xoxos

xoxos · Post by **xoxos** » Thu Sep 15, 2011 5:37 am

i'm developing a modal synth to the radian bundle, similar to radian but for rectangular plates.

i thought i'd add 1d structures as well since it was easy to do.

this is 1 dimension modulated (phase and gain) by striking position with partial stretching implemented. there's no gain (spectral gain, eg. different parts of the spectrum being louder or quieter yet) so i threw a quick enveloped sv lowpass on it.

http://www.xoxos.net/temp/modal01.mp3

it's got a ways to go before it sounds like stuff, to my ears it sounds like it's going to get closer there than karplus-strong.

xoxos · Post by **xoxos** » Sat Sep 17, 2011 12:08 am

not convinced, eh..

http://www.xoxos.net/temp/modal02.mp3

low notes with strong coupling to self causing lovely rippling effect..

gradually coaxing partials to behave more like physical object, i thought this sounded keen.

xoxos · Post by **xoxos** » Sat Sep 17, 2011 6:41 am

http://www.xoxos.net/temp/modal03.mp3

still not convinced?

xoxos · Post by **xoxos** » Sun Sep 18, 2011 3:38 am

http://www.xoxos.net/temp/modalbass2.mp3

tough audience

that still doesn't say "thick & rigid metal string" to you??

man, i bet the waveform even looks right.

(whoever took my sweet customed fretless sx go to

)

Genetic_Junk · Post by **Genetic_Junk** » Sun Sep 18, 2011 3:51 am

Cool! The lowest notes on modal3 sound pretty realistic. I like the xylophone and vibraphone sounds too.

If you need some waveforms for comparisons, I have an electric bass I can record.

xoxos · Post by **xoxos** » Sun Sep 18, 2011 4:10 am

for bass, any way it's patched isn't going to sound good for more than ~half a dozen keys. the partial stretch and scaling would need to be adjusted for the change in string length to make it sound the same, et al., atm it's a basic dsp core, not an instrument.

thanks, it's nice to have someone else hear the potential, or i'm chasing my own tail.

Genetic_Junk · Post by **Genetic_Junk** » Sun Sep 18, 2011 4:25 am

No problem, in fact, seeing as I am a potential beneficiary of the result, it's in my interest to give feedback.

paulhart · Post by **paulhart** » Sun Sep 18, 2011 8:19 pm

Xoxos, Hello... I just bumped into this thread. I have valued your contributions and make a point of clicking through on your link at KVR whenever I see a posting. I went for the Bass, excellent, my limited capacity of discernment is very comfortable imagining that this is a sample, not a creation of your fertile mind. Well done, once again. I we left only with the tantalizing sound bytes, or can others make creations with modal??? Thank you again.
Paul

jdt · Post by **jdt** » Sun Sep 18, 2011 8:35 pm

Sounds pretty damn good to me! The modal bass mp3 sounds so close to my 6-string bass it was hard to believe it was synthetic.

xoxos · Post by **xoxos** » Sun Sep 18, 2011 8:50 pm

thank you

my perception is fairly comfortable seeing the ideal content between real and synthesized forms..

tbh it probably won't go too much further.. atm it's a dsp core with a few sliders and a keyboard connected to it in synthedit. the actual instrument is/was intended to focus on 2d modal synthesis, with the 1d as an addition. i'd expect about 3 weeks until release.

so the bass here will be augmented as an instrument with tracking and velocity modulations, the end result being something that sounds similar to an ideal string.

usually when i use my modeled strings, i'll use a different instance of the plugin for each string, and split the sequence up. that way each string has better dispersion and damping settings.

this will please me, but i couldn't market it to people wanting bass emulation (unless they only enjoy 'perfect' playing).

what i am hearing in this that i haven't heard in other emulations is that the partials are phase independent.. like on the very low notes in some of the other examples you have that 'percussed metal wire fence' effect. to me it sounds like a bass when the partials are slightly stretched.. real basses aren't perfect due to resonances/coupling.. which would be an immense cpu load to model to expectation of course.

(i may also be able to improve this by applying a slight pitch envelope on the attack like radian uses).

what i am *not* hearing here are those coupled resonances, and the imperfections of complex human contact..

yesterday i was playing guitar (people keep stealing my instruments, but i managed to afford a used bullet strat recently) and noticed one of the main tones that emulations miss (they miss loads and loads of things of course..) is the part of the string from the finger to the neck.. (which i poorly modeled in pling 2??)

..when you have one fret held then press another, say two frets higher.. if you release the first fret, there's usually a fairly audible vibration of the higher string section (especially in fretted bass if i remember, it's been so long..).

i was thinking to myself.. the intervals going *upwards* from the frets aren't scaled the same, so you can't accurately emulate them by quietly triggering another note..

..and you'd need the composer to manually enter an entirely different sequence for string releases.. the same way i haven't seen any "satisfactory" engine for discerning which string someone is meaning to play when playing a guitar emu on a keyboard.. those are decisions that should be left to the performer.. and current vst users largely aren't prepared for the intense sequencing that modeling requires (i recently saw this cited as the rationale for the decline of physical modeling).

but thanks

i was thinking that it sounded like a bass, now someone else is saying it too

g.j. on the recording offer.. i think the focus here will be on the overall 1d/2d synthesizer. i'm making plans to obtain a bass by the end of the year, douglas if i can't find another SX sb 301. i *really* liked the one i had....

Fritze · Post by **Fritze** » Sun Sep 18, 2011 9:11 pm

Wow! The first two mp3s didn't really click with me (though interesting) but modal03 and modalbass2 have VERY cool sounds in it. Some of the bass notes sound stunningly real to me. Some of the to me ideal bass sounds are in there. Not kidding. I'm used to hear very good audio demos from you and these two are no exceptions. I would love to have a plugin that can do these bass sounds in a predictable and playable/programmable way. Thanx for working on that.

xoxos · Post by **xoxos** » Sun Sep 18, 2011 11:30 pm

maybe with some attack shaping it might be equivalent to a rompler (with some added dynamic behaviour) in this form. we'll see!

xoxos · Post by **xoxos** » Mon Sep 19, 2011 9:18 pm

aah

up until 2am last night finding and fixing a bug with this.. i don't know when i'll have some more audio but it is certainly more accurate an emulation now

xoxos · Post by **xoxos** » Tue Sep 20, 2011 5:59 am

okay, duh... fixed the bug, this should sound immensely more real...

http://www.xoxos.net/temp/modalbass3.mp3

excuse the clicky attacks... the thing about comparing it to samples is that, while samples can be analysed to give information about a system, what i'm building here (at least at this extent) is pretty much a known system.. rendering partials, whoopee :p i can get the phase, amplitude and frequency right, then i can augment it with an attack, coupling, filtering, stuff like that..

at this point it is pretty much psychoacoustic because the audio this renders doesn't have any kind of dimensional point, it doesn't correlate to a pickup position, and i'm way out of knowing what a pickup is going to make an unknown waveform look like

anyway.. hopefully that sounds something more like a thick metal string!

and it loses that emphasis tone on the plucking position that sounded wrong with the last audio... (there are some mellower notes in the middle that don't sound so emulative but i was fascinated that the waveform looked like an ogee)

xoxos · Post by **xoxos** » Sat Oct 15, 2011 11:48 pm

where to hit a square membrane (to 64 harmonics on each axis)