Software vs. Analog in 2025 – Has the Balance Shifted?

[_leras]

The only thing that we have all learned from this thread so far is; that Software emulations sound almost indistinguishable from their Hardware counterpart, yet the new Behringer never sounds like the Original. Which is something to think about.

that's not what I'm taking from all those examples - to me the hardware has properties that software still doesn't get. it's clear as day from comparing even youtube videos.

that's by no means to say software can't be used to make great music. of course software has been the core of many whole genres for nearly 20 years,

[_leras]

I'm sure we can keep going of examples where software can't match.

I am still waiting for you to post a single example as those sounds are easily done with software

And again I ask what can't software match? Your personal tastes, likes, and dislikes since those can't be measured or quantified is irrelevant, they are just your personal tastes

but you haven't given examples that match any of these - and I doubt you can

that prophet video lets say at 3:00 or 5:45
the crave video at 0:45 where the filter pops open
the neutron video at 5:45
that pro 2 video - from almost anywhere


look, if you don't think it makes much difference, fine... - but if you can't hear it thats on you tbh

[_leras]

I demoed UAD's Anthem yesterday, and it's basically as simple as most hardware analogs, but the sound is fantastic, as is their Model D software. Anthem is based on a mod'd Korg filter topology and it screams like a teenager in a horror movie. I did a full forensic autopsy, and no digital artifacts were detected. I'd put it up against any analog synth sold today.

Having given this a proper demo now - I can only agree Anthem is an absolutely top tier and amazing sounding synth. Could perhaps be the best soft synth sonically.

I'd say it's right up there with the others I think sound best: the other UAD synths, Phosycon 2, The Legend HZ, Massive X.

I'm not sure why it's called Anthem though - no super saw..

[IvyBirds]

I'm sure we can keep going of examples where software can't match.

I am still waiting for you to post a single example as those sounds are easily done with software

And again I ask what can't software match? Your personal tastes, likes, and dislikes since those can't be measured or quantified is irrelevant, they are just your personal tastes

but you haven't given examples that match any of these - and I doubt you can

that prophet video lets say at 3:00 or 5:45
the crave video at 0:45 where the filter pops open
the neutron video at 5:45
that pro 2 video - from almost anywhere


look, if you don't think it makes much difference, fine... - but if you can't hear it thats on you tbh

What exactly am I supposed to be hearing that makes any difference? It's not on me as you are unable to quantify what you mean

Again I asked you to measure and quantify what you mean when you say a synth has "life" and "presence". When I listen to those examples I don't hear any "life" of "presence" most certainly I don't hear anything really all the special

All you are doing is throwing out examples of synths you like, that are playing sounds you like and then declaring that somehow they are superior because of "reasons" but never define what those reasons are or how we would measure it

Beyond that we also don't know how they were recorded, what the gain staging was like, what kind of preamps were used, what effects were used, if they had any EQ applied, what kind of compression, what kind of mastering, etc.

How can anyone make any kind of A/B comparison unless they know that information?

[_leras]

Again I asked you to measure and quantify what you mean when you say a synth has "life" and "presence". When I listen to those examples I don't hear any "life" of "presence" most certainly I don't hear anything really all the special

Beyond that we also don't know how they were recorded, what the gain staging was like, what kind of preamps were used, what effects were used, if they had any EQ applied, what kind of compression, what kind of mastering, etc.

How can anyone make any kind of A/B comparison unless they know that information?

I've given you several exact examples in some of those demos.

If you can't hear the difference between the demos e.g. the repro and prophet 5, then can't help you. Sorry.

[ghettosynth]

Right, I agree with you on most of your points, but I'll differ with you as to whether it can't be done as well, or just hasn't been done as well on the models you listed. Maybe there is something about the topology of those filters that is extra difficult to model. I don't know, but I don't think there is an emulation of either of those filters that is very good either.

Could be, I don't know either. I wanted Urs to make the Cat, I get it, business decisions and all that. But my offer to buy it sight unseen first day of release is still on.

Those are just examples though. The Neutron has a unique filter that I really like. I have two to set them up in stereo. I'd buy that filter in a module. I also really like the Steiner Parker Synthacon filter, the weird 7-pole in the FR-777, I could go on.

So, you might say that it's not the technology, just that they haven't been modeled yet. However, as I was asserting, I think that when we're asking if the balance has shifted in 2025 I think that's an important consideration, this is a practical limitation. You seem to like Softube Modular. I nearly bought it when they came out with the IR3109 filter (Roland). However, in my demo, I realized that they stuck with being true to the hardware and did not allow you to modulate the resonance. This was a deal breaker for me. I get it, you can't do that in an SH-101, but, I can do it in my modular. I can also do it in Blocks using the extracted filter from Super8. An important point here, however, is that they (softube, blocks) cannot be combined in a manner that does not introduce latency. So, asking whether an analog synth can be reasonably replicated is not the same thing as asking if a family of analog synths working together can be replicated.

Also, all of this A/B comparison stuff is just argument rocket fuel. Don't get me wrong, I get it, it's hard to make an objective claim without doing the work. Nobody cares though, nobody is going to do that work who isn't trying to model a device because it is exceedingly difficult and time consuming and, ultimately, unconvincing. In order to make a true objective claim, you need constraints. Those constraints are kindling for objections.

A part of the reason for this is that the limitations always emerge during modulation. Static sounds are not the challenge anymore. So, when we talk about in context usage using similar digital and analog systems, e.g., Reaktor (blocks) vs analog modular, I get more for less (time/irritation) out of analog. It sounds better for less effort. The equation can flip for many kinds of sounds, and certainly it completely flips if our measure is cost per unit quality.

I still prefer analog for analog sounds, especially when modulated. I can only say that it sounds better to my ears in context. I can also say that some of my experiments have been tried, actively, in both domains. I have built hundreds of Reaktor synths (ok, maybe dozens, but many dozens, maybe one to two hundred) both with blocks and with the same techniques used in blocks (that aren't really special, per se), and to me, it's not the same at the edges. I can also say that, objectively, that it fits my production style more naturally for several styles. I do think that it matters that the number of software models of different filters is limited and I also think that it matters that, at the plugin level, you cannot combine plugins with zero latency.

So, are many, if not most, sounds that most people want achievable in a plugin, sure. Has the "balance" flipped in 2025? I don't think so, not yet.

[IvyBirds]

Again I asked you to measure and quantify what you mean when you say a synth has "life" and "presence". When I listen to those examples I don't hear any "life" of "presence" most certainly I don't hear anything really all the special

Beyond that we also don't know how they were recorded, what the gain staging was like, what kind of preamps were used, what effects were used, if they had any EQ applied, what kind of compression, what kind of mastering, etc.

How can anyone make any kind of A/B comparison unless they know that information?

I've given you several exact examples in some of those demos.

If you can't hear the difference between the demos e.g. the repro and prophet 5, then can't help you. Sorry.

Several examples of what? All you did was post a video of a hardware synth, one you didn't even make with no information on how it was recorded and just declared that it was superior, without any explanation as to why you just decided it as so. I have asked you repeatedly to share what you are basing on that can be measured and quantified and you refuse and just double down on your statement that it's better

And it's not just that hardware synth A sounds different than Software Synth B, it's that you claim its better without any context other than you say it is

That's like saying Pepsi is better than Coke because you like it better

[IvyBirds]

The Neutron has a unique filter that I really like. I have two to set them up in stereo. I'd buy that filter in a module. I also really like the Steiner Parker Synthacon filter, the weird 7-pole in the FR-777, I could go on.

The Neutron filter is unique to the Neutron, i was designed by a Behringer engineer on his own and he brought it to Uli, you can't reproduce that with any other synth, so if that is the sound you want you have to use not only hardware but that specific piece of hardware

You could fill your studio up with synths like a CS80, Jupiter 8, or whatever and still not be able to replicate that

I don't think modeling that in software however would be any more difficult than modeling any other analog filter, I just think no one has really tried as the Neutron is not a classic synth, and most companies don't see the value in it.

[ghettosynth]

The Neutron has a unique filter that I really like. I have two to set them up in stereo. I'd buy that filter in a module. I also really like the Steiner Parker Synthacon filter, the weird 7-pole in the FR-777, I could go on.

The Neutron filter is unique to the Neutron, i was designed by a Behringer engineer on his own and he brought it to Uli, you can't reproduce that with any other synth, so if that is the sound you want you have to use not only hardware but that specific piece of hardware

You could fill your studio up with synths like a CS80, Jupiter 8, or whatever and still not be able to replicate that

I don't think modeling that in software however would be any more difficult than modeling any other analog filter, I just think no one has really tried as the Neutron is not a classic synth, and most companies don't see the value in it.

It was an example. It's commonly referred to as the Moffat filter, because it was designed by Keith Moffat. However, it's a common 2164 based filter with some differences that could be determined by tracing the circuit board. A similar core is used in the Blades filter, another great filter. There are also other classic synth filters that aren't modeled in software. So, it's not as unique as you think that it is, but, at the same time, the details, that could be traced in the schematic are unique.

It doesn't matter that it "could" be modeled. It isn't. Someone will model it eventually, when there is desperation for new ground. But you can't get it today. I have had the FR777 since release. It's definitely considered a classic in some circles. There wouldn't be much point in trying to model that because it would push modeling limits and wouldn't drive many sales. This is a marketing limitation, but that limitation is valid.

What's this reminds me of is that period in the 80s when everyone was building CEM based synths. It didn't matter much what you did with it, a CEM 3387 sounds like almost all of the late model CEM filters. I remember ditching pretty much all of my simple CEM based synths at some point in the 90s because they were all very similar.

This is a limitation with modeling. It's either fast with approximations that nobody wants, or, it's slow and accurate, but, it's hard work so there is significantly less variety. You're pointing out a marketing limitation as if that limitation isn't a limitation in a practical sense. Of course it is. If one is asking if software can replace hardware, but in the next sentence states that, well, not all hardware, then the answer is functionally no.

I don't think that this will ever change. Ask me if a piano can be replaced by samples, hell to the f**king yeah! Now go ask a concert pianist. The entire debate is largely a waste of time. I'm not wrong, and neither is the concert pianist, it's all about context. If your context is limited with respect to certain kinds of variety and modulation, sure, plugins will work fine for you. I've tried to help you see why not everyone will agree with you, however, and also why your solution of midi controllers doesn't work for everyone in every context.

[Uncle E]

Beyond that we also don't know how they were recorded, what the gain staging was like, what kind of preamps were used, what effects were used, if they had any EQ applied, what kind of compression, what kind of mastering, etc.

Aren't we all, though? To anyone still running a clean main buss:

[zerocrossing]

To me it's pretty clear the softube doesn't have the same presence,

Repro sounds good but overall the hardware still has more life an character (to me)

What are doing to measure life and presence other than just saying it's things you like so therefore it has it? Because if you can't measure and quantity something it doesn't exist. Synths in hardware and software are inanimate objects they don't have "life"

In the end your entire argument for hardware is that you personally like it, however that is really pointless as every human being likes it dislikes things differently than any other human being

I'm sure we can keep going of examples where software can't match.

I am still waiting for you to post a single example as those sounds are easily done with software

And again I ask what can't software match? Your personal tastes, likes, and dislikes since those can't be measured or quantified is irrelevant, they are just your personal tastes

All I can say is, either _leras has exceptional hearing up to 20 kHz that I don't have, or he's completely imagining this "life" and presence. That's the last I'm going to say about it, because I can't hear what he's talking about, or see it on a scope.

[zerocrossing]

Right, I agree with you on most of your points, but I'll differ with you as to whether it can't be done as well, or just hasn't been done as well on the models you listed. Maybe there is something about the topology of those filters that is extra difficult to model. I don't know, but I don't think there is an emulation of either of those filters that is very good either.

Could be, I don't know either. I wanted Urs to make the Cat, I get it, business decisions and all that. But my offer to buy it sight unseen first day of release is still on.

Those are just examples though. The Neutron has a unique filter that I really like. I have two to set them up in stereo. I'd buy that filter in a module. I also really like the Steiner Parker Synthacon filter, the weird 7-pole in the FR-777, I could go on.

So, you might say that it's not the technology, just that they haven't been modeled yet. However, as I was asserting, I think that when we're asking if the balance has shifted in 2025 I think that's an important consideration, this is a practical limitation. You seem to like Softube Modular. I nearly bought it when they came out with the IR3109 filter (Roland). However, in my demo, I realized that they stuck with being true to the hardware and did not allow you to modulate the resonance. This was a deal breaker for me. I get it, you can't do that in an SH-101, but, I can do it in my modular. I can also do it in Blocks using the extracted filter from Super8. An important point here, however, is that they (softube, blocks) cannot be combined in a manner that does not introduce latency. So, asking whether an analog synth can be reasonably replicated is not the same thing as asking if a family of analog synths working together can be replicated.

Also, all of this A/B comparison stuff is just argument rocket fuel. Don't get me wrong, I get it, it's hard to make an objective claim without doing the work. Nobody cares though, nobody is going to do that work who isn't trying to model a device because it is exceedingly difficult and time consuming and, ultimately, unconvincing. In order to make a true objective claim, you need constraints. Those constraints are kindling for objections.

A part of the reason for this is that the limitations always emerge during modulation. Static sounds are not the challenge anymore. So, when we talk about in context usage using similar digital and analog systems, e.g., Reaktor (blocks) vs analog modular, I get more for less (time/irritation) out of analog. It sounds better for less effort. The equation can flip for many kinds of sounds, and certainly it completely flips if our measure is cost per unit quality.

I still prefer analog for analog sounds, especially when modulated. I can only say that it sounds better to my ears in context. I can also say that some of my experiments have been tried, actively, in both domains. I have built hundreds of Reaktor synths (ok, maybe dozens, but many dozens, maybe one to two hundred) both with blocks and with the same techniques used in blocks (that aren't really special, per se), and to me, it's not the same at the edges. I can also say that, objectively, that it fits my production style more naturally for several styles. I do think that it matters that the number of software models of different filters is limited and I also think that it matters that, at the plugin level, you cannot combine plugins with zero latency.

So, are many, if not most, sounds that most people want achievable in a plugin, sure. Has the "balance" flipped in 2025? I don't think so, not yet.

Well, I'm going to have to end this here, because I think the Neutron is one of the crappiest sounding synths to come out in the last decade. If that's what you think is great, have at it.

[ghettosynth]

Well, I'm going to have to end this here, because I think the Neutron is one of the crappiest sounding synths to come out in the last decade. If that's what you think is great, have at it.

LOL, well, there's no accounting for taste. I said that the filter is great, but, subjective opinion about sound quality is of no moment with respect to the conversation about modeling.

However, if we're calling into question each other's taste, I can hardly see how you would think a (semi) modular synth sounds "crappy?" How any modular synth sounds is a function of its individual components and how you connect them. The filter is a highly resonant 2164 based dual state variable and the oscillators are CEM3340s. When I picked mine up they were about $200 and I used the pair of them to reboot my analog modular interests.

[IvyBirds]

The Neutron has a unique filter that I really like. I have two to set them up in stereo. I'd buy that filter in a module. I also really like the Steiner Parker Synthacon filter, the weird 7-pole in the FR-777, I could go on.

The Neutron filter is unique to the Neutron, i was designed by a Behringer engineer on his own and he brought it to Uli, you can't reproduce that with any other synth, so if that is the sound you want you have to use not only hardware but that specific piece of hardware

You could fill your studio up with synths like a CS80, Jupiter 8, or whatever and still not be able to replicate that

I don't think modeling that in software however would be any more difficult than modeling any other analog filter, I just think no one has really tried as the Neutron is not a classic synth, and most companies don't see the value in it.

It was an example. It's commonly referred to as the Moffat filter, because it was designed by Keith Moffat. However, it's a common 2164 based filter with some differences.....So, it's not as unique as you think that it is,

In other words it's unique to the Neutron as there is no other synth that has what you call those same "differences'" which is what makes it unique

And please tell me how unique do I think it is? Because I never communicated that.

[_leras]

Several examples of what? All you did was post a video of a hardware synth, one you didn't even make with no information on how it was recorded and just declared that it was superior, without any explanation as to why you just decided it as so.

That's like saying Pepsi is better than Coke because you like it better

Several examples of great synth demos where software can't match the realness/presence/life of the hardware version. And it's not really that close.

Simple sounds may come close, but I honestly struggle to find demos of vets to the level of demos that people do of hardware synths. Vsts usually just gets crappy run through the presets. Read into that what you may.

People have said, yeah can do that with this synth, but not many examples posted tbh.

If you'd like to pick one sound to try and recreate, perhaps the sound right at the start of the Leipzig video? It's simple enough sound with the filter being opened a bit, and though a couple of synths may get close I'd be amazed if any software can get as much meatiness/richness and realism into a sound.

I doubt you'll try though...