Topic: Stringteq

Hi all!

I have a dream and a wish for Modartt which has a long and successful experience on modeling:
Modeled string instrument. (And I am sure this has been asked before...)

First, few reasons for making it:
To my experience sampled strings are either lacking nuances, sounding dull or they are very unpredictable and complicated to use. The latter means that samples have variety but it's difficult to control them. They may have different attack time and sound, different volume levels, different tone qualities and they are sometimes out of tune etc. Large string sample libraries may sound real to a certain level but unfortunately they sound like bad orchestra. Large library is also of course very demanding for computer.

What a string instrument should have:
Obviously good enough basic sound model to reproduce string tone (easier said than done of course).
Lot of controllability to have natural crescendos and diminuendos, staccato-marcato-legato, vibrato etc. Very important thing is also to have natural transitions from a note to another note, from one dynamic to another.
It would be also a marvelous thing have a chance to have one solo intrument or bigger section.

Any comments? Is there some model already available or is this perhaps an impossible mission?

Re: Stringteq

Ecaroh wrote:

Large string sample libraries may sound real to a certain level but unfortunately they sound like bad orchestra.

I quite disagree with this statement...

http://spitfire-webassets.s3.amazonaws....emo_AB.mp3


Spitfire sample libraries sound absolutely awesome, their Albion/BML range is recorded in one of the greatest halls in the world (AIR studios, where Hans Zimmer loves to record his scores), and the result is nothing short of amazing.


If that demo above sounds like a bad orchestra to you, well... All I can say is that modelled strings would not sound nearly as good. Case in point - check out Synful Orchestra, which is sample based but employs a variant of modeling (sadly, developer seems to be MIA and there haven't been any updates in years). It sounds so thin and lifeless, despite its apparent expressivity.

We're a long, long way from a properly modelled orchestra that can sound as lush as Spitfire libraries do.

Last edited by EvilDragon (06-03-2015 12:54)
Hard work and guts!

Re: Stringteq

Well IMHO the "key" (pun intended) is the controller... How could you give to a note played on a keyboard all the variety you can with a bow ???

Re: Stringteq

MIDI CCs is the usual suspect. That'd be a lot of foot controllers.


Then again, there's this: https://www.youtube.com/watch?v=kLkPvmr93K8

Hard work and guts!

Re: Stringteq

Wahooo !!

Re: Stringteq

On dit "wahooo" en français aussi?! Anyway, "wahooo" indeed !!!

Re: Stringteq

Ok guys,

I have to agree that some of those libraries sound pretty impressive but I fear situation is quite similar with sampled pianos. They sound realistic simply because they are real in sense that they are real instruments recorded in real space. Problem arises when you start making music with them. It's hard to make them sound exactly how you like as performer or composer. I mean many small imperfections which make them "alive" start to control you instead you controlling them. Maybe I have had worst string libraries (which btw sounded great on demos) cause I've needed too much time to make them sound right. I have been scracthing my head how to make good sounding crescendos or tweaking pianoroll to make attacks times even etc. IMO Pianoteq isn't the most realistic piano sound but it's one of the most playable and flexible. With strings I could give up little of realism if I could get easier and more flexible strings. Quite often you would like to have strings "feeling" instead of ultimate realism, get kind of a perfect orchestra to play exactly how you want as a composer. And put your time and creative power to composing, not to endless tweaking of CCs, velocities, start times etc. And this is not to say that other times you may want to have realism which needs sample libraries.

Anyway many thanks for interesting tips! I am quite newbie here with string software...

P.S. I believe sample libraries perform best for percussive style (staccato, marcato) but have most trouble in long legato lines. Situation is quite equivalent with sampled vs. modeled piano sounds.

Last edited by Ecaroh (07-03-2015 15:59)

Re: Stringteq

Luc Henrion wrote:

Well IMHO the "key" (pun intended) is the controller... How could you give to a note played on a keyboard all the variety you can with a bow ???

Another example here, other style, other skill but very impressive!

https://www.youtube.com/watch?v=fyorUijach0

Re: Stringteq

I would love a modelled string program similar to pianoteq. Especially if you could layer and split with the piano. I am supposing (far from an expert) that some of the bowing effects could be accomplished by wheels, pedals, and ribbon controllers for those who need added realism.

Pianoteq, Ravenscroft 275

Re: Stringteq

Ecaroh wrote:

Problem arises when you start making music with them. It's hard to make them sound exactly how you like as performer or composer. I mean many small imperfections which make them "alive" start to control you instead you controlling them. Maybe I have had worst string libraries (which btw sounded great on demos) cause I've needed too much time to make them sound right.

This is exactly the beauty of Spitfire libraries - they sound ALIVE, there are imperfections in samples (because it was recorded with vintage microphones directly onto 2" tape!), and every instrument was recorded positioned as it should be, so you really don't need to work a lot at the mix stage to make the orchestra gel together.

I mean just go to the Spitfire Audio website and listen to those demos for Sable and Mural. Then tell me that it doesn't sound good. It sounds extremely good - and very plausible!

Hard work and guts!

Re: Stringteq

I hope that Modartt continue to develop these pianosounds(and other sounds they have) and get some new pianosounds, but dont split on to many different sounds. But, because of the string examples here, I attach my example of a cello. It is a short piece from my composition ”Dreaming” for cello, flute and saxophone. I made this before I found Pianoteq. Nowdays working with Ptq( and this years competition).

http://www.forum-pianoteq.com/uploads.p...aming_.mp3

Re: Stringteq

I see no harm in create tring essamble basic intrument for pinoteq, a simple thing, like they did with toy pianos, just to have option to tweak to many diferente things.

A string essamble like double bass, to cello, viola, and  up to violin, along keyrange.


The problem is that Modartt reached a level of perfection, quality and reputation, that nearly forbides them to create a nice simple thing for fun.

Re: Stringteq

EvilDragon wrote:

This is exactly the beauty of Spitfire libraries - they sound ALIVE, there are imperfections in samples (because it was recorded with vintage microphones directly onto 2" tape!), and every instrument was recorded positioned as it should be, so you really don't need to work a lot at the mix stage to make the orchestra gel together.

I mean just go to the Spitfire Audio website and listen to those demos for Sable and Mural. Then tell me that it doesn't sound good. It sounds extremely good - and very plausible!

You're absolutely right, they sound amazing. Maybe best strings out there.

But in case I wasn't clear enough, I will one more time clarify my main point: because of those imperfections those libraries can be difficult or time consuming to use. This does not mean that you cannot make them sound good or great. I will give you an example. With my string library (which is Cinestrings Core btw) even 1 bar phrase with 16th notes can be hard to make it "groove". Some notes have different attack times than others and it's difficult to say what is the one you should put earlier or later. And if some notes are little out of tune there's nothing you can do it fix it. I am not saying that I am searching for perfect music - in perfect/quantized time or in perfect tune - quite contrary, but I would like to be in charge to decide when something is out of rhythm or out of tune and how much. With too much "imperfection" libraries can be annoying to use.

Maybe modeling would give us at least an alternative to them. As I said, in my opinion situation is quite same as it is with pianos...

Re: Stringteq

EvilDragon wrote:

MIDI CCs is the usual suspect. That'd be a lot of foot controllers.


Then again, there's this: https://www.youtube.com/watch?v=kLkPvmr93K8


Wow. That's the best implementation of the Leap Motion that I've seen. Thanks for posting that.

Re: Stringteq

The best "strings" I've ever heard were the sounds of the old analog synths back in the day, and I think I liked it more when they used some chorus effect. A good example is "Everyone's a Winner" (Hot Chocolate). Such a "wet" rich sound - better than the real thing. 

Greg.

Re: Stringteq

Here are various modelled instrument VSTs:

https://www.applied-acoustics.com/modeling-collection/
http://www.wallanderinstruments.com/?mo...mp;lang=en
http://www.samplemodeling.com/en/products.php

I claim no experience with any other than listening to a few snippets on Soundcloud.

Last edited by Mossy (08-03-2015 13:08)

Re: Stringteq

They are very different instruments. So far Pianoteq developed percussion instruments and strings belong to another family which would probably require new mathematical models from scratch.

Last edited by Chopin87 (09-03-2015 01:13)
"And live to be the show and gaze o' the time."  (William Shakespeare)