side. On the general side you have wave field synthesis and ambisonics, demos of which I recently heard in person at an acoustics conference. Very impressive technology and interesting effects. But it's still very clearly "just an effect" and nothing really like the real experience.
At my university we have one auditorium fitted out with 1,600 individual speakers that—in theory—can reproduce the ambience of most any place, real or not; e.g. a particular gothic cathedral. It's very convincing! But still clearly just an imitation, however impressive it might be. And let's not think about the kind of money this kind of sound system might cost!

[BobTheDog]

Well it's a bit like the argument of listening to an Orchestra.

On one side you have the actual experience of listening to live players in a live space (hopefully an excellent concert hall), or listening to a recording of the same event.

Now for me nothing beats the live space, no recording system is going to give me the same auditorial experience.

But we also have reality, most listeners of music listen to music from a recorded medium. So from an "emulation" point of view the developers of "virtual instruments" need to look at the instrument they are emulating from a recorded perspective.

Nowadays guitar amp simulators are amazing, they simulate the pre amp, power amp, cabinet (and interaction between these 3), microphone, microphone position and ambient space amazingly well. On a recording 99.99999% could not tell the difference. Playing live through a PA no one could tell the difference, but still there are a subset of guitarists that would never use this technology.

[Fuseball]

I think there's a separation between what we, as musicians, hear and feel when we play our instruments and what an audience hears and feels.

As day-to-day life becomes more and more digitised and abstracted, I want the things I do for pleasure to be as authentic and sensory as possible. For example, I choose to drive a manual car with a lot of 'feel' and relatively unencumbered with gadgets. I like to deal with people face-to-face when I can. I'm one of those guitarists that relishes the unpredictable interaction between guitar and amp. I'm absolutely no Luddite but emulation or filtering of these things is profoundly unsatisfying to me.

For all their positives, in my mind digital synths are still introducing a layer of abstraction to the process of creating sound. Software synths doubly so. What drew me to analog synthesisers was that it was as close to the metal as I could get while still remaining modern and electronic. When I'm shaping a sound I can picture how I am directly influencing the electrical circuit. A huge part of the experience is interface and the responsiveness of that interface. I recently got to spend some time with a Minimoog Model D and the thing felt so alive. It felt like every switch and knob was coursing with electricity. I never get that from digital instruments.

I know that, for many, instruments are tools to serve a specific musical purpose. I guess I'm fortunate enough not to need to see them that way. I make music purely as a hobbyist and the way a guitar or a synth makes me feel when I play it is, if anything, more important than its utility. I don't derive much pleasure from the few digital synths I have, but the analogs never fail to make me smile.

[Sacred Synthesis]

Well said, Fuseball.

[BobTheDog]

I think there's a separation between what we, as musicians, hear and feel when we play our instruments and what an audience hears and feels.

I totally agree, depressing isn't it.

As day-to-day life becomes more and more digitised and abstracted, I want the things I do for pleasure to be as authentic and sensory as possible. For example, I choose to drive a manual car with a lot of 'feel' and relatively unencumbered with gadgets. I like to deal with people face-to-face when I can. I'm one of those guitarists that relishes the unpredictable interaction between guitar and amp. I'm absolutely no Luddite but emulation or filtering of these things is profoundly unsatisfying to me.

Totally agree on the car front, manual car, mid engine, high powered and rear wheel drive. The only car worth driving

The interaction between guitar and amp though has been nailed in the digital world. For example I use an Orange TH30 when I (rarely) play live but there is no way I can play this at home even though it is only 30watts, dialing it to get the sort of power amp distortion I want and all the neighbours within 200 yards will be banging on the door. So I use a power soak and to be honest that doesn't get the same feel, the air being moved by the speaker isn't the same, the interaction between the speakers in the cab and the power amp isn't the same. We are already moving away from the full blown experience.

What You can do though is buy a Kemper and profile it down the practice studio (along with my other amps), now I can get that power amp distortion but at much lower volume levels without using a power soak and without loosing the interaction between the cab and the power amp. Now I can't tell the difference between the interaction between the guitar and the Kemper compared to the guitar and the amp, apart from the air being moved that is.

Also with the Kemper I can profile all my other amps, and also I can download profiles of amps I would't even begin to try to afford. A whole world opens up.

For all their positives, in my mind digital synths are still introducing a layer of abstraction to the process of creating sound. Software synths doubly so. What drew me to analog synthesisers was that it was as close to the metal as I could get while still remaining modern and electronic. When I'm shaping a sound I can picture how I am directly influencing the electrical circuit. A huge part of the experience is interface and the responsiveness of that interface. I recently got to spend some time with a Minimoog Model D and the thing felt so alive. It felt like every switch and knob was coursing with electricity. I never get that from digital instruments.

I find it hard to see analog synths as modern, most of them are severely limited. Just taking a simple part such as the oscillators, what do you get, four distinct waveforms if you are lucky. Compare that to Omnisphere, I have no idea how many oscillator types it has, too many to comprehend. Same with filters, same with modulation ability, same with envelopes, really same with anything. The DX7 is getting old but even that is far more modern that all the recent analog synths, a different world.

For me I think in general the allure of analog synths is their simplicity not their modernity, and there is nothing wrong with that.

I know that, for many, instruments are tools to serve a specific musical purpose. I guess I'm fortunate enough not to need to see them that way. I make music purely as a hobbyist and the way a guitar or a synth makes me feel when I play it is, if anything, more important than its utility. I don't derive much pleasure from the few digital synths I have, but the analogs never fail to make me smile.

And thats where it comes down to, each man to themselves.

Everything I have said in this post only has meaning to the way I see things, I would hope others would see these truths the way I see them but I am old enough to know this will never be the case.

[chaparral]

I have been entertained by this discussion, but my answer to 'What makes analog (or digital) better would be 'The person playing it'.
I don't mean to be flippant. I spend many hours perfecting my synth voices and all too often - especially when playing with a rock band or in a place with dodgy acoustics - all my subtlety is wasted. I doubt that most of the audience would be wondering the mechanics of my synth as long as they like what I play.
My main requirement for a synth is that I can make sounds that are NOT like the ones you usually hear. Tim Turan (a famous mastering engineer) once played some of my work to a number of recording engineers and none of them could guess that I was playing a DX7 (I had programmed all my own voices). I now have a Prophet 12 and would want my sound to be equally unique. I do not believe that I could easily push analog sounds to such extremes, but If I had an analog synth I would give it a go.
I used to play with a guy who played a Woolworths electric guitar which I am sure you would all agree is no Gibson Les Paul, but I tell you sincerely, he could trash most other guitarists no matter what their equipment.

[tumble2k]

I think this thread has gone dormant, but I'll jump in. I have two comments that I may expand on more if there's interest.

Sacred Synthesis keeps saying that analog sounds warm and lush compared to digital emulations. I agree with that. There's something to be said for a system that has nothing generating frequencies other than the ones you intend to generate. This is true of acoustical instruments and of analog instruments. Not so true of digital instruments, which generates aliasing (that is filtered out) and other digital frequencies that can intermodulate down to audio range. I think these may add high frequencies that subjectively thin out the sound. There's something to be said for a system that doesn't generate extra audio signals in the first place.

The second comment I have is that analog synthesis is easy to understand and very immediate. In the end we are shaping sound. The goal is to get the sound the way we want it, and in the case of analog synthesis, the tools are oscillators, low pass filters, envelopes, and LFOs. This is a great set of tools that maps well onto what we hear and how we analyze sounds. To have an instrument with which we can modify these parameters directly is amazing. This is not so true for digital instruments, where the controller inputs go into a big black box that generates waveforms that are digitally converted back to analog. That digital instruments can actually be enjoyable to use is what is surprising to me.