Acustica Audio debuts THING - Iconic '80s Synth Reimagined for a New Era

[plexuss]

I would be surprised if the basic oscillator waveforms weren't sampled, and the filters and EQ weren't convolution-based. They sound too good and close to the real thing for coming out of a digital algorithm.

PWM can't be done with samples. It has PWM.

Wavetables

[Uncle E]

Good point. How about cross mod?

[db3]

I'm torn on this - it's only worth it for the Black Friday price, and it does sound good but I wound up running into crippling CPU overloads on M1 Max and M1 Ultra Macs. It works fine in general (CPU usage etc.) but there are regular and unavoidable pops that would be impossible to work around and I don't know whether to expect any improvement on that front.

Most of the preset switching CPU spikes are due to preset specific quality settings. From the dev:

"Between today and tomorrow, we’ll address the menu issues and improve preset switching. To mitigate clicks, we’ll introduce an anti-click protection system, and by the end of the week, we might even have a smoother transition system in place."

[briefcasemanx]

If you don't want to return to the thread, stop typing in the thread. This is weird and neurotic--trying to get other people to comply when it's you that has the issue, and the solution is completely controllable by you. This is very weird, very controlling behavior. NON ONE is forcing you to type replies

No one is forcing you to constantly reply to my posts in order to insult my intelligence. I'm not going to just let you take shots at me without responding. So please stop quoting me, otherwise it is you who is "weird and neurotic" and quite frankly being a troll.

It still sounds like you don't actually know what convolution is.The rep here telling you the synth uses "samples" might not mean what you think it means.

It still sounds like you are missing the point. The question was does the synth engine use samples or is it 100% generative as are the vast majority of other analog emulations? It's really a very simple question and I can't make it any simpler for you to understand.

I'm not the only one who asked the question and as for the rep's response:

is this sampled or dsp

So does it use samples or not?

Yes it does.

Granted the rep is being obtuse by not clearly explaining how it uses samples which seems odd to be ashamed of their fancy Voltera system. Seems like they'd be proud to explain the details especially given their propensity for using buzzwords and acronyms to describe their products.

Now I'll ask you again politely to please stop quoting me. Please find someone else to argue with.

I'm having a conversation in a thread with people who replied to the thread. I'm not complaining and whining about people quoting me or responding to me. You are the one complaining and whining about that, and then being weirdly controlling and childish and going "I'll keep responding to you but you have to stop responding to me because I say so".

This is personality disorder-level behavior. If YOU don't want the conversation to continue, you have the power to stop it. The locus of control is not on me to stop you from typing. Unbelievable.

The person you quoted asked if it was "DSP or samples" which really means is it algorithmic or is it "sampling" hardware, which can mean a lot of things including impulse responses. Again, it seems like you don't know what convolution is.

[zerocrossing]

You didn't answer the question. Is some of the synth sample based or "convolution" based ?

No but it certainly was a great session of buzzword bingo I'm surprised they didn't use Heisenberg compensators too.

IMG_0298.jpeg

[zerocrossing]

Ok I say the following as someone who is a huge fan of sampled based synthesis. It is without a doubt capable of a wider range of sounds than any other method.

But sampled or not I've tried the demo and have already uninstalled it since I found the overall sound unpleasant and the CPU demand way too much despite claims to the contrary.

The fact that it's limited to 8 note poly was the nail in the coffin but I suppose any higher poly would melt down most top end systems.

Good luck with the project but it's just not for me.

I’m listening to a preset demo right now and frankly I am unimpressed. Sounds like someone tossed a thick blanket over a monitor.

[Teksonik]

I'm having a conversation in a thread with people who replied to the thread.

No, you're being a troll. Your "conversation" started with being insulting and was utterly and completely pointless:

This question doesn't even make sense. It specifically says in your quoted text that it's convolution-based. The question itself makes it seem as if you don't know what convolution is.

Even after it has been explained to you time and time again you still don't get how the question did indeed make sense. I understand perfectly what convolution is but you still fail to understand the difference between generative or "DSP" synthesis as is used in the vast majority of the other analog emulations on the market and sample based synthesis whether you want to call it convolution or whatever. If you can't understand the subject then stay out of the conversation or you just end up looking stupid.

This is personality disorder-level behavior. If YOU don't want the conversation to continue, you have the power to stop it. The locus of control is not on me to stop you from typing. Unbelievable.

No, the personality disorder is on you after being asked politely to stop quoting me you still had to persist. That is pathological behavior. Please seek help.

Ok so let's see if you have the restraint that you expect me to have when being insulted. You are a reprehensible reprobate and a gutter crawling troll.

Now once you get done looking up what those words mean try to restrain yourself from responding. I can feel your fingers trembling, unable to resist pounding away at the keyboard to post another steaming load of drivel.

So can you do it? Can you resist responding? I'm betting the house that you can't.......

[davidka]

I just tried it briefly. Despite that I've some experience with J8 based soft-synths, and a few analogs, there were sounds I never heard in any of them. It feels different in a good way. Really enjoyed it, and I encourage everyone to try it. Devs also mentioned, they will patch up the majority of known bugs.
Not in time for me to buy, but I'll keep my eye on where this technology will go. I am curious, how long will be the R&D life cycle for other products. Wish all the best for the developers, I think they've really done something new here.

[briefcasemanx]

I'm having a conversation in a thread with people who replied to the thread.

No, you're being a troll. Your "conversation" started with being insulting and was utterly and completely pointless:

This question doesn't even make sense. It specifically says in your quoted text that it's convolution-based. The question itself makes it seem as if you don't know what convolution is.

Even after it has been explained to you time and time again you still don't get how the question did indeed make sense. I understand perfectly what convolution is but you still fail to understand the difference between generative or "DSP" synthesis as is used in the vast majority of the other analog emulations on the market and sample based synthesis whether you want to call it convolution or whatever. If you can't understand the subject then stay out of the conversation or you just end up looking stupid.

This is personality disorder-level behavior. If YOU don't want the conversation to continue, you have the power to stop it. The locus of control is not on me to stop you from typing. Unbelievable.

No, the personality disorder is on you after being asked politely to stop quoting me you still had to persist. That is pathological behavior. Please seek help.

Ok so let's see if you have the restraint that you expect me to have when being insulted. You are a reprehensible reprobate and a gutter crawling troll.

Now once you get done looking up what those words mean try to restrain yourself from responding. I can feel your fingers trembling, unable to resist pounding away at the keyboard to post another steaming load of drivel.

So can you do it? Can you resist responding? I'm betting the house that you can't.......

Why would I resist responding? That makes no sense. You're the one with the issue. Once again, you're strangely offloading your own issues onto others. YOU are the one that is having issues here. YOU are the one that is pretending yo not want to comment on the thread. So stop. I'm perfectly fine continuing to comment in this thread.

The question didn't make sense and despite you now claiming that it was explained "time and time again" no actual logical argument was made. The phrase "Time and time again" also makes no sense, how many replies have I gotten? It says in the quoted text when you asked the original question that its convolution-based. You then asked if it was "some" of the synth was sample-based or convolution based, which makes no sense. There are multiple ways to try to re-interpret your terribly phrased question, snd none of them make logical sense to anyone that realizes that this type of convolution is sampling. So it makes it seem like you don't know what convolution is and nothing you've written, despite all the weird and neurotic projecting, has made it appear that you understand. Maybe you do understand it, maybe you dont, but either way the question is nonsense. Write with more clarity when asking other people questions.

It's funny that you claim I don't know the difference between a "generative" (not the word typically used, but I'll take it, it's not worse than "algorithmic" ) and a convolution algorithm when I've created DSP algorithms for both .

But if you think I'm somehow obliged to stop responding to you because you say so, because somehow I'm forcing you to keep posting and whatever other childish, transparent nonsense that you are getting mad at because I'm calling it out--follow your own logic:

Please stop responding to me. I'm asking you to stop. Now that I'm asking you to stop you have to do what I say, isn't that right??

[burkek]

TH8 is ~20% CPU in Ableton 12 on a M1 Max MBP for anyone curious - not the lightest weight synth I've got installed but not crippling either.

A lot of people don't know how to read CPU percent in Activity Monitor (if that's what you're using). CPU% shown in Activity Monitor is total across all cores. So for the sake of argument, if you have 10 cores and Activity Monitor shows 20%, that equates to 2% per core. A 10 core system could theoretically show a CPU% of 1000 (10 x 100%). So when you see 20% in Activity Monitor, it's 20 of 1000 (in this example). In other words, nearly nothing.

KEv

[Zaphod (giancarlo)]

Hi everyone, and thank you for your interest. I’ll try to address some of the most frequently asked questions and share a few observations for those curious about the technology behind the synth and looking for precise answers.

First, it’s important to note that this type of synthesizer doesn’t perform well with wavetables and is certainly not ideal for sampling. This is because, beyond the oscillator sync—which is not just about PWM—the real challenge lies in the cross modulation. While wavetables can technically achieve these operations, replicating the level of aliasing that this synthesizer achieves would be an extremely difficult task. That said, we are conducting extensive research in this area. To stay focused on the topic, a typical wavetable-based synthesizer will never be able to achieve this kind of synthesis.

Regarding the technology we use, it’s worth mentioning that we typically incorporate Volterra Kernels, which are also present here. These are primarily utilized in the MEGA mode. However, keep in mind that convolution operations in this synthesizer are minimal. The decision isn’t about whether to use convolution but rather where it makes sense to apply it in the context of the synthesizer. In simpler terms, you can have Volterra Kernels without requiring direct convolution operations. For those curious, we’ve managed to include convolution not only in the effects section but within the synthesizer itself. For example, one of the oversampling modes uses linear phase filters, which is quite a novelty in this field.

What makes this synth particularly faithful to the original hardware is our approach to sampling. For instance, when sampling slider values, where some competitors rely on a handful of points or a single equation, we use multi-sample techniques to capture every point of the range and reproduce them as accurately as possible. Additionally, we assembled a large team of experts to help us match some of the more challenging aspects, like cross modulation.

You’ll find more detailed explanations in some of my Italian interviews, online resources, or even in the manual. Feel free to reach out with specific questions, and I’ll do my best to answer them.

Regarding CPU consumption, the latest version has already reduced clicks and CPU usage, even when switching presets. During normal operation (not changing presets), there should be no clicks, noise, or similar issues. We’re actively working toward eliminating potential clicks in extraordinary scenarios, such as switching quality modes. Furthermore, we’re refining the ECO quality mode to ensure it offers a CPU consumption comparable to competitors. Naturally, sacrificing some quality results in lower CPU usage, and vice versa. What sets this synth apart is that, if you want higher quality, we can provide it—unlike some competitors who don’t offer such settings.

As for the presets, they were created by industry experts. Based on the videos I’ve seen, if you take the original hardware and set the sliders to the same positions, you’ll achieve the same sound. However, when evaluating presets on platforms like YouTube, it’s easy to assume that the machine isn’t flexible or doesn’t produce outstanding sounds. I can assure you that increasing the quality mode delivers incredible sound. That’s my perspective, but I encourage you to try it yourself.

What truly modernizes this synth are features like the super stereo mode, which creates stereo sounds that the original hardware could never produce, as well as the sine wave sub oscillator synchronized with the main oscillators. This feature brings the synth into the modern era, making it perfectly capable of generating trap and EDM sounds—where the original hardware would have been limited to genres like synthwave.

Give it a try, and let us know what you think!

[trusampler]

I'm demoing TH8T and it's tuning is completely off, and it won't layer with any other synth. You guys might want to check out your tunings. They're completely wrong in the demo.

[miloszz]

A lot of people don't know how to read CPU percent in Activity Monitor (if that's what you're using). CPU% shown in Activity Monitor is total across all cores. So for the sake of argument, if you have 10 cores and Activity Monitor shows 20%, that equates to 2% per core. A 10 core system could theoretically show a CPU% of 1000 (10 x 100%). So when you see 20% in Activity Monitor, it's 20 of 1000 (in this example). In other words, nearly nothing.

KEv

I've never seen the point of looking at activity monitor, what matters is DAW performance. If a plugin spikes past 100% in Ableton or Logic's CPU monitor you get pops. Thing runs at 15-25% in Logic and Ableton's CPU monitor (ie ~20% of one core) on both my M1 Max MBP and M1 Ultra Studio, but it also spikes (on my machines) at up to 3000% and causes pops regardless.

[JanuszPelc]

Furthermore, we’re refining the ECO quality mode to ensure it offers a CPU consumption comparable to competitors. [...] Naturally, sacrificing some quality results in lower CPU usage, and vice versa.

Lower CPU requirements in ECO mode are a godsend for composing and arranging, and the quality tradeoff is completely acceptable in that context. However, most real-life projects involve numerous tracks with multiple synth instances, making freezing, rendering, or bouncing tracks "in place" a necessity.

But the current UX approach is a real workflow killer in these situations. Manually switching to high-quality mode before freezing each track and then reverting to ECO mode after unfreezing to make adjustments is tedious and disrupts the creative process.

This is why many CPU-intensive competitor products include a setting like "Always render in best quality."

Based on the videos I’ve seen, if you take the original hardware and set the sliders to the same positions, you’ll achieve the same sound.

Watching YouTube videos is fun, and tweaking such a wonderfully sounding synth is pure joy. Many of us could happily do it all day.

However, for some of us, using synths in real projects is essential to justify their cost. This demands top-tier sound quality, not compromises made for eco-friendly settings.

So, Giancarlo, please consider implementing an "Always render in best quality" setting.

[Zaphod (giancarlo)]

Let me try to provide a cumulative response to the recent observations.

First, regarding the creation of a parameter for high-quality rendering during audio export: this is technically possible, but we’ve given it a lower priority for a simple reason—the sound changes. From our perspective, building and playing the sound in a specific way and then automatically rendering it at a higher quality does not yield the same result.

To explain this further and give an example of our reasoning, in ECO mode (low-resource mode), we don’t account for the fact that the oscillators are always active and continue running, and oversampling is kept at a lower level. In MEGA quality mode, however, we not only apply a different, more accurate oversampling but also introduce harmonic distortion and similar elements. As a result, the synthesizer behaves differently—it’s not just a matter of higher quality like sharpening a pixelated photo. The content of the “photo” itself changes. This is why we’ve deprioritized this feature for now.

The second question concerns the plugin’s tuning. Initially, the plugin did have tuning issues, but these were resolved based on user feedback. However, note that some presets are intentionally in a different pitch, particularly those utilizing cross modulation. By its nature, cross modulation alters the tonal structure of a preset, leading to differences in pitch.

In the future, we might consider adding a transpose function to the plugin, but there’s a potential drawback: users might rely on it for transposing, which could conflict with the default settings influenced by cross modulation. It’s a bit of a “chicken-and-egg” situation.

I hope this clarifies things!