Strings and saw waves

Anything modular synth related that is not format specific.

Moderators: Joe., lisa, luketeaford, Kent

User avatar
cyberdine
Wiggling with Experience
Posts: 306
Joined: Wed Jan 06, 2016 8:13 pm
Location: New York

Strings and saw waves

Post by cyberdine » Mon May 11, 2020 7:40 pm

I’m sitting here at home, day blah, listening to the Chromatics track ‘Saturday - Instrumental’ and idly pondering: if the saw wave sounds like a bowed string instrument, how do I synthesize the lushest string sound using saw waves? I love a good saw wave sound, but I’ve never made it sound like a violin. Suppose a few saws, subtly detuned and then with a brace of attenuated LFOs modulating the pitch around a centre frequency? Loads of reverb is a given. Probably need an attenuated LFO to modulate the volume of each saw too. Maybe a few subtly modulated low pass filters on each. Have I just answered my own question in posting this?

User avatar
cyberdine
Wiggling with Experience
Posts: 306
Joined: Wed Jan 06, 2016 8:13 pm
Location: New York

Re: Strings and saw waves

Post by cyberdine » Mon May 11, 2020 7:42 pm

Realise I’m forgetting the overall amp envelope, slow attack, long decay, at least played legato.

User avatar
wuff_miggler
Wiggling with Experience
Posts: 496
Joined: Wed Feb 19, 2020 11:11 pm
Location: Disney, Australia

Re: Strings and saw waves

Post by wuff_miggler » Mon May 11, 2020 7:52 pm

i think you need quite a bit more than what you've mentioned so far - try the following:

*True Stereo Chorus pedal
*A formant filter - to help simulate the wooden body of violin.
*White noise - modulated with envelope follower (super important)
*Subtle phasing to help with the rosin sound.
*Phrasing your performance as a string player would - a ribbon controller is handy for this.
*The Steiner String filter
*A natural sounding reverb - to place it in acoustic space.
*Further equalisation to boost the high frequency content past 15khz (whatever of it exists anyway)

User avatar
trentpmcd
Wiggling with Experience
Posts: 350
Joined: Fri Mar 20, 2015 4:00 pm
Location: NH

Re: Strings and saw waves

Post by trentpmcd » Mon May 11, 2020 8:02 pm

Over the winter I experimented with trying to get a real string sound. I didn’t have some tools I wish I did (i.e., FFB). I did pretty much what you said – different saws (and sometimes the same saw split several ways) but a little bit of square going through three different filters. One saw was also FMed with noise. I used velocity on a CV envelope for articulation and mod wheel for volume. I changed the filter slightly different on each filter with the mod wheel to change the timbre when the volume grew. Same with velocity – changed different filters as well as the attack. Overall, I used 6 envelopes, three filters, six VCOs (sometimes - see later), 4 VCAs, 2 attenuverters, 4 mixers and I don’t remember what all else for each single instrument , of course changing filters to make different sounds (Violin, viola, cello, bass). This gave a background ensemble. I then went down to just two saws split through all of this - one with noise, and played each line three or four times over to make it sound like a large ensemble, making sure the articulation, etc., was slightly different each time. That is also the solo sound - two saws, one FMed by noise. You can tell me if you think it was worth all of time... (The harpsichord was a Kawai K5000, the rest my modular). (I've posted this like 4 times on the forum, so I will just put up a link: https://youtu.be/EnV0z_K7xNY) The front panel shot of the modular was pretty much the string sound I used.

User avatar
commodorejohn
Ultra Wiggler
Posts: 975
Joined: Fri May 03, 2013 4:19 pm
Location: Placerville, CA

Re: Strings and saw waves

Post by commodorejohn » Mon May 11, 2020 8:19 pm

Have you tried running it through a wooden resonator box? Perhaps a maple or pearwood structure with a spruce top? Maybe with a secondary structure made of a really hard wood like ebony? Surely there must be a module for this, although you might need to run it as an outboard processor rather than try to fit it in Eurorack dimensions...
Computers: Amiga 1200, DEC VAXStation 4000/60, DEC MicroPDP-11/73
Synthesizers: Roland JX-10/SH-09/MT-32/D-50, Yamaha DX7/V50/TX7/TG33/FB-01, Korg MS-20 Mini/ARP Odyssey/DW-8000, Ensoniq SQ-80

"'Legacy code' often differs from its suggested alternative by actually working and scaling." - Bjarne Stroustrup

KSS
Super Deluxe Wiggler
Posts: 2045
Joined: Mon Jan 25, 2016 7:28 am

Re: Strings and saw waves

Post by KSS » Mon May 11, 2020 8:31 pm

Missing from many string patches are pulse waves. Saws alone often lack basic string tone which is then -tried to- fix using items further down the chain.

Pretty sure it was a Bob moog column where I first read this. A very long time ago. He was right.

Think about the physical model often used to justify ramp-saws as correct starting WF. Bow grabs string, rising resistance of string until -supposedly- instant release, repeat. = ramp-saw.

But the release isn't instant. And it's not necessarily a linear buildup. Mix a narrow pulse into your saw, reduce the saw level compared to the pulse and go from there. Assumes the pulse and fat part of the saw are at the same 'end' of the WF. Your PWM becomes powerful in this patch too.

And the type of PWM matters. PWM generated from a saw core will extend one direction, while a tri-core PWM -if using the triangle as its basis- will extend in both directions. PWM from a sine -use a comparator if your VCO doesn't have this natively- also extends both directions, but in a non-linear way. All these PWM variations matter, and can 'fill-in' for some of the 'necessary' and 'expected' string sound modules and patch points.

edit: The rising resistance of the string to bow motion also imparts a very slight pitch increase -or modulation- also missing in naive string patches.
Last edited by KSS on Mon May 11, 2020 8:35 pm, edited 1 time in total.

KSS
Super Deluxe Wiggler
Posts: 2045
Joined: Mon Jan 25, 2016 7:28 am

Re: Strings and saw waves

Post by KSS » Mon May 11, 2020 8:33 pm

commodorejohn wrote:
Mon May 11, 2020 8:19 pm
Have you tried running it through a wooden resonator box? Surely there must be a module for this,
That's what FFB's are for and what Param EQ's can shine at doing.

User avatar
cyberdine
Wiggling with Experience
Posts: 306
Joined: Wed Jan 06, 2016 8:13 pm
Location: New York

Re: Strings and saw waves

Post by cyberdine » Mon May 11, 2020 10:25 pm

trentpmcd wrote:
Mon May 11, 2020 8:02 pm
Over the winter I experimented with trying to get a real string sound. I didn’t have some tools I wish I did (i.e., FFB). I did pretty much what you said – different saws (and sometimes the same saw split several ways) but a little bit of square going through three different filters. One saw was also FMed with noise. I used velocity on a CV envelope for articulation and mod wheel for volume. I changed the filter slightly different on each filter with the mod wheel to change the timbre when the volume grew. Same with velocity – changed different filters as well as the attack. Overall, I used 6 envelopes, three filters, six VCOs (sometimes - see later), 4 VCAs, 2 attenuverters, 4 mixers and I don’t remember what all else for each single instrument , of course changing filters to make different sounds (Violin, viola, cello, bass). This gave a background ensemble. I then went down to just two saws split through all of this - one with noise, and played each line three or four times over to make it sound like a large ensemble, making sure the articulation, etc., was slightly different each time. That is also the solo sound - two saws, one FMed by noise. You can tell me if you think it was worth all of time... (The harpsichord was a Kawai K5000, the rest my modular). (I've posted this like 4 times on the forum, so I will just put up a link: https://youtu.be/EnV0z_K7xNY) The front panel shot of the modular was pretty much the string sound I used.
Wow.. the solo sounds like a synth but the staccato violas(?) are pretty convincing in the mix. This is really excellent. Yes, it was worth the time in my opinion! Thank you for the detailed notes.

User avatar
cyberdine
Wiggling with Experience
Posts: 306
Joined: Wed Jan 06, 2016 8:13 pm
Location: New York

Re: Strings and saw waves

Post by cyberdine » Mon May 11, 2020 10:28 pm

KSS wrote:
Mon May 11, 2020 8:31 pm
Missing from many string patches are pulse waves. Saws alone often lack basic string tone which is then -tried to- fix using items further down the chain.

Pretty sure it was a Bob moog column where I first read this. A very long time ago. He was right.

Think about the physical model often used to justify ramp-saws as correct starting WF. Bow grabs string, rising resistance of string until -supposedly- instant release, repeat. = ramp-saw.

But the release isn't instant. And it's not necessarily a linear buildup. Mix a narrow pulse into your saw, reduce the saw level compared to the pulse and go from there. Assumes the pulse and fat part of the saw are at the same 'end' of the WF. Your PWM becomes powerful in this patch too.

And the type of PWM matters. PWM generated from a saw core will extend one direction, while a tri-core PWM -if using the triangle as its basis- will extend in both directions. PWM from a sine -use a comparator if your VCO doesn't have this natively- also extends both directions, but in a non-linear way. All these PWM variations matter, and can 'fill-in' for some of the 'necessary' and 'expected' string sound modules and patch points.

edit: The rising resistance of the string to bow motion also imparts a very slight pitch increase -or modulation- also missing in naive string patches.
Thank you for the detailed reply. What I read here is the importance of noise - in the sense of a somewhat random, unpredictable (but attenuated) response. The noise has to be like the breath, the uncertain flutter of the bow imperfectly drawn across the strings, to make them sing.

User avatar
cyberdine
Wiggling with Experience
Posts: 306
Joined: Wed Jan 06, 2016 8:13 pm
Location: New York

Re: Strings and saw waves

Post by cyberdine » Mon May 11, 2020 10:30 pm

commodorejohn wrote:
Mon May 11, 2020 8:19 pm
Have you tried running it through a wooden resonator box? Perhaps a maple or pearwood structure with a spruce top? Maybe with a secondary structure made of a really hard wood like ebony? Surely there must be a module for this, although you might need to run it as an outboard processor rather than try to fit it in Eurorack dimensions...
It sounds like a physical modeling module, like the one from Mutable instruments which name escapes me right now. I take your point though: it’s not just the overtones in the saw that give the bowed string its character, it’s the resonance in the body, the way the organic matter feels the sound in its own due time.

User avatar
cyberdine
Wiggling with Experience
Posts: 306
Joined: Wed Jan 06, 2016 8:13 pm
Location: New York

Re: Strings and saw waves

Post by cyberdine » Mon May 11, 2020 10:32 pm

wuff_miggler wrote:
Mon May 11, 2020 7:52 pm
i think you need quite a bit more than what you've mentioned so far - try the following:

*True Stereo Chorus pedal
*A formant filter - to help simulate the wooden body of violin.
*White noise - modulated with envelope follower (super important)
*Subtle phasing to help with the rosin sound.
*Phrasing your performance as a string player would - a ribbon controller is handy for this.
*The Steiner String filter
*A natural sounding reverb - to place it in acoustic space.
*Further equalisation to boost the high frequency content past 15khz (whatever of it exists anyway)
Interested in the idea of boosting past 15khz. At the risk of derailing the thread, I can’t hear that high (anymore, maybe ever?) if you can’t hear it, do you think it still adds something to the sound?

User avatar
cyberdine
Wiggling with Experience
Posts: 306
Joined: Wed Jan 06, 2016 8:13 pm
Location: New York

Re: Strings and saw waves

Post by cyberdine » Mon May 11, 2020 10:35 pm

KSS wrote:
Mon May 11, 2020 8:31 pm
Missing from many string patches are pulse waves. Saws alone often lack basic string tone which is then -tried to- fix using items further down the chain.

Pretty sure it was a Bob moog column where I first read this. A very long time ago. He was right.

Think about the physical model often used to justify ramp-saws as correct starting WF. Bow grabs string, rising resistance of string until -supposedly- instant release, repeat. = ramp-saw.

But the release isn't instant. And it's not necessarily a linear buildup. Mix a narrow pulse into your saw, reduce the saw level compared to the pulse and go from there. Assumes the pulse and fat part of the saw are at the same 'end' of the WF. Your PWM becomes powerful in this patch too.

And the type of PWM matters. PWM generated from a saw core will extend one direction, while a tri-core PWM -if using the triangle as its basis- will extend in both directions. PWM from a sine -use a comparator if your VCO doesn't have this natively- also extends both directions, but in a non-linear way. All these PWM variations matter, and can 'fill-in' for some of the 'necessary' and 'expected' string sound modules and patch points.

edit: The rising resistance of the string to bow motion also imparts a very slight pitch increase -or modulation- also missing in naive string patches.
Also, to respond to your point about a PWM pulse- I hadn’t even considered that, but it makes sense in principle- the square wave simulates hollowness of the body of the stringed instrument, and I can imagine how a careful flutter of the pulse width could make the sound ring in a sympathetic way.

User avatar
cyberdine
Wiggling with Experience
Posts: 306
Joined: Wed Jan 06, 2016 8:13 pm
Location: New York

Re: Strings and saw waves

Post by cyberdine » Mon May 11, 2020 10:37 pm

cyberdine wrote:
Mon May 11, 2020 10:32 pm
wuff_miggler wrote:
Mon May 11, 2020 7:52 pm
i think you need quite a bit more than what you've mentioned so far - try the following:

*True Stereo Chorus pedal
*A formant filter - to help simulate the wooden body of violin.
*White noise - modulated with envelope follower (super important)
*Subtle phasing to help with the rosin sound.
*Phrasing your performance as a string player would - a ribbon controller is handy for this.
*The Steiner String filter
*A natural sounding reverb - to place it in acoustic space.
*Further equalisation to boost the high frequency content past 15khz (whatever of it exists anyway)
Interested in the idea of boosting past 15khz. At the risk of derailing the thread, I can’t hear that high (anymore, maybe ever?) if you can’t hear it, do you think it still adds something to the sound?

Also, a ribbon controller makes sense too- the more I synthesize, the more I appreciate how important the amp envelope is the the feeling of the sound.

User avatar
wuff_miggler
Wiggling with Experience
Posts: 496
Joined: Wed Feb 19, 2020 11:11 pm
Location: Disney, Australia

Re: Strings and saw waves

Post by wuff_miggler » Mon May 11, 2020 10:59 pm

cyberdine wrote:
Mon May 11, 2020 10:32 pm
wuff_miggler wrote:
Mon May 11, 2020 7:52 pm
Interested in the idea of boosting past 15khz. At the risk of derailing the thread, I can’t hear that high (anymore, maybe ever?) if you can’t hear it, do you think it still adds something to the sound?
well - a gradual HP shelf boosting gradually upwards - perhaps AFTER combining the saw element with a Low passed noise element may still do something you can hear?

also ...wanting to add to my list above - some kind of resonance elements in a filter - perhaps using a low pass gate instead of a VCA..doepfer's LPG self resonates - using that somewhere in your signal path could provide some added realism adding to the instrument body resonance .

User avatar
wuff_miggler
Wiggling with Experience
Posts: 496
Joined: Wed Feb 19, 2020 11:11 pm
Location: Disney, Australia

Re: Strings and saw waves

Post by wuff_miggler » Mon May 11, 2020 11:03 pm

commodorejohn wrote:
Mon May 11, 2020 8:19 pm
Have you tried running it through a wooden resonator box? Perhaps a maple or pearwood structure with a spruce top? Maybe with a secondary structure made of a really hard wood like ebony? Surely there must be a module for this, although you might need to run it as an outboard processor rather than try to fit it in Eurorack dimensions...
interested in this idea commodorejohn - have you tied it?

assuming on one side of a wooden box - there's a transducer being amped - feeding signal - and on the otherside - a piezo - going back into a preamp....back into the eurorack system. i have ben thinking of several ideas of this ilk - have not embarkred on any of the diy work - would love to hear if you've had success.

Sorry to hijack cyberdine - just very timely topic as i've been mulling over the same question in my head for a few months.

User avatar
commodorejohn
Ultra Wiggler
Posts: 975
Joined: Fri May 03, 2013 4:19 pm
Location: Placerville, CA

Re: Strings and saw waves

Post by commodorejohn » Tue May 12, 2020 2:10 am

I'd love to hear someone try that :lol:

In all seriousness, though, yeah, I think it depends to a great degree on how realistic "realistic" is for you. Trent's demo gets surprisingly close to "passable," without getting too crazy deep into full-fledged modelling (though it could really use either a nice lush chorus or a pile of multi-tracking.) And, as he mentions, adding a bit of modulated pulse really brings in a level of animation that plain sawtooth just doesn't have. But how much is "enough" for what you, specifically, need? That's a bit more complex of a question.
Computers: Amiga 1200, DEC VAXStation 4000/60, DEC MicroPDP-11/73
Synthesizers: Roland JX-10/SH-09/MT-32/D-50, Yamaha DX7/V50/TX7/TG33/FB-01, Korg MS-20 Mini/ARP Odyssey/DW-8000, Ensoniq SQ-80

"'Legacy code' often differs from its suggested alternative by actually working and scaling." - Bjarne Stroustrup

User avatar
cptnal
Super Deluxe Wiggler
Posts: 3747
Joined: Wed Jun 14, 2017 2:48 am
Location: People's Republic of Scotland

Re: Strings and saw waves

Post by cptnal » Tue May 12, 2020 3:46 am

Would it be off topic to suggest a triangle wave?

I did this a few months ago, more as an exercise in expression and dynamics than trying to achieve a specific sound. I can't remember the exact signal path, but the main idea was using the difference patch to control modulation depth. You can hear that higher harmonics come in with louder and higher notes.


User avatar
wuff_miggler
Wiggling with Experience
Posts: 496
Joined: Wed Feb 19, 2020 11:11 pm
Location: Disney, Australia

Re: Strings and saw waves

Post by wuff_miggler » Tue May 12, 2020 4:02 am

cptnal wrote:
Tue May 12, 2020 3:46 am
very impressive cptnal!
no need for exacts - but if you can try - would love some more details regarding the patch. what do you mean by a difference patch in particular.

also - doubt the OP will mind if the goal is to get some sick strings! :D

User avatar
galanter2
Common Wiggler
Posts: 168
Joined: Mon Aug 04, 2014 8:32 pm
Location: Texas A&M University
Contact:

Re: Strings and saw waves

Post by galanter2 » Tue May 12, 2020 4:48 am

I’ve been using a couple Spitfire string sample libraries that support multiple articulations and controllers. If the goal is to add as realistic a string section as possible you aren’t going to synthesize something better. The point of using a synth here should be to make intentionally string-LIKE timbres that knowingly go to a different place.

IMHO

User avatar
wuff_miggler
Wiggling with Experience
Posts: 496
Joined: Wed Feb 19, 2020 11:11 pm
Location: Disney, Australia

Re: Strings and saw waves

Post by wuff_miggler » Tue May 12, 2020 4:51 am

really good point galanter2. infact while thinking about adding resonating speakers and such to my setup - i actually said to myself - "ffs if im just trying to get real shit....sample it mutherfucker".

really good to remember your point - "try to synthesize something different"...

x 100000..which reminds me of a user her bartlebooth - amazing "modular string" stuff here....just blows my mind:


Parnelli
Super Deluxe Wiggler
Posts: 1278
Joined: Wed May 03, 2017 6:04 pm
Location: Utah

Re: Strings and saw waves

Post by Parnelli » Tue May 12, 2020 6:09 am

It never fails to amaze me the golden nuggets of information available here, thanks for this!

Keltie
Veteran Wiggler
Posts: 672
Joined: Fri Apr 14, 2017 11:33 am

Re: Strings and saw waves

Post by Keltie » Tue May 12, 2020 6:27 am

The reason that saw is the canonical starting point for strings is that the harmonic series contains all harmonics, even and odd, as does a tensioned string immobile at both ends. Whilst it’s useful to visualise the process of horsehair gripping string etc etc, that isn’t really relevant to the wave shape, and isn’t a process that repeats x100 times a second through the course of the note. It will tension the string sharp, which is why a slight touch of pitch enveloping can work, at the start of a note.

There’s an awful lot more to it than that, of course, and mixing in another wave shape is entirely valid, but in basic terms, and leaving aside complicating factors, if you wanted to synthesise just about any kind of string sound, and you just had one choice of classical waveform to do it, it’s saw all the way.

ricko
Common Wiggler
Posts: 229
Joined: Sat Dec 03, 2016 5:36 am
Location: Sydney, Australia

Re: Strings and saw waves

Post by ricko » Tue May 12, 2020 9:52 am

Keltie wrote:
Tue May 12, 2020 6:27 am
Whilst it’s useful to visualise the process of horsehair gripping string etc etc, that isn’t really relevant to the wave shape, and isn’t a process that repeats x100 times a second through the course of the note.
Interestingly, as well as this saw/tri motion from the bow, there is also a rotational saw/tri happening which has a measurable effect on the sound: the bow pulls the string but it also twists, and the rotational grip/slippage can be a higher frequency, but synched to the bow slip to one extent or another.

Regards
Rick

Keltie
Veteran Wiggler
Posts: 672
Joined: Fri Apr 14, 2017 11:33 am

Re: Strings and saw waves

Post by Keltie » Tue May 12, 2020 10:59 am

ricko wrote:
Tue May 12, 2020 9:52 am
Keltie wrote:
Tue May 12, 2020 6:27 am
Whilst it’s useful to visualise the process of horsehair gripping string etc etc, that isn’t really relevant to the wave shape, and isn’t a process that repeats x100 times a second through the course of the note.
Interestingly, as well as this saw/tri motion from the bow, there is also a rotational saw/tri happening which has a measurable effect on the sound: the bow pulls the string but it also twists, and the rotational grip/slippage can be a higher frequency, but synched to the bow slip to one extent or another.

Regards
Rick
That’s interesting, and not something I knew about. It goes to show just how complex acoustic instrument sounds are....

User avatar
cptnal
Super Deluxe Wiggler
Posts: 3747
Joined: Wed Jun 14, 2017 2:48 am
Location: People's Republic of Scotland

Re: Strings and saw waves

Post by cptnal » Tue May 12, 2020 11:25 am

wuff_miggler wrote:
Tue May 12, 2020 4:02 am
cptnal wrote:
Tue May 12, 2020 3:46 am
very impressive cptnal!
no need for exacts - but if you can try - would love some more details regarding the patch. what do you mean by a difference patch in particular.

also - doubt the OP will mind if the goal is to get some sick strings! :D
Thank you, and sorry - my bad. I meant rate-of-change patch. It's described in the manual for the ADDAC215 (on whose panel it's marked "Difference") way better than I could...

http://www.addacsystem.com/contents/pro ... &H_A_1.pdf (page 5)

Post Reply

Return to “Modular Synth General Discussion”