Tags: posts polarity-music Bitwig Physical-Modeling Poly-Grid Sound-Design Replace VST

Creating a DIY Physical Modeling Synth in Reason Grid

Tutorial | May 11, 2023

In this video, I explain how to make a physical modeling synth in the Reason software. Physical modeling is a way of replicating real-world objects in a synth by modeling their resonances and frequencies. To start building the physical modeling synth, we need to introduce an exciter and use multiple bandpass filters. The resonance frequency of objects is specific to their material, and the frequencies of the overtone change with time. We need to introduce envelopes for each overtone that changes over time. We can use multiple partials and tune the frequencies of the bandpass filters to replicate specific objects' resonances. Loudness is also crucial, and we can use multiple envelopes for each harmonic. We can use a comp filter or use a ratio knob to target specific harmonics. Alternatively, we can introduce all harmonics and remove the ones we don't want. It's an intricate process, but the result is worth the effort.

You can watch the Video on Youtube - support me on Patreon

Questions & Answers

Maybe you dont watch the video, here are some important takeaways:

Questions and Answers

1. How can one start making an off-brand physical modeling synth using recent studio objects in the grid?

To create an off-brand physical modeling synth using recent studio objects in the grid, one can start by understanding the concept of physical modeling. This involves replicating the fundamental frequency and overtones of a sound and adding movement over time through envelopes. The process can be simplified with a combination of steep bandpass filters and an exciter function, which simulates the effect of hitting or plucking an object. Additionally, the resonating frequency of an object needs to be taken into account and modeled through resonator banks, which are essentially bandpass filters. By dialing in specific frequencies, one can replicate the resonating frequency of different materials or objects.

2. How does the object synthesizer work and what are its components?

The object synthesizer works by combining an exciter function, which simulates the action of hitting or plucking an object, and a resonator function, which replicates the resonating frequency of a physical object. The resonator function is usually modeled through bandpass filters, and the exciter function can be anything from noise to a modeled mallet. Object synthesizer also features different resonance modes, such as string and beam, which simulate different types of objects. Additionally, there is an object delay function, which provides a delay effect that can be modulated by the input signal.

3. What are the key components of physical modeling and how are they used in synthesizers?

The key components of physical modeling are the fundamental frequency and overtones of a sound, the resonating frequency of an object, and movement over time through envelopes. These components are used to replicate the sound of real-world objects, such as musical instruments, in synthesizers. In synthesizers, bandpass filters are used to model the resonating frequency of objects, while an exciter function


This is what im talking about in this video. The text is transcribed by AI, so it might not be perfect. If you find any mistakes, please let me know.
You can also click on the timestamps to jump to the right part of the video, which should be helpful.

[00:00.000] Hey folks, welcome back to another video.
[00:03.280] Someone on Reddit asked, I'm a huge fan of physical modeling, but also poor right now.
[00:09.000] I mean, I can relate to that.
[00:10.920] If hypothetically, I'd like to make an off brand physical modeling sit based on recent
[00:14.840] studios object in the grid, any tips on where to start.
[00:18.520] So I just wrote here, yeah, it's blah blah blah, it's just steep band pass filters, targeting
[00:23.920] the right frequencies and so on.
[00:25.840] So that's basically very, very short answer to that.
[00:29.640] But it's kind of true.
[00:31.680] And I've watched here the video of reason introducing the synthesizer called object.
[00:37.920] And they also took basically the same approach as every other physical modeling sit where
[00:43.400] you need to create brought, brought results or brought results of different instruments.
[00:51.960] So it's not like that you only want to have a flute or saxophone, right?
[00:58.000] When you do something like this, you probably want to really model the real world inside
[01:06.360] of your synthesizer and it's not really possible because it's a lot of math involved.
[01:13.080] But using here basically a bunch of band pass filters, so we have your three tabs of band
[01:18.080] pass filters where you can dial in offsets to the frequencies.
[01:25.080] And we have also an exciting part, which is the exciter.
[01:29.120] And I explained to you this in a minute how this kind of works.
[01:32.600] It's not really hard to get.
[01:35.720] So what you need to know basically is that every instrument, every sound has a fundamental
[01:41.200] frequency and has overtones, every sound.
[01:45.400] And you can replicate every sound by just replicating these overtones and the fundamental
[01:50.720] frequency.
[01:51.880] And of course, the movement over time.
[01:54.120] So when you play, for instance, a guitar, right, not every overtone is at the same load
[01:59.760] ness at the same point in time.
[02:01.800] So you need to introduce envelopes to each of these overtones.
[02:07.360] So they change over time and maybe also in frequency over time.
[02:11.360] So this is the complicated part, right?
[02:14.280] So here we have an exciting part.
[02:16.360] This is basically when you pluck the string on another guitar or when you hit the desk with
[02:20.280] your hand or when you hit the drum plate with the drumstick, right?
[02:26.520] The drumstick itself is modeled here.
[02:29.200] And this is the resonance of the material you are hitting or the object you are hitting.
[02:36.800] This is the frequency response.
[02:38.920] And every material, every object reacts with different frequencies.
[02:44.840] So you can imagine if you hit your desk with your hand, you get a different sound, different
[02:49.560] harmonics back, then hitting a metal plate with your hand, right?
[02:55.360] It sounds completely different because the resonance of what you get back is different.
[03:01.720] It's also the same with light.
[03:05.920] The color on your wall is basically not on the wall itself.
[03:11.320] It's the frequencies you are reflecting.
[03:15.280] So when you put light on a wall, you get all the frequencies or color frequencies back
[03:23.640] from the wall because the wall is exactly that material and that reflects only different
[03:29.680] frequencies.
[03:31.200] I don't want to get into that.
[03:32.960] But it's basically the same thing.
[03:35.200] It's all frequencies.
[03:36.200] It's just another different spectrum.
[03:39.040] So objects have different resonating frequencies.
[03:42.120] You get back different harmonics from an exciting moment.
[03:48.480] Okay.
[03:49.960] So this object synthesizer works exactly like, for instance, a chroma phone.
[03:59.360] Chroma phone three here.
[04:01.000] Let's open this one.
[04:04.960] So just remember this, we have an exciter and we have a resonator simulating the object.
[04:10.680] Here we have the same thing in the editor editor.
[04:13.080] We have on the synth part, you have some different modes here, it's just modulation.
[04:17.200] The synth part is the main part.
[04:19.000] We have here a noise part, which is the exciter or mallet.
[04:22.680] You can mix this together so you have a different exciter sound or make the exciter sound
[04:28.040] modeled to a specific sound.
[04:31.280] And then we put this noise, this exciter noise through resonator banks.
[04:35.960] We have here two resonator banks on the synth tab one.
[04:40.880] And we can change here the resonating part.
[04:44.400] We can also switch this to manual if you want to, maybe switch to the second part of.
[04:50.800] And then...
[04:53.800] Right.
[05:04.040] You can model basically with this resonator some kind of object behavior, maybe I'll put this down.
[05:17.040] And then we have these other models here, string, for instance, or beam.
[05:25.040] But really these models are nothing else than just this manual tab here, basically.
[05:29.840] I think it's just a bunch of band pass filters.
[05:34.840] But they are pretty tuned, so you don't have to manually dial in, dial in, but the manual
[05:42.840] part is basically that gives away how this works.
[05:45.680] We have multiple partials, and these partials are basically the frequencies of these band
[05:50.440] pass filters.
[05:51.440] So we have four band pass filters here, maybe.
[05:58.440] So we have exciter, and then we have here resonators.
[06:01.440] It's the same concept as here, exciter, band pass filters.
[06:07.440] So here we have three, and chromophone we have one, two, and also four here.
[06:13.440] And so we have two synths with each two resonators that you can bring in.
[06:20.440] And then you can combine them, mix them together, and it gives you a much, much...
[06:26.440] richer result.
[06:27.440] But the concept is the same, basically.
[06:31.440] Then we have object delay.
[06:35.440] Object delay is from the same company, AAS.
[06:42.440] And in here, it's more aimed at having some kind of delay.
[06:49.440] So you have to put something through that.
[06:51.440] It's an effect.
[06:52.440] So we have your polymer.
[06:59.440] But when you basically turn off LFO, pre-filter, and the delay part, I think this gives us
[07:12.440] here the whole wet signal.
[07:14.440] So now we can define basically with the synthesizer itself kind of an exciter noise.
[07:21.440] In this case, you would just take noise.
[07:27.440] So this is basically your hand slapping the desk.
[07:34.440] And when we put this here through the resonator bank,
[07:41.440] get the drum head sound.
[07:54.440] But the problem here is we can't change the frequency dependent on the key we are pressing
[07:58.440] on the keyboard.
[07:59.440] So it's a fixed frequency, basically.
[08:01.440] But it's the same concept also here.
[08:04.440] We can change to a plate with a metal plate.
[08:17.440] So what you do basically with material, it sounds like you change the resonance of upper frequencies
[08:24.440] of the upper panpass filters.
[08:27.440] Also here the K just increases probably the resonance.
[08:33.440] And this changes the frequencies of all panpass filters.
[08:38.440] And this is the form.
[08:39.440] And this is interesting here.
[08:41.440] Formants are usually fixed frequencies.
[08:47.440] I'll explain you this here also, of course, quickly in this one here.
[08:55.440] So what you do or what you want to model with physical modeling is,
[08:59.440] first one, the drumstick itself, so how it sounds when you hit something.
[09:04.440] And only that without the resonances of the object.
[09:07.440] So you hit something in our case here, maybe a metal plate.
[09:11.440] And the metal plate resonates then.
[09:14.440] And the resonances of this metal plate are basically exactly the right resonances of this material.
[09:21.440] And maybe also the physical size, the physical size of this object,
[09:29.440] so it gives you back different resonances.
[09:32.440] And this is basically the panpass part where you dial in the specific frequencies for specific objects.
[09:41.440] And then when you play, for instance, if this is your hand and this is maybe a string on a guitar,
[09:48.440] then you also have to put this through an object, another object,
[09:52.440] because the guitar, for instance, has a guitar body, right?
[09:57.440] You have this nice, formed, yeah, that's a guitar.
[10:03.440] You get the idea, right?
[10:05.440] So this body itself, also the size of this and what kind of material you use,
[10:10.440] what kind of wood you use, decides on how the sound is shaped,
[10:15.440] or how it sounds, basically, how the sound sounds.
[10:19.440] You know what I mean?
[10:21.440] And these frequencies from this body usually don't change.
[10:25.440] They are fixed frequencies.
[10:27.440] It's called formance, right?
[10:29.440] So we have formance here, and then we have resonances here,
[10:32.440] and then we have here maybe noise, which is the exciting part.
[10:39.440] So and this one is nicely, most of the time,
[10:44.440] nicely modeled with convolution reverbs,
[10:47.440] where you have like a short impulse response from an object.
[10:50.440] And it's basically also just filtering the sound in a way.
[10:54.440] You have filter out different overtones,
[10:57.440] but these frequencies never change when you change the pitch.
[11:01.440] So when you play a different melody on the guitar or when you play a different melody
[11:06.440] on the piano itself, the resonances of the piano body never change.
[11:12.440] They're always the same.
[11:14.440] They are fixed frequencies, so you can dial in just also with the EQ
[11:18.440] these fixed frequencies to simulate some kind of body.
[11:22.440] And I want to give you this here in the grid, an idea about this,
[11:27.440] so that we have multiple approaches to that.
[11:30.440] So we can just start very simple.
[11:34.440] So we introduce a noise, or the exciting part here with an AD,
[11:41.440] pink noise, and then go to output here.
[11:48.440] So now we arm that.
[11:51.440] So we get the click sound.
[11:54.440] So this is basically our exciter part,
[11:57.440] and you can put a lot of work into that.
[12:01.440] You can shape it, you can put some filters on there just to shape the exciting part,
[12:06.440] exactly how you want it.
[12:08.440] And we also use here, this way we label this,
[12:11.440] so you have an easy way to understand this.
[12:14.440] So this is the exciter.
[12:18.440] And then we send this through a body.
[12:21.440] And the body is usually simulated here with a filter,
[12:27.440] with a bandpass filter, so we go into that.
[12:30.440] And then we can use here the resonator or the resonance.
[12:37.440] So it already sounds like an object a little bit.
[12:47.440] So that's this part.
[12:50.440] And we also want to introduce here probably a macro
[12:56.440] to change the resonance here.
[12:59.440] But it's likely over the edge.
[13:06.440] Okay, fine.
[13:13.440] And in the end here, we probably want to have some kind of distortion
[13:16.440] because when you record and physical object with your microphone,
[13:20.440] you have some kind of distortion.
[13:22.440] And when you hit the desk, right, pretty hard.
[13:25.440] And then you have a microphone right beside it.
[13:28.440] Besides that, you probably have some distortion on there
[13:30.440] because the microphone doesn't get a whole dynamic range.
[13:35.440] So we simulate this here.
[13:37.440] And it's also nice for you to hear it better.
[13:42.440] Okay.
[13:43.440] So then we need, of course, multiple of these.
[13:46.440] And we have to summarize this.
[13:49.440] So let's go with, I think, let's see, they had here.
[13:53.440] One, two, three, four, five, six, seven, eight.
[13:56.440] Okay, eight, one, so three, four, five, six, seven, eight.
[14:05.440] Right.
[14:06.440] So we have this kind of battery of one, two, three, four, five, six, seven, eight.
[14:13.440] Okay.
[14:15.440] So we sum these all together here with the sum.
[14:25.440] Hopefully I can get eight in there.
[14:27.440] Yeah.
[14:30.440] And then let this into the output.
[14:32.440] So now all these band pass filters basically have the same resonance.
[14:37.440] We can change here.
[14:40.440] And all these band pass filters have the same frequency.
[14:43.440] So what you can do now is you can dial in specific frequencies for an object.
[14:51.440] So let's see how this sounds.
[14:56.440] Maybe I'll leave this open.
[14:58.440] Maybe a scale is a bit down.
[15:00.440] Something like this so I can increase here this.
[15:03.440] Nice.
[15:05.440] So now we can dial in specific frequencies.
[15:08.440] I just dial in here some random stuff.
[15:10.440] Let's see how this sounds.
[15:23.440] Right.
[15:26.440] If you put too much resonance on that, it's exciting too much.
[15:32.440] You can also put this here.
[15:34.440] Maybe, nah, I'll leave it at that.
[15:36.440] I'll put this here into maybe a logarithmic.
[15:42.440] So we have all exponentials, maybe better.
[15:51.440] Yeah, it's better to have exponential here.
[15:56.440] So we have more fine-crain control at the end of the macro.
[16:07.440] Look, a logarithmic puppy probably better.
[16:10.440] Okay, so now it's already sounds like it's some kind of random object.
[16:15.440] And to replicate specific objects, you have to analyze these objects
[16:20.440] and have to see what kind of frequencies they're resonating with.
[16:24.440] Right, so also the problem here now is that we can't change the pitch
[16:37.440] because we just disabled here the pre-cords.
[16:44.440] So we can change the frequencies depending on the input pitch.
[16:54.440] So if you're not happy with the excite or sound here,
[16:59.440] you can also switch the noise, for instance, to stereo,
[17:04.440] which sounds much better.
[17:11.440] And then you can try to shape it here maybe with...
[17:15.440] Yes, we have filter.
[17:45.440] You can see it gives you a lot of different variety on how to replicate certain sounds.
[17:56.440] That's always sounds like a tom.
[17:59.440] So now you come to the second problem,
[18:14.440] because not only did frequencies are important for certain objects,
[18:18.440] also the loudness of these overtones.
[18:22.440] So you have to decide if you want to amplify certain sounds or certain overtones
[18:31.440] and if so also how does it change over time?
[18:37.440] This is also important.
[18:39.440] We basically have to introduce not only loudness knobs here.
[18:52.440] Right, so when I dial down basically at the fundamental frequency,
[19:06.440] you have just a mind frequency.
[19:09.440] It comes automatically a bit more dissonant.
[19:17.440] So this is something you can do.
[19:20.440] And then you can also, if you want to introduce your segments,
[19:26.440] but not only for everything together for all these different partials,
[19:31.440] if you want to.
[19:32.440] So how they behave over time.
[19:38.440] If you want to do that, you don't need to do that,
[19:41.440] but you can do it. That's the thing.
[19:46.440] Maybe it.
[19:52.440] So they all kind of give you different results.
[20:08.440] So this is also something you can do.
[20:11.440] Interesting would be also to have something like.
[20:17.440] Where you spread the frequencies apart with the knob, right?
[20:21.440] So we have this, we have this here and then we maybe change either the frequencies itself.
[20:32.440] Something like this.
[20:35.440] I just do random stuff here.
[20:37.440] And usually I leave here the root untouched,
[20:42.440] but you can analyze certain parts and can really replicate something with this.
[20:47.440] But sometimes it's just fun to play around with this.
[20:53.440] So you can spread frequencies in different directions apart.
[21:07.440] And you bring the fundamental back in.
[21:15.440] So you get this kind of stuff, right?
[21:18.440] So spreading the frequencies.
[21:20.440] So loudness is important.
[21:22.440] Frequencies is important.
[21:24.440] The excited is important.
[21:25.440] You have so much freedom, but there's in the grid here that it's almost impossible to explain everything.
[21:34.440] So you can see you can also use different envelopes for different harmonics.
[21:40.440] So it's yeah, endless possibilities.
[21:43.440] You can also use more than just eight.
[21:45.440] You can use infinite amount of number of bandpass filters in here.
[21:53.440] And also did this already, of course.
[21:56.440] So here with the, yeah, I think we have only more, it's eight.
[22:00.440] So we have eight here, eight there.
[22:02.440] So this is also something you can do.
[22:03.440] You can use multiple of these in rows.
[22:05.440] So if you are done with this, you can do another bunch or just use your polygrid and duplicate it
[22:10.440] and use it more than input instead of an instrument here.
[22:13.440] Use it more than effect and then put stuff through that.
[22:18.440] Another idea is instead of using specific frequencies,
[22:26.440] you can also just say I want to introduce all the harmonics and then remove the harmonics I don't want to use.
[22:38.440] So I want to give you here also a quick overview of just double click here.
[22:43.440] This to get this back at C3.
[22:46.440] So now we have all the same frequencies everywhere.
[22:49.440] So let's say you want to,
[22:52.440] you want to target specific harmonics, right?
[22:55.440] So you want to say I want to have this is the fundamental and this is the second harmonic here.
[23:01.440] You can use the ratio for that.
[23:03.440] So put in here ratio.
[23:06.440] Pull this down, right?
[23:08.440] So zero.
[23:09.440] So it doesn't, isn't influenced by the keyboard.
[23:13.440] And then use a pitch in.
[23:18.440] Oh, maybe let's start over the bottom one.
[23:22.440] Turn this up, pull this down.
[23:25.440] So this is the first harmonic here.
[23:28.440] This is the second harmonic and you probably guessed it.
[23:32.440] We have to dial in to one.
[23:36.440] This is the third harmonic.
[23:42.440] Because this is here at C3,
[23:44.440] we modulate exactly to the right amount of frequency with us.
[23:50.440] So let's duplicate this here quickly.
[23:55.440] So this gives you also some freedom.
[24:02.440] If you want to target specific frequencies,
[24:06.440] or harmonics, a number of harmonics.
[24:09.440] So there are three, four, five, six, seven, and eight.
[24:18.440] Okay.
[24:19.440] So we have to start in here.
[24:21.440] Quick check.
[24:29.440] I need to back here to move all this.
[24:34.440] So now we have all the harmonics.
[24:40.440] And no surprise, this sounds exactly like a comp filter.
[24:44.440] Because the comp filter gives you exactly that.
[24:48.440] So when we go into the comp filter here,
[24:52.440] pitch on 262 hertz all the way up with the resonance,
[24:56.440] and then disconnect here our first idea,
[24:59.440] and just connect this here to output.
[25:06.440] Sounds almost, you have more overtones with this here.
[25:14.440] But this also gives you all the overtones.
[25:17.440] And with this approach, you basically can easily model harmonic,
[25:22.440] harmonic physical objects like guitar sounds or piano sounds,
[25:30.440] where you probably want to have all the harmonics in there,
[25:33.440] but some harmonics are a bit off in loudness.
[25:37.440] So you can target this with the SVF, for instance,
[25:40.440] and then remove with the notch certain frequencies.
[25:46.440] So right, if I use here an EQ,
[25:53.440] right, you can then pull in, put in a notch shared certain frequencies
[26:04.440] and remove the harmonics.
[26:07.440] So this is basically the opposite approach.
[26:09.440] Instead of introducing the harmonics you want,
[26:11.440] you introduce all the harmonics and to remove these harmonics you don't want.
[26:16.440] So it's the opposite approach, basically.
[26:20.440] There's also a nice filter in here where you can pull off all the upper harmonics.
[26:29.440] So this is more like for, like I said, for harmonic instruments.
[26:40.440] And this approach here is more for this harmonic instruments,
[26:46.440] because you can target specific frequencies.
[26:49.440] So here, this one makes no sense.
[26:51.440] One, two, three, four, five, six, seven, eight.
[26:55.440] Of course, we introduce here all the harmonics.
[26:58.440] And we can just fall back to comfort.
[27:01.440] So this only makes sense if you really target specific frequencies.
[27:05.440] So for instance, we don't want to go here maybe to 30,
[27:08.440] and here it was something like this.
[27:26.440] Now we probably need to introduce here also a different envelope.
[27:47.440] White noise.
[27:56.440] You can also use instead of SVF,
[28:01.440] you can use all the other filters, for instance, XP,
[28:04.440] where you have more options here for bandpass,
[28:07.440] here's deeper bandpass filter.
[28:12.440] It's a bit of work to exchange all of this, but you know.
[28:17.440] What gives you all different sounds depends,
[28:28.440] if you need steeper filters or more broader filters.
[28:39.440] XP4, XP bandpass 4.
[28:47.440] XP5, XP5, XP5, XP5, XP5, XP5, XP5.
[28:52.440] XP5, XP5, XP5, XP5, XP5, XP5, XP5, XP5, XP5, XP5, XP5, XP5, XP5.
[28:57.440] Maybe a bit more excitement here.
[29:10.440] So this is basically a rough introduce,
[29:13.440] introduce introduction to physical modeling, this is more
[29:19.040] kind of modal synthesis here. We have plant best filters and
[29:24.280] create sounds with that. You can also remove this whole part
[29:29.840] here and do something like the object delay, where instead of
[29:33.680] using here noise burst, it just use sounds coming in audio
[29:42.960] in, right? You use audio in and go here with a different
[29:47.120] sound from a synthesizer in or maybe you want to hit the desk,
[29:50.160] right? You have a microphone and then hit the desk with that
[29:53.840] and you can, yeah, pass that through these bandpass filters,
[29:59.840] which gives you all interesting results because they react
[30:02.680] to every sound you hit the desk, probably much, much different
[30:06.520] than having this noise sample. Actually, I could try this.
[30:10.840] It's actually not a bad idea. So we have the microphone here.
[30:15.880] And then I put this here on the second track.
[30:20.600] And then I have here this part.
[30:26.520] And this needs to be removed for us.
[30:41.080] Something like this.
[30:48.040] You probably want to have also here some kind of gate.
[30:58.520] Yeah, something like this.
[31:11.400] It needs a pitch.
[31:21.960] Um, pitch in.
[31:31.640] Yeah, it gives you also some kind of sounds that react a bit better
[31:36.840] or a bit more dynamic on more alive because the input signal
[31:41.800] is not always the same as here with the excited, with the noise burst.
[31:48.680] So yeah, that's also something you can do.
[31:51.960] And like I explained to you with the comfortors,
[31:54.520] also an easy part if you want to replicate certain things
[32:00.760] that are harmonic.
[32:03.400] So, um, let's you lose that here again.
[32:07.320] Or maybe go back here to the first instrument.
[32:13.960] To an instrument track.
[32:16.120] And in here,
[32:22.120] I limit the Q limit.
[32:33.160] So this sounds like guitar, but not really.
[32:44.360] Because now, um,
[32:47.400] yeah, the resonance is the fixed resonances are missing from the guitar body,
[32:50.680] which is the most important part of a, of a guitar, I think.
[32:55.480] And here we can try to figure it with an EQ.
[33:02.680] I have to bring certain frequencies out.
[33:14.920] Maybe also make this here polyphonic.
[33:29.240] Right, so you get this type of sound.
[33:33.640] Uh, but really, really cool gets, is this getting by using, um, a convolution,
[33:42.360] uh, convolution reverb here, and then using some real bodies.
[33:48.760] Not sure if I have something here.
[33:53.240] Um, TV scoop.
[34:03.560] But I have some kind of,
[34:08.040] some kind of your air responses on my desk.
[34:10.840] I can show you how the sounds, let me see.
[34:15.000] Um,
[34:19.880] yeah, I have something here downloaded, classical guitar body.
[34:24.440] So this is basically an air response through a guitar body.
[34:33.080] This is without, this is with,
[34:54.920] and this only introduces basically some kind of, uh, fixed frequencies from the guitar body itself.
[35:01.560] So it makes a huge difference having these, uh,
[35:04.360] four months on there.
[35:07.000] If you want to replicate certain instruments.
[35:16.200] Also probably the nice part.
[35:27.480] You bring in a bit of here.
[35:31.080] And then use Velo mode, which is influenced by the velocity of the keyboard, right?
[35:51.080] So how much you change basically at the, um, the frequency offset.
[36:01.720] Maybe also here.
[36:06.680] Maybe also here.
[36:11.000] Also here.
[36:24.280] So you get the idea.
[36:25.800] You try to make it then as dynamic as possible.
[36:30.920] Maybe let's try a different one here.
[37:01.880] Okay, so you can take hours to tune this.
[37:21.000] So it sounds like a real guitar.
[37:22.600] And at the end you don't even know if it sounds like a guitar at all.
[37:27.640] So, um, it's really, it's, that's a hard part basically,
[37:33.560] dialing all the fine nuances to get the sound, right?
[37:38.200] When it sounds, it's pretty, I think it sounds really okay.
[37:45.160] All it needs is probably here a transient control, no transient split.
[37:50.120] So we get a transient out and only put the rework on the tonal part here, super massive.
[37:57.640] It sounds nice, huh?
[38:19.560] I think there's still too much low end in there from the, let me see,
[38:25.960] from the comfort there, yeah, down here.
[38:55.960] Okay, yeah, that's the rough idea about physical modeling
[39:25.400] in physics studio, I think, um, replicating stuff like object from reason,
[39:31.000] replicating something like, um, chroma phone, uh, yeah, I think that's it.
[39:37.800] It's already a pretty long video, right?
[39:39.800] Let me see, 40 minutes.
[39:42.280] Okay, so if you liked the video, leave a like.
[39:45.720] If you have some questions, leave it in the comments.
[39:47.960] You can also, um, write there if I talked a lot of shit in this video.
[39:52.440] I have no problems with that.
[39:55.800] And I'll see you in the next video.
[39:57.080] Thanks for watching and bye.