Mark Mothersbaugh on Classic Hardware vs Software Emulations

[jamcat]

Herbert von Karajan sold over 200 million records of music he didn’t write or even play on.

Sure but did he brag about charting on some obscure chart while offering criticism for those who didn't?

I don’t know. He wasn’t on KVR.

[D-Fusion]

I wonder how much of the perceived fatness difference between hardware and emu here is the filter being actually too clean and not having any slight distortion nonlinearities the original hardware may have?

Based on what i hear on my headphones it is more to do with the digital stiffness that gives the sound a 2d machine gun effect sound in the memorymoog comparison video and the sounds that get close enough are probably sounds that only uses 1 oscillator.

It is the moving oscillators that add some thickness to the sound and the memorymoog would probably sound the same as the plugin if it had the same stiff pitch perfect oscillators.

[Aloysius]

Not a synth but when I listen to Overloud TH-U I sense that I am hearing a very choked output. It seems unnatural to me. A picture has been taken but it seems very forced for want of a better word. It does not replace the real thing imo. It may be state of the art but perhaps the art is in a state.

[zerocrossing]

I wonder how much of the perceived fatness difference between hardware and emu here is the filter being actually too clean and not having any slight distortion nonlinearities the original hardware may have?

Based on what i hear on my headphones it is more to do with the digital stiffness that gives the sound a 2d machine gun effect sound in the memorymoog comparison video and the sounds that get close enough are probably sounds that only uses 1 oscillator.

It is the moving oscillators that add some thickness to the sound and the memorymoog would probably sound the same as the plugin if it had the same stiff pitch perfect oscillators.

But Memorymoog V has a per-voice control on how loose the oscillators, and actually all the main components, are.

[D-Fusion]

I wonder how much of the perceived fatness difference between hardware and emu here is the filter being actually too clean and not having any slight distortion nonlinearities the original hardware may have?

Based on what i hear on my headphones it is more to do with the digital stiffness that gives the sound a 2d machine gun effect sound in the memorymoog comparison video and the sounds that get close enough are probably sounds that only uses 1 oscillator.

It is the moving oscillators that add some thickness to the sound and the memorymoog would probably sound the same as the plugin if it had the same stiff pitch perfect oscillators.

But Memorymoog V has a per-voice control on how loose the oscillators, and actually all the main components, are.

That is true but it doesn't sound like that feature is used in the comparison video on some of the sounds so maybe it would get closer if it was set correctly?

[whassup]

I'll be in the basement inhaling flux core fumes for the next 6 months, because I'M going DIY modular. Fool me once...

Hm, flux core fumes...

[El°HYM]

This is what happened to Tangerine Dream. They went from wonderful hypnotic analog soundscapes to pure cheese. Le Parc and Underwater Sunlight were their last good albums.

Aww, I like Raum.

Oh the new Tangerine Dream with Thorsten Quaeschning is great. Quantum Gate is a good album too. I was thinking of that period between the late 80s and early 2000s.

To be fair, Quaeschning's Rig is still quite Analog. When TD transitioned to fully digital, with all those Workstations and digital effects only. Quite horrible and also sad at the same time, kind of similar to folks today feeling the need the replace everything analog with a plugin, because it looks almost the same. Eno somehow managed it better when he fell in love with his DX7, but kept some of the old stuff around, better transition at least. Last Studio session I have seen him was mainly Logic, yet still some racks to the side.

[zerocrossing]

I wonder how much of the perceived fatness difference between hardware and emu here is the filter being actually too clean and not having any slight distortion nonlinearities the original hardware may have?

Based on what i hear on my headphones it is more to do with the digital stiffness that gives the sound a 2d machine gun effect sound in the memorymoog comparison video and the sounds that get close enough are probably sounds that only uses 1 oscillator.

It is the moving oscillators that add some thickness to the sound and the memorymoog would probably sound the same as the plugin if it had the same stiff pitch perfect oscillators.

But Memorymoog V has a per-voice control on how loose the oscillators, and actually all the main components, are.

That is true but it doesn't sound like that feature is used in the comparison video on some of the sounds so maybe it would get closer if it was set correctly?

Hard for me to say, as I’ve got no access to the hardware. All I can surmise is that one might be able to get closer if they build a patch from scratch and carefully match the settings by ear with visual confirmation. Getting per voice offsets to match would be extremely challenging, so your best bet is to just use your own judgement.

Of course, if you’re happy with your Memorymoog, there’s no reason to pick this up. I doubt the market for this is full of Memorymoog owners. As always, it’s best to judge these things like unique instruments that share their feature sets and general characteristics with a classic synthesizer, but also extend the idea. Instead of asking how much it sounds like the original, as how well it can make sounds that serve your music.

[sellyoursoul]

Not a synth but when I listen to Overloud TH-U I sense that I am hearing a very choked output. It seems unnatural to me. A picture has been taken but it seems very forced for want of a better word. It does not replace the real thing imo. It may be state of the art but perhaps the art is in a state.

Have you tried amp profiles? Profiles are a huge sonic improvement over modelers to my ears. Some people just don't hear it, happy with their modelers, and that's okay, which is the same with synths. To each their own. Personally, I would never use an amp modeler again. That ship has sailed. I hope to see profiling permeate other areas of plugins, such as compressors and non-linearities in synths and other things.

By the way, NAM A2 just dropped, and I think every profile on Tone3000 has been retrained to A2 by now. A2 improves NAM accuracy (which was often already indistinguishable from the real recorded amps used for profiling), reduces latency by quite a bit, improves dynamics, and is significantly more cpu efficient. And now standard and lite versions are included in every profile.

Back on topic...

[TechHaus]

Uncanny valley.

Like talking to a human vs a chatbot.

[D-Fusion]

I wonder how much of the perceived fatness difference between hardware and emu here is the filter being actually too clean and not having any slight distortion nonlinearities the original hardware may have?

Based on what i hear on my headphones it is more to do with the digital stiffness that gives the sound a 2d machine gun effect sound in the memorymoog comparison video and the sounds that get close enough are probably sounds that only uses 1 oscillator.

It is the moving oscillators that add some thickness to the sound and the memorymoog would probably sound the same as the plugin if it had the same stiff pitch perfect oscillators.

But Memorymoog V has a per-voice control on how loose the oscillators, and actually all the main components, are.

That is true but it doesn't sound like that feature is used in the comparison video on some of the sounds so maybe it would get closer if it was set correctly?

Hard for me to say, as I’ve got no access to the hardware. All I can surmise is that one might be able to get closer if they build a patch from scratch and carefully match the settings by ear with visual confirmation. Getting per voice offsets to match would be extremely challenging, so your best bet is to just use your own judgement.

Of course, if you’re happy with your Memorymoog, there’s no reason to pick this up. I doubt the market for this is full of Memorymoog owners. As always, it’s best to judge these things like unique instruments that share their feature sets and general characteristics with a classic synthesizer, but also extend the idea. Instead of asking how much it sounds like the original, as how well it can make sounds that serve your music.

I don't have the Memorymoog to compare it to either so for me it would be more important that I like the sounds i can get out of it

I will wait until it is released in the V-Collection since i have more than enough to play with atm.

[stoopicus]

I don't have the Memorymoog to compare it to either so for me it would be more important that I like the sounds i can get out of it

This is figuratively all that matters. The rest is just winesnobbing.

[BONES]

Now I'm depressed that I live in a world where the poles are melting, fascism is on the rise, and BONES has music that charts. Unless that chart is BONES's top hits.

(D)Anger (album, 2016) - two weeks at no. 1 on the DAC.

Iconoclast (album, 2020) - high point no. 4 on the DAC.

Artifice (album, 2024) - high point no. 2 on the DAC.

BeLIEve (album, 2025) - four weeks at no. 4 on the DAC.

I don't really keep track of the singles but, with the exception of the last one, they've all charted on the DAC Top 10 singles chart, as well as on two different DJ charts in Germany. Once we were only two places behind Taylor Swift on one of them. f**k knows why or how.

We have very high hopes for the new one, which is probably not a very good idea. But we'll be very disappointed if it doesn't spend a few weeks at no. 1 because it is so clearly better in pretty much every way to anything we've done before. People want to see using AI as cheating or taking shortcuts but the reality is that this album has taken longer to get through than the previous one, which did all on our own. (It's 8 months since Tunee spat out the first usable tune and the mastering won't be finished until next week.) Anyway, enough about me...

[BONES]

If you stack the exact same sawtooth wave that doesn't really move at all a thousand times you're not going to hear much difference, in other words, with no variation stacking doesn't do anything, and detuning will just make it muddy.

Detuning creates phasing, which gives the sound movement. It's the very essence of Unison. Without detuning it just makes it louder, which is useless. bx_oberhausen doesn't have a detune control but if you turn on TMT, which simulates analogue instability in the signal path, you get the same effect and it is perfectly judged on that synth to make it sound abso-f**king-lutely huge! Four voices is usually plenty (which is good because it's a CPU hog) but it's nice to know there are 32 available if you need them.

Atomika oscs wide open with as little filter influence on the sound as possible etc. still sound much bigger than the Memory V, or most soft synths. This, is the mythical analog (™Bones) hardware sound that some people complain about, and again we have examples where developers pull it off, it's not impossible.

I think GForce's SEM does a pretty good job here, too, whereas Cherry's SEM definitely needs unison (which it has), as does Arturia's (which it lacks). DForce's OB-E and OB-EZ, of course, have it all - big oscillators and 8 stackable voices. If only OB-E was a little easier to work with or OB-EZ offered a few more controls.

I just get sick of a skeuomorphic send up with mediocre osc and filter emulaitons of the real thing, they screw up the most important parts pretty often.

That doesn't bother me at all, as long as the instrument sounds good and is useful. e.g. Full Bucket's Mono/Fury doesn't really sound anything like a Mono/Poly but it still sounds great in its own right and it's a pleasure to work with. The only thing that stops me using it all the time is that it's not very reliable - I often load up a song and it won't make any sound.

OMG. I hadn't been on here in a while. Re-affirmed: KVR = a whole legion of lifeforms who have dedicated their life to splitting the hairs of vaguely music related arguments down into their constituent atoms; then further, until the original thoughts behind the arguments break down into their quantum components, release small amounts of energy and disappear into entropy itself.

Yeah, it kinda surprised me. I thought we'd moved past this and, honestly, I did not expect the video to stir up a hornet's nest like this. I was expecting more of a consensus on how far we'd come. Shows what I know.

But here's the thing. We are discussing the absolute value of the instruments and their fit for making music. Yes, the Arturia or Cherry emulations of the classic synths may have some minor differences - but they are still outstanding instruments just as fit for making music as the real things were.

Here's the thing, though, what was "outstanding" in 1976 is hardly much chop in 2026. Not really. I find the V Collection (which I have owned for several years now) to be pretty lame compared to what else is out there. The example I used above is a perfect one - their SEM doesn't have unison and it sounds weak next to GForce's version. Even Cherry's SEM has unison to give it a boost. I also compared Airwave to Arturia's revamped JP8000 (or whichever thing it is) and Airwave sounds better to me. I have no idea how faithful either is to the original and I don't care, but Airwave sounds better so it's what I use. If you're new to the VST game, the V Collection is probably a good starter pack - they're not terrible and they cover a LOT of ground - but there are better things out there.

When making an album, professional artists (or their record companies) will ask, "Do we need to hire a producer? An engineer? A sound designer? A session player? A ghost singer? An orchestra? An arranger? A composer? A songwriter? A librettist? Diane Warren?"

When you're creating commercial music, the only thing that really matters is the end product, not your ego.

Well, that's feeding egos in a different way, lavishing all that money and attention to make someone famous. (BTW, I had to look up who Diane Warren is, it sounded vaguely like some American politician's name. I only recognised two of the song names mentioned on her WIkipedia page.)

Are we supposed to be impressed that the AI you use wrote a song that charted on the German Industrial Charts?

1. There is no "German Industrial Chart". Weve charted against bands like Red Hot Chili Peppers and The Psychedelic Furs.
2. If we were using AI to write songs for us in 2016 then, yes, you should be incredibly impressed (and no, you cannot borrow our time machine).

I mean you are obviously so successful at it that you can't even bother or afford to renew your domain

It's 2026, band websites are so 2006 - search for "novakill" and the top result is our BandCamp page. It's interesting that the domain is up for sale because as soon as we ditched it, it reappeared as some weird music site within a matter of weeks.

Sure but did he brag about charting on some obscure chart while offering criticism for those who didn't?

I wasn't bragging, just pointing out facts.

I should also probably have mentioned that I have owned many hardware synths and own one now as well. Maybe I did? In any case I love them too and love hybrid (though mainly ITB) workflows.

I don't. I have kept a couple around, even though I am absolutely sure I will never, ever use them again, but it's out of sheer habit. In fact, I couldn't even tell you where my Rocket is at the moment but my Uno Synth Pro is buried under a pile of shit on my settee, I can just see one corner of it's bag hanging out. I reckon it's probably two years since I have switched it on and it's probably more like 5 years for the Rocket. It took me an awfully long time to get out of that habit but I've not even been tempted to buy any hardware for ages now.

[IvyBirds]

Are we supposed to be impressed that the AI you use wrote a song that charted on the German Industrial Charts?

1. There is no "German Industrial Chart". Weve charted against bands like Red Hot Chili Peppers and The Psychedelic Furs.
2. If we were using AI to write songs for us in 2016 then, yes, you should be incredibly impressed (and no, you cannot borrow our time machine).

Wow so you are taking about things from 10 years ago

In 2016 Red Hot Chili Peppers released an album called "The Getaway" it went #1 in 10 countries and #2 in the USA, UK, and Germany on the Billboard 200 chart. It sold over 500,000 units in America and is certified Gold, and another 500,000+ globally. It's a great album, I own it and went to that tour in 2017 and saw them headline Lollapalooza festival in July 2016.

The fact you are comparing yourself with them is cute. Novakill has never been on the Billboard 200

Now if you want to talk about the DAC or the German Alternative Chart that's a different story. That's not based on album sales or streams. It's a closed-panel chart compiled by publicists determined by points awarded from weekly DJ playlists, not from actual sales. You can reach #1 on that chart without actual selling much of anything. It's pretty meaningless in reality

Again are we supposed to be impressed by that? But hey cool beans that 10 years ago some DJs in Germany played your music in a few clubs for a few weeks making more money than you in the process while you were half a world away in Australia

Shockingly those same DJs and same clubs also play music by the Red Hot Chili Peppers and The Psychedelic Furs. Again cute you want to compare yourself with them

But that was 10 years ago, and it's 2026. The State of the Music Industry and Music Production is vastly different now than it was then which is really the topic of this thread