#568

In the Prologue (and the Minilogue XD as you mention) you have analogue and digital versions of the same waveforms to play with in the same unit, they both have their pros and cons, I wouldn’t want to be without either.

#569

Completely disagree, haha. They sound different, but not equal, IMVHO. I find analogue a more engaging listen than digital, 100% of the time. I say this as someone who has spent the last three decades and more trying to find the best in audio quality, in both domains. I love them both very much, but I’d never say one has the other “beat”. They are different, and can be exceedingly complimentary. No need to chose one over the other. Enjoy both. I think it’s a fool’s errand to try and pick one over the other. My UDO Super 6 is a perfect example of this credo. :wink:

7 Likes
#570

The distinction is going to slowly diminish over time, but never disappear. Digital recreations of analog eccentricities are improving, and I think the list of things that analog truly sounds different with are going to become more circumscribed. That’s certainly been the case in the last 50 years. What’s really interesting is how vintage digital is often harder to recreate than vintage analog.

BTW, I am actively developing simulations of analog gear right now, and the biggest differences happen not because things are impossible, but more software developers are fine with abstractions and “good enough”. Most people are just copying freely available papers and not actually testing them against actual analog parts. Hell, that’s probably the biggest difference between the Prophet 6 and the Prophet 5 rev 4, from everything I’ve heard from people at Sequential; they had multiple P5s to test against.

3 Likes
#571

I covered this in the “audio voodoo” bullet point :smiling_imp:

Joking aside, in the context of a mix, say, Moog’s “Minimoog Model D” app won’t be distinguishable from Native Instruments’ Monark or the real deal. If any of those “sounds better” then it’s fully due to mastering differences.

Then again, maybe we’re actually agreeing with each other and what you mean is this: analogue hardware, due to its tactile look and feel, invites more modulation by the musician. So digital counterparts sound more boring. Then you’ve got things signal cross-bleed, unstable oscillators, non-linear sweeps, and so on. Those can be fully recreated in software but as @Mangrove is saying, often this level of detail is no longer interesting.

You can say there’s no need to choose digital over analog but realistically I think everybody does. For example, the Moog One is an expensive piece of gear, big and heavy, has cooling fans, and a built-in keyboard. Each of those characteristics might be a deal breaker for somebody. Or maybe you just need more than 16 voices?

I mean, in the end if you’re recording onto a bunch of bits on an SSD, it becomes digital anyway. There is literally no reason for analog sources to be somehow magically better. So I think in the end software will entirely dominate this industry.

(I’m coming back to the Moog examples since I’m most familiar with their hardware.)

#572

I thought the use of digital and analogue in a single unit was really intriguing when I first encountered the Peak/Summit. It very much felt like it was digital where it was most effective and analogue where it counted most (but I appreciate that this is entirely subjective!)

2 Likes
#573

Not sure really. If I wanted to boil it down, I’d say I can definitely hear a difference, don’t generally prefer one over the other, enjoy them both immensely, but find analogue a more engaging listen. Just IMHO.

1 Like
#574

Using absolute terms - “period”- to state a subjective belief - “digital has beat analog”, is not good practice. It generates needless controversy.

Characterizing a subjective belief that you do not hold by calling it “voodoo”, or characterizing it as a form of nostalgia (your first three points) is also bad practice. It comes off as arrogant and dismissive of other people’s taste and aesthetic. Preferring analog for some applications myself, it certainly comes off that way to me. I have no interest in nostalgia and no belief in voodoo, audio or otherwise.

10 Likes
#575

I certainly agree that analog vs digital is a matter of taste, and I’m willing to believe that virtual analog doesn’t sound exactly like real analog. For me, though, the reason I like digital is because of the forms of synthesis that are purely digital, such as granular synthesis, digital FM (very different to my ears than analog FM), and wavetable synthesis. There’s no analog version of, say, the audio rate wavetable morphing in Iridium or Disting. A lot of people aren’t into that kind of weird heavily digital sound, and that’s totally ok, but I love that stuff!

4 Likes
#576

Yeah I love that stuff also. Things that straddle the line, like Lunettas and many of the NLC modules also fascinate me.

1 Like
#577

I am not arguing about taste and aesthetic. I am arguing that modern VA in software is not distinguishable from pure analogue signal chains. In other words, I am not arguing about taste but about tap water tasting which lends different results based on what logo is on the bottle. (Penn and Teller had a great episode on it in their skeptics show “Bullshit!” 15 years back.)

Looking at how far physical modeling has come with, say, Pianoteq sounding good enough for Steinway to license them the name of its flagship grand piano model, we’ve reached the point where it’s hard to believe claims of audible differences between a VCO and an NCO.

Now, sure, there are numerous different filter designs. If you compare different designs, they will sound different. If you carefully model a given analogue filter in software though, there is nothing to be gained from sticking to hardware.

Again, I own quite a few hardware analogue synthesizers myself and I love them. Do they sound good? Yes, they sound fantastic. I’m a collector in this sense, I enjoy the feeling that “this is raw electricity singing”, I like to know “this is the same filter they used on the TRON soundtrack”, and I love the tactile look & feel. There’s nothing dismissive about this.

At some point all this hits a DAC* anyway and ends up being a bunch of numbers. We can generate those same numbers in software.

* ADC to be exact.

#578

There are non-subtle differences with analog synth emulation that I’d venture to say most people can objectively hear. I feel like it is a bit inaccurate to dismiss that fact. To me it’s not on the level of a lossless WAV vs a 320kbps MP3. For example, the overload circuit emulation on the model d app…while sounding great, it doesn’t get as thick and growly as the real thing. That is not to say one is better than the other…for example, digital stuff can have a huge advantage for delicate material that would get lost in the natural noise floor inherent in the analog circuit (if the dsp doesn’t code that noise floor in of course)

I have a hard time with the idea of comparing instruments with qualifiers in general. I don’t think you can boil an instrument down to it’s feature set and say certain things are objectively better than others. And I don’t find what is “better” would hold true across all people.

That being said I completely agree that the kinds of things you can make with digital synths is incredible and it’s cool it continues to move forward every day…take the CS-80, where “presets” were tiny copies of the faceplate underneath a panel, hah!

EDIT: to be clear, by sounding different I mean if you were to a/b them right next to each other with as many variables eliminated as possible…once things are in a mix, definitely would be harder to tell. On that note, one time I emailed someone to ask how they got some guitar tones that I thought were great…and I found out it was all amp sims! So just to reiterate not passing value judgments on if analog is better than digital or vice versa.

1 Like
#579

The categories of “analog” and “digital” can be more context dependent than empirical.

I think one pocket of hardware synthesizer design that doesn’t seem to exist much are digital synthesizers that adhere to the same practical limitations of analog synthesizers.

Like, the Juno 106 is an analog synth that’s digitally controlled and is relatively limited in it’s abilities (i.e. modulation), but the digital controls still feel closer to using an analog synth in it’s immediacy and gestural capabilities. If you were to marry digital synthesis and their timbral possibilities within an interface that is closer to the limitations of vintage subtractive synths, then you might get the same feeling as

“this is raw electricity singing”

even though it wouldn’t be analog.

2 Likes
#580

:100:

I totally agree. It’s how you control the instrument that defines the experience. Natural-feeling tactile controllers with digital hearts: that’s the Golden Grail.

(…and when the heart is already digital, you can just put it in a computer, where it gives you the most flexibility.)

A great example of this is the Korg Wavestate which is based on a Raspberry Pi, you can even connect to it via SSH. It is a computer with a dedicated synthesizer controller which itself is an evolution of the Korg Wavestation.

#581

This is a bit of a pointless discussion since subjective opinions about sound can’t be meaningfully debated, but I do want to point out that you posts are not completely coherent on the subject of digital and analog. What companies put their names on has nothing to do with whether something sounds different.

I am writing my dissertation on analog simulation, and, again, there are objective differences between analog and digital devices. You can see them on a scope, you can hear them on a recording, etc. They will never be the same, because they are not constructed in the same manner.

With all of that said, we have had over half a century in improvements in digital audio, every year this difference is going to decrease. I would say the difference in oscillators only matter in certain circumstances. As analog technology has improved (great stability, purer waveforms, improved tracking), the difference between the two can be truly minimal. The problem is that, of course, the best analog synths are not simply reproducing waveforms, or reproducing the transfer functions of ideal filters.

This brings us back to the work being done with simulation. It doesn’t only matter for digital synths either. Much of the sound of prophet 5 rev 4 comes from the digital control of the analog circuits, which is modeled on the behavior of the old systems.

I listen to my digital Hydrasynth, and find that it sometimes sounds more “analog” than my prophet 12, which is a hybrid digital analog poly, but there are sounds my hydrasynth will never be able to reproduce, simply because of the behavior of the analog filters/distortion. The various filter simulations that are included sound vaguely similar to what they are trying to simulate (I recognize the character of the MS-20 filter, even if it doesn’t subjectively sound the same).

Point being: analog means a variety of things and some of them are more complicated than others.

It is obviously theoretically possible to completely simulate the characteristics of an analog synth, but it currently requires a lot of programming skill, a lot of attention to detail, access to a lot of analog equipment to reference and extensive knowledge of the components involved. You also need quite a bit of computing power to be able to reproduce these circuits. The person I work with has to use supercomputing clusters for days to render a few hours of the output of modeled analog chaotic circuits, to give one example. Things like filters that use feedback are very sensitive to minute rounding errors, etc.

Anyway, if you record an analog sound into your computer, obviously it’s possible to reproduce that since every part of that waveform has been quantized, it’s just that getting from the theory, and abstract modules to something that sounds convincing is not that simple.

There’s a reason my fairly powerful MacBook Pro (6 cores, 2.6ghz, hyper threading, etc.) starts to chug when I run VCV.

Also, unrelated, but there’s a lot of work being done with analog co-processors, and I suspect that computers will begin to integrate more analog processing power for faster differential equations, etc. Ironically we might be able to simulate analog circuits more easily with integrated analog computing.

5 Likes
#582

whatever advances in DSP will not be able to carefully track and replicate the ever-changing deterioration of our space echo tape machine. it is a beautiful, perpetually cranky, idiosyncratic member of our band. this is not a universally desirable quality.

the digital/analog debate implicates a rather boring fixation within music making. it is interesting technically, and i would prefer to see any conversation of this more in the context of a conference with white papers, code examples, sound samples, etc— otherwise it’s just flinging opinions and pairs wonderfully with GAS.

24 Likes
#583

Thanks for your detailed post @Mangrove, I appreciate it. I essentially agree with everything you’re saying.

Agreed. I never said it was easy, there are entire companies specializing in just the thing, and as you’re saying there will always be details either too expensive CPU-wise or too minute to get right.

As I pointed out earlier, my interest is usability in the context of digital mixing and mastering. Layer a few tracks in the mix, add reverb and delay, compress and limit here and there, and the minute differences between using a real Model D vs. a Monark are gone.

I agree that 100% faithful reproduction is a challenge computationally and programatically, I also agree that we’re getting there. As a classically trained pianist I always like to give Pianoteq as an example as I was watching it evolve over the past 13 years.

I agree. Ironically my initial post was an attempt to communicate just that, because from where I’m standing we are already here: digital is not only good, it’s good enough to replace expensive and bulky analog circuitry. So to me the real challenge now is controllers, the interface to the instrument.

I anticipated it might be somewhat controversial but I still underestimated to what extent. Let me then just repeat what I said before:

Cheers and sorry if I pushed anybody’s buttons with my rambling.

1 Like
#584

I 2nd this - My thing is, when I’m recording (or bouncing down), I sometimes like every instance of a wave to look and feel just slightly different. A lot of times that’s meant analog equipment where nothing can ever truly be done twice, sometimes it’s meant round-robins, velocity laters, filtered noise and such until whatever part i’m hyper-fixated on is doing what I need it to do. Sometimes, I want strict perfection.

It’s really all about getting the vibe you want at the end of the day!

1 Like
#585

Makes me wonder if an adventurous person could get HDMI out of it? Because the included screen is ridiculously tiny!

2 Likes
#586

At some point all this hits an ADC anyway and ends up being a bunch of numbers. We can generate those same numbers in software.

I guess you’ve never used tape? Or just listened to it with your ears?

#587

The final listening medium being digital isn’t super relevant to the discussion of analog emulations, in my opinion, in that the ability of digital to capture and faithfully reproduce an analog source is certainly necessary, but not sufficient for analog emulation. For example movies are projected / viewed digitally, which would along the same lines suggest that all of the actors, etc. could just be rendered completely digitally. But of course in practice completely constructing something digitally is more difficult than simple capture and reproduction. Of course emulating an analog synthesizer circuit is easier than a full physical environment, and may well be perceptibly indistinguishable from the real thing at this point.

2 Likes