How would you create a sitar sound in Rack?

I just noticed that the center frequency of the 4th Combulator is an octave too low. Now C4 (same as the 2nd). Should be C5…

Whatever…it’s just a POC…but still…annoying.

Tuned Comb Filters.vcv (20.9 KB)

Here’s another approach (which doesn’t sound like a sitar but shows resonance). You need Surge, BaconPlugs, Bog, MixMaster.

In the top row I make a voice playing random notes fine

In the second row I flatten the polyphony to a single signal and pass that into a pair of band pass and comb filters. But those bpf and comb are set up polyphonically and each polyphonic pair is tuned consistently with the modulation

May be an interesting technique to explore?

1 Like

It reminds me a bit of a kalimba, probably due to the resonance of the kalimba box. Thanks.

First of all…many thanks for your invaluable work on the Surge Synthesizer since it went free and open source and now also for the Surge VCV Rack modules…and the Airwindows collaberation.

Anyway…

The Surge Filter has several resonating filter models that we can use either as (shaper(s) for) Exciter(s), or as (shaper(s) for) Resonator(s).

The Surge Filter FX Comb + and FX Comb - models at near 100% feedback are pretty capable of Karplus Strong (Tuned Delay) resonation.

Effectively creating a feedback loop from output to input:

  • The Comb + is indeed just output to input. E.g. suitable for Plucked Strings (with vary short pulses of noise, if thats the excitement signal). Or Drums, depending on the Exciter spectrum.
  • The Comb - is simply inverted output to input, resulting in odd harmonics only and at half the frequency. E.g. for blown Tubes like a Clarinet (some longer attacks/holds/decays for the incoming noise, if that’s the exciter signal).

As said, several other Surge Filter models also can resonate towards longer decaying sounds. But you can also use them at static (resonant) frequencies to function as (part of) Resonator(s). There are more clean and more driven/dirty models, each with their own characteristics.

I guess I won’t have to tell you how Surge works…but to make the Surge Filter(s) follow 1 V/Oct (pitch in tune), first turn the Frequency all the way down, then press mod(x) and turn the Frequency all the way open (100% modulation). Then pres mod(x) again (to exit edit mod matrix mode). Now set the desired base frequency (default C4 = double click).

Of course you can also use filters to just shape the Exciter spectrum/behaviour and/or the output signal(s). Either at fixed frequencies or following pitch. E.g. Noise → Some filter → VCA x EG → (tuned) BPF → (tuned) Rez Comb.

For tight plucks/hits, you do need very fast attacks. Not all Envelope Generators offer fast enough attack/decay/release times/curves (they should preferable go down to 0 ms). Also, some VCA’s might not open/close fast enough.

One last thing: you might want to add compressor(s)/limiter(s) or a shaper like Instruo tanh in the loop since ‘external’ feedback loops can soon spiral out of control.

Such an external feedback loop could look like:

some signal → input mixer → into some module → output → to audio output (e.g. main mixer) AND back into the input mixer.

1 Like

Well…it’s definitely very possible to create beautiful Exciters and Resonators using Surge XT Filter(s).

I just used Surge XT Filter BPF (24 dB/Oct, Driven) into Surge XT Filter Comb + (at near 100% feedback to create a single resonator unit. Then copied that combo a few times and tuned them to various slightly detuned harmonic frequencies/notes to create a sort of multiband Modal Resonator.

I used a Noise burst (through VCA x (fast) EG (Vult Spank) into Surge XT Comb+ as the input signal (varying from Plucked String to Blown tube).

Then mixed the lot back together. The results are way above my initial expectations.

EDIT: Here’s the result. I guess the concept could be adapted to do a Sitar emulation with sympathetic strings…

4 Likes

That sounds like a very good acoustic guitar emulation…

This is pretty cool. Although, I will say it sounds more like a sarod to me than a sitar.

The tuned delay + module could also do a lot of these routings too and may be useful.

Thanks for the kind words also

most of the wave shapers I’ve tried have alias supression. Except Instruo is one of the few catering to the high aliasing market.

Haha… If anyone would address the aliasing component of waveshaping, it would be you!

Some generic stuff on waveshaping and aliasing for those who care…

Yes, as soon as waveshaping algorithms stray away from harmonic ratios, we soon are confronted with aliasing when operating at audiorate signals. Not just think of obvious so called waveshapers. Also consider stuff like compressors/limiters that effectively also bend/shape waves and can and do effect the waveshape and therefore the spectrum (very dependent on depth/knee and attack/release times).

Other (potential) offenders are waveshaping algorithms that might in principle respect harmonic relations, but that create sidebands that might reach '‘past’ Nyquist, like AM and FM/PM.

Indeed, since aliasing frequencies bounce back from Nyquist back into the audible spectrum, they get mixed up with the original audible spectrum, we can not really easily filter them out after the fact. Bit like sorting out white and black beans that got mixed up in a jar.

So, prevention at the source is the better option. Good to hear that many waveshapers incorporate ways to suppress aliasing. And that some don’t…

Although, there is room for some aliasing, as long as levels/effects are limited to adding some liveliness in the higher end of the spectrum (think original SuperSaw or all sorts of ‘imperfect’ analog circuits).

Things don’t have to cause panic as long as the aliasing frequencies don’t introduce very distinguishable and audible inharmonics where not wanted. Especially annoying if they move around the spectrum as pitch varies, but in inverse direction.

The Surge XT Tuned Delay and Tuned Delay+ are, as their name suggests, capable of Karplus Strong synthesis, which is basically tuned delay. The + version incorporates feedback with filtering options, offering the basic components for Karplus-Strong in a single module.

Oooh… I really like this. Like a steel guitar/banjo through delays, very nice!

yeah. I made “Shaper” years ago because I was tired of waveshapers that didn’t do anything about aliasing. But soon after that a primary offender came out with an update. These days many/most waveshapers have alias suppression. I was surprised tanh doesn’t, since many of the Instruo modules are so well thought out. I guarantee you the analog original doesn’t alias!

One more option to create a Sitar with 12 sympathetic strings…

Again using Surge XT. But this time, using the Surge XT String VCO . This module can do both the main plucked string (that plays the notes) and the 12 sympathetic strings (tuned as mentioned before).

The Surge XT String module offers 2 strings. So, at 2 strings per module, we need at least 6 modules for the 12 sympathetic strings. And 1 module as the main (pitched) plucked string.

For the 1 plucked string, set the second string to no decay = silence/deactivate, since we want 1 string here.

The 2nd string can be detuned by a maximum of +/- 100 cents (an octave up/down). Each has its own delay time, which in this case are set to near 100%.

The module has trigger in and an audio in (that must be selected via the menu). For the plucked string we can either trigger it externally with a noise burst, or use the trigger (loosing control over the exciter envelope). Also, for the plucked string we connect the V/Oct. The sympathetic string modules receive only audio in from the plucked string.

Fiddle with the delay times, stiffness (stiffer = high frequencies linger for longer, so brighter ).

As before: mix the lot together at the desired levels.

Here is my latest which adds a 2nd octave of sympathetic resonator strings, a plucked drone and quite a bit of CV automation. This is a 30 minute performance.

Pseudoraga Minimal-Jawari Sitar Drone and Tabla w Polyrhythms and Automation-1.vcv (12.9 KB)

1 Like

These modules are all polyphonic so if you need 12 sympathetic strings you can also just run it 12-way poly (and then actually get 24 ‘strings’ which are slightly detuned).

Sure, can do. But…GUI wise, setting and interactiing with the various module parameters will be a bit of a challenge. :grin:

That’s why I use polycon a lot. It basically lets me have offsets per channel via th modulation system.

Here is my latest pseudo-raga patch and is the last I plan to post on this, for now. This patch makes use of Vult for most voices, now including Vult Trummor for the tabla emulation as Trummer along with Tangents and Decline allowed me to tweak the two tabla voices to a degree that I am satisfied with. Proteus and Jawari are still important elements of this patch.

Pseudoraga Minimal-Jawari Sitar Drone and Trummer Tabla w Polyrhythms and Automation-1.vcv (12.7 KB)

7 Likes

I love seeing the evolution and progress of the question – more enlightening than just a final polished patch!

3 Likes