Modelling a guitar's dynamics - Reaktor? Anything else?

For computer based music makers. Discussions about plug-ins and stand alone computer synth gear.
Post Reply
mikejm
Newbie
Newbie
Posts: 2
Joined: Sun Sep 21, 2014 7:09 pm

Modelling a guitar's dynamics - Reaktor? Anything else?

Post by mikejm » Sun Sep 21, 2014 7:23 pm

I am trying to build a synth patch that effectively emulates the feel of an electric guitar.

The most difficult part of this so far seems to be that a guitar does not follow a simple ADSR envelope. In theory, it's just an attack with a decay. But in reality it's more complex. The attack is not linear exactly. In addition, the pickups compress a bit as you get to the peak of the attack, so it rounds out a bit at the top. Then there's a quick decay, followed by a fade (not a true sustain).

In other words, the envelope is too complex to captured by a simple 4 phase envelope. I am therefore needing something that can produce more complex envelopes. I have been looking at this:

http://www.nireaktor.com/reaktor-tutori ... n-reaktor/

Which seems to suggest that in Reaktor you can build envelopes with multiple phases and customize them to a detailed extent. I get the impression it already comes with multi-break envelope modules premade. It looks to me like this would possibly work, but I don't understand Reaktor so I am speculating.

Further adding complexity to what I want to accomplish, a guitar's dynamics change depending on whether it is played hard or soft. I am hoping to use velocity data from each MIDI note to change the shape of the envelope. ie. The greater the velocity, the stronger the attack with greater the "compression" at the peak of the attack of the envelope.

Would Reaktor be a good candidate for the job?

I really don't know anything about physical modelling. This will be my first foray. But I've tried a number of softsynths and the main failing seems to be the simple ADSR (or ADSFR) envelopes simply can't convey the energy and explosiveness of a heavy guitar pluck (vs the gentle softness of a light one).

Any thoughts or suggestions?

gcoudert
Expert Member
Expert Member
Posts: 1198
Joined: Sat Dec 29, 2007 1:24 pm
Real name: Gilles
Gear: Roland Fantom X7 with Ultimate Keys expansion, Yamaha TX802, Roland JP-8080, Tascam DP-24SD.
Location: SE England

Re: Modelling a guitar's dynamics - Reaktor? Anything else?

Post by gcoudert » Sun Sep 21, 2014 7:35 pm

I use a real guitar!
GC

mikejm
Newbie
Newbie
Posts: 2
Joined: Sun Sep 21, 2014 7:09 pm

Re: Modelling a guitar's dynamics - Reaktor? Anything else?

Post by mikejm » Sun Sep 21, 2014 7:38 pm

gcoudert wrote:I use a real guitar!
Thanks. I have lots of guitars. I'm a guitar player. I'm trying to do things you can't do with guitars though and also get a slightly different sound than a true guitar.

I really just want to capture the dynamics of a guitar and apply them to a synth. It seems complex though. I still have to sit down and record a bunch of samples of playing my guitars and study the waveforms to know exactly how I need to emulate it. But I know already a simple ADSR isn't going to cut it.

mute
Active Member
Active Member
Posts: 640
Joined: Thu May 08, 2008 6:51 pm

Re: Modelling a guitar's dynamics - Reaktor? Anything else?

Post by mute » Sun Sep 21, 2014 11:22 pm

The newest version of String Studio by AAS looks pretty bad a*s... was mulling it over the other day.

https://www.applied-acoustics.com/string-studio-vs-2/

but be warned.. the demo video is really, really, really, really lame and not a fair representation (all of AAS's videos are horrible). There's a shining example of a bass guitar though if you skip to 5:40. You'd be better off lookin @ youtube for examples.

gcoudert
Expert Member
Expert Member
Posts: 1198
Joined: Sat Dec 29, 2007 1:24 pm
Real name: Gilles
Gear: Roland Fantom X7 with Ultimate Keys expansion, Yamaha TX802, Roland JP-8080, Tascam DP-24SD.
Location: SE England

Re: Modelling a guitar's dynamics - Reaktor? Anything else?

Post by gcoudert » Mon Sep 22, 2014 9:09 pm

mikejm wrote:
gcoudert wrote:I use a real guitar!
Thanks. I have lots of guitars. I'm a guitar player. I'm trying to do things you can't do with guitars though and also get a slightly different sound than a true guitar.

I really just want to capture the dynamics of a guitar and apply them to a synth. It seems complex though. I still have to sit down and record a bunch of samples of playing my guitars and study the waveforms to know exactly how I need to emulate it. But I know already a simple ADSR isn't going to cut it.
OK, I see what you're trying to do. This is going to be very complicated because of the multiple factors that shape the sound of a guitar from the string type (wound or not), gauge, palm muting strength, fret noise, finger pressure to picking strength and harmonics and I could go on forever. I would love to hear the result of your efforts though.
GC

c-level
Active Member
Active Member
Posts: 394
Joined: Tue Jan 31, 2012 1:57 am
Real name: Mike
Gear: Juno106, JX10, JX8P, JX3P, MKS50, SY22, SH201, Micromoog, CP 251, Korg Legacy, Triton Rack, Ableton 9, APC40, Max/MSP, Focusrite Pro40, Jazz Bass.
Band: C-Level
Location: Baltimore
Contact:

Re: Modelling a guitar's dynamics - Reaktor? Anything else?

Post by c-level » Tue Sep 23, 2014 12:41 am

i would say maybe start by analyzing guitar patches on synths that you have? when i tend to use guitar i use nylon guitar patches on my triton bc i dont have a real nylon. i know that probably doesnt help you to try to purely synthesize a guitar. my observation of a guitar strings tone fundamentally would that it would prob be more of a saw/tri wave than a square OSC waveform for one. both the volume and filter envelopes would have to be dynamic. id use a quick dynamic/key related filter envelope 'snap' to simulate the pick attack, with the env depth being varied, higher value for pick sounds, slower value and darker timbre for fingerstyle sounds. immediate attack for the amp, with a slower decay, some sustain and tiny release. keep in mind most of the guitar your used to hearing goes thru an amplifier. so amp/pedal modeling can help. maybe some resonant eqing and/or comb filtering to simulate the body of the acoustic... hope that helps...

MitchK1989
Active Member
Active Member
Posts: 663
Joined: Mon Apr 24, 2006 1:46 am
Location: Nova Scotia

Re: Modelling a guitar's dynamics - Reaktor? Anything else?

Post by MitchK1989 » Tue Sep 23, 2014 10:33 am

one thing you could do in a modular environment that you couldn't do elsewhere is run a sample of a guitar pluck through an envelope follower, then record the output of the envelope follower and use that as your attack/decay.

User avatar
ninja6485
Synth Explorer
Synth Explorer
Posts: 2766
Joined: Sun Apr 19, 2009 10:13 pm
Gear: Virus Ti, Jx-8p, Juno 60, Radias, Maschine, 101,303,606,707,727,808,909, odyssey, mirage, akai s5K/s2K/s1k, drumtraks, E6400ult, M1R, rx5, fizmo,d50
Band: Lyra, The Sun Worshipers
Location: Exton/ westchester

Re: Modelling a guitar's dynamics - Reaktor? Anything else?

Post by ninja6485 » Tue Sep 23, 2014 10:08 pm

You might also want to look into something that can give you multiple key groups or something that can mimic the way that only one note per string will play at a time, and playing a second note will mute the first on that string. It gets complicated since on a guitar you can often choose to play that note on a different string, so you get the option of having the previous note either ring through the change, or mute depending on which string you play it on. A velocity setting might be helpful there too, or a controller that lets you sweep up and down the key.
This looks like a psychotropic reaction. No wonder it's so popular...

Post Reply