Music Player Network Home Guitar Player Magazine Keyboard Magazine Bass Player Magazine EQ Magazine
Page 2 of 3 < 1 2 3 >
Topic Options
#1984209 - 08/26/08 04:31 AM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: orangefunk]
Tusker Online   content
MP Hall of Fame Member

Registered: 08/02/00
Posts: 5997
Loc: Saddle Ridge
I agree the ES effects are very helpful to the PLG. I use the pitch change algorithm with very subtle detuning to accomplish a smoothed/enlarged sound without adding artifacts.

For what it's worth, this video (with Sancious firing up the PLG VL at about 3.08) was one of the reasons I started working with the PLG/ES combination ...

http://www.youtube.com/watch?v=RByaN8w5mso

For (non-imitative) synthesis applications like these .... it still works remarkably well.
_________________________
"Musical ideas are prisoners, more than one might believe, of musical devices." Pierre Schaeffer

Top
KC Island
#1984352 - 08/26/08 11:05 AM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Tusker]
Mark Schmieder Offline
MP Hall of Fame Member

Registered: 09/08/05
Posts: 6646
Loc: San Francisco Bay Area
I can't watch YouTube videos, but I'm going to guess you are referring to Dave Sancious, one of my favourite players?
_________________________
Eugenio Upright, Dingwall ABI, Yamaha BB1025X, 50th T-Bass, 60th P-Bass, Select J-Bass
Yamaha WX5, Hammond 44 Melodion & XK1c, Moog Voyager, DSI Prophet12 & Evolver

Top
#1984371 - 08/26/08 11:40 AM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Tusker]
JohnH Offline
Platinum Member

Registered: 05/05/06
Posts: 1061
Loc: Whittier Ca, Los Angeles
 Originally Posted By: Tusker
 Originally Posted By: analogholic
But in "theory", could I progam those sounds with the VL-plugin in my Motif ES?


I There are some recordings made in the 80's with the original VL1 (which had more resolution). Some of these recordings hold up pretty well, even today. Jerry


I don't know how that would be possible, since the VL-1 came out in 1994.


http://www.sonicstate.com/synth/yamaha_vl1.cfm

John

Top
#1984383 - 08/26/08 12:12 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: JohnH]
Mark Schmieder Offline
MP Hall of Fame Member

Registered: 09/08/05
Posts: 6646
Loc: San Francisco Bay Area
While we're at it (that is, disspelling false notions), it's worth bringing up that someone else on another forum insists that the VL technology is FM-based. I argued until I was blue in the face, then gave up. Whoever shouts the loudest and longest wins arguments, right? :-)
_________________________
Eugenio Upright, Dingwall ABI, Yamaha BB1025X, 50th T-Bass, 60th P-Bass, Select J-Bass
Yamaha WX5, Hammond 44 Melodion & XK1c, Moog Voyager, DSI Prophet12 & Evolver

Top
#1984402 - 08/26/08 01:02 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: JohnH]
Tusker Online   content
MP Hall of Fame Member

Registered: 08/02/00
Posts: 5997
Loc: Saddle Ridge
 Originally Posted By: JohnH
I don't know how that would be possible, since the VL-1 came out in 1994.




The Deep Forest recordings I was referring to were made in the mid 90's.

And yes, it's a David Sancious video. According to SOS, there is some two operator FM in one of the modifiers that is downstream of the model ...

http://www.soundonsound.com/sos/1994_articles/jul94/yamahavl1.html

although the basic instrument model is ... a model.

Jerry
_________________________
"Musical ideas are prisoners, more than one might believe, of musical devices." Pierre Schaeffer

Top
#1984422 - 08/26/08 02:17 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Tusker]
Yoozer Offline
Gold Member

Registered: 07/12/04
Posts: 728
Loc: The Netherlands
This is what Reaktor can do, by the way (excuse the sloppy playing) using the Steampipe (ex. 1 and 2 and Silverwood (ex. 3 and 4, substituting the breath-controller with the modwheel) ensembles:

http://theheartcore.com/music/reaktorphys.mp3

Top
#1984439 - 08/26/08 03:16 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Yoozer]
Mark Schmieder Offline
MP Hall of Fame Member

Registered: 09/08/05
Posts: 6646
Loc: San Francisco Bay Area
Tusker, thanks for clearing that up.
_________________________
Eugenio Upright, Dingwall ABI, Yamaha BB1025X, 50th T-Bass, 60th P-Bass, Select J-Bass
Yamaha WX5, Hammond 44 Melodion & XK1c, Moog Voyager, DSI Prophet12 & Evolver

Top
#1984508 - 08/26/08 11:38 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: analogholic]
DrSynth Offline
Senior Member

Registered: 11/03/06
Posts: 40
Loc: Woodland Hills, CA
Yes, it's still the best ...

The point of the VL is that it's musicality is still unsurpassed. 100% timbral accuracy is not what it was/is about, it's playability. If you're into BC and especially wind controllers it is still the king of expressive solo sounds.

None of the modern 'modeling' sample slice & dicers are as fluid or organic. There's a big difference between what the VL does and analyzing note phrasing and some controller inputs and deciding what would instrument X do in that situation and spitting out a particular sample of instrument X doing that phrase and articulation.

I've always felt the best explanation of how to approach the VL is that you're synthesizing behaviors, and timbres are a side effect. All the cool stuff is 'in the cracks' as it were, and the VL lets you go continuously in, out and through them. And if you're lucky, you'll like the timbre too!

About some other details noted on the thread -- Yes the VL7 is absolutely identical in the synthesis engine, just a single element/voice instead of the 2 in the VL1. The main reason for most 2 element voices was convenience. The VL70m is a reduced subset of the model because of less DSP power, but it's core playability is nearly identical, and since so much more had been learned about programming the model, some VL70m voices are actually 'better' -- as noted Patchman's Turbo chip took it even further. The VL expert editors are Mac OS 7 & 8, with a Windows 95 reverse engineered version originated by a bunch of dedicated VL users. The Version 2 software added some characteristics specifically for brass instruments. There is no "FM" in the VL, though not to say some portion of the VL 'equation' may contain formulas that are mathematically similar.

The VP1 story came down to one thing -- price. It was 8x the hardware of the VL1, and priced accordingly. Also, its true power only shines with sequencer aided performance, as all those sweet controllers can't be manipulated while you're playing with both hands! An 8 voice version was proposed some time after the VL70m came out, in a simple case without all the cool controllers to hit a $2000 price, but by then it was obvious the controllers were what made the VP in the first place, not the model (essentially plucked string). Polyphony is a weird thing in the modeling idiom, as playing say 5 notes at a time ends up being more akin to one person playing, for example, 5 identical violins with the same bow, as opposed to 5 people playing 5 different violins. Which is in turn completely different when you talk about polyphony in regards to behaviors of a guitar or piano etc. What was so cool with the VP is things like changing the density, tension and inharmonicity of a guitar's string in real time while plucking it with the side of your finger or the edge of a brick.

Manny

Top
#1984744 - 08/27/08 04:08 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Mark Schmieder]
Dan Phillips Offline
Senior Member

Registered: 02/12/05
Posts: 92
Loc: California, USA
I've enjoyed reading Mark's detailed comments! A few historic and technical notes:

 Originally Posted By: Mark Schmieder
The modeling algorithms are from the 70's and earlier. Korg uses them too, but took a different approach. They are Public Domain, and very little has been done with them since then.


I've worked at Korg R&D since 1990, a bit before the original OASYS project began, so I've been involved in physical modeling for a while. Many of the basic concepts for delay-based physical modeling (such as used in the VL series, and some of the algorithms in the original OASYS and OASYS PCI) were developed at Stanford in the mid to late 1980's by researchers such Julius Smith and Perry Cook (see this article from Wired, hosted at Harmony Central. Many other researchers contributed as well (for instance, see www.acoustics.hut.fi/research/asp/aguitar/ ). My group did a good amount of original work which built on these concepts, and Yamaha did a lot as well (with many patents to show for it). The bulk of the Stanford research is not public domain; instead, it is administered as part of the Sondius XG patent portfolio.


Edited by Dan Phillips (08/27/08 04:09 PM)
_________________________
Dan Phillips
Product Manager, Korg R&D

Top
#1984745 - 08/27/08 04:15 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: DrSynth]
Mark Schmieder Offline
MP Hall of Fame Member

Registered: 09/08/05
Posts: 6646
Loc: San Francisco Bay Area
Well, as of last night, I am fully decided on parting with my VL70m/Turbo, which has done me well for probably around 10 years (maybe a bit less). The last thing to go were the saxophones, and I am now getting good phrasing and articulation from other sources. I really never expected to see the day when the VL70m would be surpassed by software VI's, but 2008 is the year!

Wallander has just announced version 2 of his hybrid instruments, and I plan to download and try the demos this weekend. He has added significant new support for expression ranges and wind controller data.

Garritan is said to have done a good job of wind controller support and sample switching with his Jazz and Big Band library, which at only $179 is tempting -- especially as it covers a lot of rare instruments (such as subcontrabass saxophone and sopranino saxophone).

Most impressive, though, is Sample Modeling's "The Trumpet", which sounds completely organic and fluid to me. I do hope they release more instruments soon. This is by far the best VI I have ever heard -- even more impressive than Modartt Pianoteq!

I have just requested a demo of the Extended License for my Vienna Instruments Saxophones library today, to see if the additional articulations (many of which involve smart-switching) work well in certain contexts vs. the VL70m/Turbo. last night I retracked some tenor sax parts using VSL, and in the context of the mix, it sounds better, but misses some of the expressivity.

The expressivity of the VL70m (if not also the VL1) is slightly overrated though, as it is more the fluidity between notes that they excel at, and not the proper note characteristic at each pitch and volume. The models just aren't quite sophisticated enough to "nail" all those details, but Sample Modeling has truly nailed every detail so that the timbre is accurate as well as the expressivity.

It's too bad most of those VL1m auctions on eBay were fake or scams. I would have loved to have had a chance to try one and see how much better it is than the VL70m/Turbo. But I figure if the technology is that good, it will be revived in some form at some time -- maybe by Yamaha, but maybe by others (the algorithms themselves, which are waveguide-based, are mostly Public Domain, and the majority come from Stanford's CCRMA Lab, run by Julian Smith -- with similar background to FM/DX7).
_________________________
Eugenio Upright, Dingwall ABI, Yamaha BB1025X, 50th T-Bass, 60th P-Bass, Select J-Bass
Yamaha WX5, Hammond 44 Melodion & XK1c, Moog Voyager, DSI Prophet12 & Evolver

Top
#1984747 - 08/27/08 04:19 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Mark Schmieder]
Mark Schmieder Offline
MP Hall of Fame Member

Registered: 09/08/05
Posts: 6646
Loc: San Francisco Bay Area
Hey, Dan, our posts collided and contain some of the same information :-).

We've met before, at NAMM, AES, and OASYS clinics at Leo's Pro Audio. I didn't know you occasionally lurk on this forum. You also work with my predecessor in my originals band, Moonlife.

Cool to have you around. I do hope the OASYS spins off some more products at some point -- especially in the area of strings, which are still underserved in the realm of hardware devices.
_________________________
Eugenio Upright, Dingwall ABI, Yamaha BB1025X, 50th T-Bass, 60th P-Bass, Select J-Bass
Yamaha WX5, Hammond 44 Melodion & XK1c, Moog Voyager, DSI Prophet12 & Evolver

Top
#1984748 - 08/27/08 04:21 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Mark Schmieder]
Mark Schmieder Offline
MP Hall of Fame Member

Registered: 09/08/05
Posts: 6646
Loc: San Francisco Bay Area
Of course, Dan's post is more accurate than my own, in terms of the dates as well as the ratio of Public Domain to patented and proprietary. As others here may not know as much about Dan as I do (well, the Keyboard editors and forum moderators obviously do), I just want to make sure others realise his word is more authoritative than my own :-).
_________________________
Eugenio Upright, Dingwall ABI, Yamaha BB1025X, 50th T-Bass, 60th P-Bass, Select J-Bass
Yamaha WX5, Hammond 44 Melodion & XK1c, Moog Voyager, DSI Prophet12 & Evolver

Top
#1984776 - 08/27/08 07:38 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Mark Schmieder]
DrSynth Offline
Senior Member

Registered: 11/03/06
Posts: 40
Loc: Woodland Hills, CA
 Originally Posted By: Mark Schmieder


The expressivity of the VL70m (if not also the VL1) is slightly overrated though, as it is more the fluidity between notes that they excel at, and not the proper note characteristic at each pitch and volume. The models just aren't quite sophisticated enough to "nail" all those details, but Sample Modeling has truly nailed every detail so that the timbre is accurate as well as the expressivity.


I have to disagree -- the VL model is quite capable. My belief is any particular details that are not "nailed" in a particular emulative patch is due to a lack of programming effort. VL voicing development stopped over 10 years ago, sad to say. There's bits and pieces of old voicing data that do exquisite realistic things -- for half an octave. With probably a month of dedicated effort, one might get it to two and a half octaves. I think Dan Phillips will agree that when you apply current knowledge to old hard/software it's amazing what you can get out of the old stuff.

The industry moved on from true modeling synthesis as there were more expedient ways to get results that appealed to more users. Which is a shame. Some, like the VLTurbo team, put in a lot of effort and made great strides, and as you noted met your needs for 10 years. Some individual users made significant progress as well, such as Bill "Burning" Busch's brass and sax work.

Sample modelling will only be able to "nail" whatever details they decide to sample. They can't do "every detail" of an instrument. They may do nearly all of one particular player's repertoire on a particular instrument. Very nice stuff, but still a one trick pony, light years ahead of the original EMu Emulator, but still just playing back gussied up recordings.

There are so many more sonic resources available in the timbral and behavior possibilities in a true synthesizer like the VL, which I think was original poster analogholics' question.

But it always comes down to what are your needs.

Guess you can tell I'm not a sample guy...

Manny


Edited by DrSynth (08/27/08 08:52 PM)
Edit Reason: typos

Top
#1984806 - 08/27/08 09:35 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: DrSynth]
Tusker Online   content
MP Hall of Fame Member

Registered: 08/02/00
Posts: 5997
Loc: Saddle Ridge
I imagine there is a continuum of people with slightly different values making these comparisons:

- "Orchestral on record" types - People who need to replicate authentic wind and string sounds with a high degree of fidelity. Playability is not an immediate concern, as long as acoustic instrument expressions can be replicated through automation or performance.

- "Synth for playing" types - People who wish to have playable synth sounds that may or may not be authentic. Expressiveness is more important than fidelity.

I imagine there are number of people with values that fall between these extremes. I admit I lean toward the "synth for playing" camp, hence my choices.

Jerry
_________________________
"Musical ideas are prisoners, more than one might believe, of musical devices." Pierre Schaeffer

Top
#1984829 - 08/28/08 03:17 AM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Tusker]
Mark Schmieder Offline
MP Hall of Fame Member

Registered: 09/08/05
Posts: 6646
Loc: San Francisco Bay Area
I would agree when it comes to FM synthesis, which sadly is associated with cheesy "acoustic" pianos and the like.

Many new technologies initially get applied to emulating real world instruments, as that's what most people want. So when the technology progresses, or at least changes, people often neglect to check out what that older technology was really capable of, when not saddled with the task of sounding like an acoustic piano :-).

The problem with the VL technology is that few people have access to the expert editors, which were pulled from Yamaha's site and other support sites eons ago. The simple editor, or whatever it's called (I forget), doesn't expose many parameters, and also doesn't run on Mac OS X (and maybe not on Windows XP either). So most of us are stuck with presets and other people's programmed voice libraries.


Edited by Mark Schmieder (08/28/08 03:17 AM)
Edit Reason: typo
_________________________
Eugenio Upright, Dingwall ABI, Yamaha BB1025X, 50th T-Bass, 60th P-Bass, Select J-Bass
Yamaha WX5, Hammond 44 Melodion & XK1c, Moog Voyager, DSI Prophet12 & Evolver

Top
#1984830 - 08/28/08 03:22 AM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Mark Schmieder]
Mark Schmieder Offline
MP Hall of Fame Member

Registered: 09/08/05
Posts: 6646
Loc: San Francisco Bay Area
Oh, the funny thing is, I wasn't a "sample guy" either -- until recently. I am at this point agnostic about technology. After all, I adopted the physically modeled Modartt Pianoteq, and went the other direction (towards samples -- albeit modified with some layers of modeling), for many other instruments.

And yes, the VL technology can be used in interesting ways for creating new sounds, vs. emulative sounds. If I had an editor I could use, maybe I'd be tempted. But I'd rather just wait until such stuff shows up in the software world, where I can edit to my heart's content.

I can't remember whether the original VL1 and VL7 had a significantly different editor from the VL70m and PLG150-VL or not. I think the architectures are significantly different in terms of the hardware, in spite of a lot of shared technology.
_________________________
Eugenio Upright, Dingwall ABI, Yamaha BB1025X, 50th T-Bass, 60th P-Bass, Select J-Bass
Yamaha WX5, Hammond 44 Melodion & XK1c, Moog Voyager, DSI Prophet12 & Evolver

Top
#1984831 - 08/28/08 03:50 AM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Mark Schmieder]
Tusker Online   content
MP Hall of Fame Member

Registered: 08/02/00
Posts: 5997
Loc: Saddle Ridge
Mark you have hit upon one of my frustrations with VL. I wish they had spent more energy on a user friendly editing environment.

Still, I rather find that even with the limited editing capability, the VL technology makes for a very good "oscillator" for creating new sounds. It's not the entire signal chain.

I sometimes take one of the more inharmonic models, like say Jurrasic and run it through a controllable lpf and a delay line. It's a whole sound concerto right there. Or go nuts and use a ring modulator and bit reducer on it. The trick works when I combine the VL with regular synthesis control of brightness, pitch, volume and wetness. It's a very good initial seed into a modular synthesis environment of your choice.

Which takes me to steampipe ...

Yoozer, you've almost sold me on Reaktor with that demo. Chet's model seems very appropriate for traditional music ensembles, but steampipe combined with the regular reaktor modules sounds like an amazing playground. I'll look into it. Thanks,

Jerry
_________________________
"Musical ideas are prisoners, more than one might believe, of musical devices." Pierre Schaeffer

Top
#1984954 - 08/28/08 12:42 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Tusker]
Mark Schmieder Offline
MP Hall of Fame Member

Registered: 09/08/05
Posts: 6646
Loc: San Francisco Bay Area
Some people say they've been successful with VL-style modeling in Nord Modular as well, but I'm skeptical. OTOH, maybe it's true, based on Clavia's work on the Electro and Stage series.
_________________________
Eugenio Upright, Dingwall ABI, Yamaha BB1025X, 50th T-Bass, 60th P-Bass, Select J-Bass
Yamaha WX5, Hammond 44 Melodion & XK1c, Moog Voyager, DSI Prophet12 & Evolver

Top
#1984981 - 08/28/08 02:35 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Mark Schmieder]
GovernorSilver Offline
MP Hall of Fame Member

Registered: 03/22/06
Posts: 4377
Loc: Washington DC
 Originally Posted By: Mark Schmieder
Some people say they've been successful with VL-style modeling in Nord Modular as well, but I'm skeptical. OTOH, maybe it's true, based on Clavia's work on the Electro and Stage series.


"Physical modeling" on the Nord Modular has been done by applying the principles taught by Ron Berry, and possibly others, of what they call "Acoustic Modeling":

http://www.soundonsound.com/sos/1997_articles/jul97/ronberry.html

So you could use a modular synth (real analog or virtual) to implement "physical modeling". Chet Singer and K. Eckhardt (sp?) have examples posted various places online - eg. the MOTM website. The Nord Modular G2 went a step further by adding a couple of physical modeling oscillator modules. But the hard core users still complain about the delay resolution being only 16-bit.

Edmund Eagan had some excellent demos utilizing the Kyma and other tools and a Haken Audio Continuum controller. I use past tense because Firefox 3 blocks the required plugins for listening/viewing the examples.


Edited by GovernorSilver (08/28/08 02:37 PM)
_________________________
https://soundcloud.com/governorsilver

Top
#1985013 - 08/28/08 05:03 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Mark Schmieder]
Dan Phillips Offline
Senior Member

Registered: 02/12/05
Posts: 92
Loc: California, USA
 Originally Posted By: Mark Schmieder
Hey, Dan, our posts collided and contain some of the same information :-).

We've met before, at NAMM, AES, and OASYS clinics at Leo's Pro Audio. I didn't know you occasionally lurk on this forum. You also work with my predecessor in my originals band, Moonlife.


Hi Mark! I'll say hi to Bill for you.

- Dan
_________________________
Dan Phillips
Product Manager, Korg R&D

Top
#1985059 - 08/28/08 10:04 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Tusker]
DrSynth Offline
Senior Member

Registered: 11/03/06
Posts: 40
Loc: Woodland Hills, CA
 Originally Posted By: Tusker
Mark you have hit upon one of my frustrations with VL. I wish they had spent more energy on a user friendly editing environment.


 Originally Posted By: Mark Schmieder
And yes, the VL technology can be used in interesting ways for creating new sounds, vs. emulative sounds. If I had an editor I could use, maybe I'd be tempted. But I'd rather just wait until such stuff shows up in the software world, where I can edit to my heart's content.

I can't remember whether the original VL1 and VL7 had a significantly different editor from the VL70m and PLG150-VL or not. I think the architectures are significantly different in terms of the hardware, in spite of a lot of shared technology.


Points very well taken. Coming up with the programming interfaces are a curious thing, especially for technologies that don't represent the typical time/frequency domain manner that we experience sounds. VL modeling, like FM, are abstractions that don't relate in a 1:1 manner to the resulting sound like the more familiar subtractive synthesis, or even additive / FFT stuff. Unfortunately, it's just the way it is -- it's just different, and you just have to commit to putting in the time until it starts to make sense. The 'friendly' visual and analog editor versions are horrible compromises to focus on the aspects of the model to which results are predicable and most would be familiar with. I don't think they touch 10% of the actual core model structure.

The only difference in the expert VL1/7 editors and the VL70m/PLG's is the 'stripped down' nature of the VL70m/PLG -- IIRC the VL1 ran the model and effects on 3 of the custom VLSI chips and the VL70m/PLG on just 1. So, some parts of the modeling algorithm were left out, or recoded to use less DSP on the VL70m. And it barely made it within a single chip so the effects are generated by a separate SPX90-II chip, which is why it's effects are not up to snuff compared to the VL1.

BTW the editors are still available around the 'net as Yamaha allowed a number of places to host them with the editing tutorials. Anyone can edit to their hearts content as long as they don't mind the risk of a coronary! You can pick up a late NUBUS or early PCI Mac on eBay for usually the price of shipping.

There were long discussions way back when about releasing the expert editor in the first place, specifically because Yamaha was concerned few would understand it. And patent concerns too -- the released version changed many of the parameter names from the correct mathematical terms to things resembling traditional acoustic terms to obfuscate the model. Which sucked for anyone who learned the original terms and had to do it all a second time.

Manny

Top
#1985060 - 08/28/08 10:15 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: DrSynth]
Mark Schmieder Offline
MP Hall of Fame Member

Registered: 09/08/05
Posts: 6646
Loc: San Francisco Bay Area
Very interesting. Thanks for taking the time to post all of that. It also points out more why the VL70m editing isn't as full-featured.
_________________________
Eugenio Upright, Dingwall ABI, Yamaha BB1025X, 50th T-Bass, 60th P-Bass, Select J-Bass
Yamaha WX5, Hammond 44 Melodion & XK1c, Moog Voyager, DSI Prophet12 & Evolver

Top
#1985100 - 08/29/08 05:58 AM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Mark Schmieder]
Tusker Online   content
MP Hall of Fame Member

Registered: 08/02/00
Posts: 5997
Loc: Saddle Ridge
The minimoog, the prophet 5, even the dx7 ... were all market successes because they made some previously difficult thing simple. The mini codified the simple subtractive model, the prophet made patch changes possible, the dx7 made struck and plucked sounds easy to use.

Lower end physical modeling (pm) interfaces and languages are aplenty at this time. (lower end = closer to the machine) The problem of good organic sound is largely solved. The issue is the physical and intellectual interface.

If there is a market opportunity for Yamaha it's in re-introducing pm with a higher end intuitive interface. Something that provides good pm with near-rompler simplicity. The V-synth has not cracked that problem yet (imo) and no other synth has taken aim at it.

Jerry
_________________________
"Musical ideas are prisoners, more than one might believe, of musical devices." Pierre Schaeffer

Top
#1985269 - 08/29/08 12:22 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Tusker]
Mark Schmieder Offline
MP Hall of Fame Member

Registered: 09/08/05
Posts: 6646
Loc: San Francisco Bay Area
Cameleon 5000 in my view has done exactly that for Additive Synthesis. A different topic, for sure, but worth noting as it shows that an older hard-to-program synth architecture (in this case, the Kawai K5000) can be emulated and given a new front-end (with some complimentary new features thrown in) that give it a new lease on life.

FM8 does this for FM as well. FM7 was more pure, but also harder to program than FM8 as a result. I complained at first that FM8 doesn't do a good job with DX7 patches, whereas FM7 did. But I've moved past that, as I no longer use vintage DX7 patches and am appreciating FM8 for what it is... which was the real intent.

That is, it is opening up the ease of using FM to create new sounds, vs. focusing on how precisely it matches the sonic character of older DX7 patches (many of which were rudimentary emulations of acoustic or electro-acoustic instruments anyway).

The reasons FM8 is not sonically backward compatible with FM7 are documented in an older thread on this forum, and it's not particularly of interest to this current topic.


Edited by Mark Schmieder (08/29/08 12:23 PM)
Edit Reason: typos
_________________________
Eugenio Upright, Dingwall ABI, Yamaha BB1025X, 50th T-Bass, 60th P-Bass, Select J-Bass
Yamaha WX5, Hammond 44 Melodion & XK1c, Moog Voyager, DSI Prophet12 & Evolver

Top
#1985347 - 08/29/08 05:10 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Mark Schmieder]
DrSynth Offline
Senior Member

Registered: 11/03/06
Posts: 40
Loc: Woodland Hills, CA
I think there's an old quote from Wendy Carlos back in her GDS Synergy days, along the lines of '.. the good news is you can control every aspect of every harmonic in a sound. The bad news is you have to."

So, sometimes capabilities come at a cost. This was a mantra repeated often during the VL voicing sessions.

Funny you mention the K5000. Loved that synth, very underated.

Manny

Top
#1985477 - 08/30/08 07:09 AM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: DrSynth]
kfffish Offline
Member

Registered: 05/28/07
Posts: 9
I find myself drawn to this form of synthesis. I am trying to get better at programming my own sounds for my vl/plg. Although the expert editor is daunting there are sections of it ( like the filter ) that can be used just like subtractive synthesis. By using the visual and analog editors as starting points I'm able to port those sounds into the expert editor and tweak further. Where I get stuck is trying to program the smoothness in my sounds when using the breath controller. BTW Manny your tutorial on the expert editor is fantastic.

Top
#1985588 - 08/30/08 02:25 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: kfffish]
Mark Schmieder Offline
MP Hall of Fame Member

Registered: 09/08/05
Posts: 6646
Loc: San Francisco Bay Area
Here's an interesting comparison (via scientific charts) of the WX-5 vs. the Akai EWI 4000s:

http://www-acad.sheridanc.on.ca/~degazio/AboutMeFolder/MusicPages/EWI-WX5.html

I still suspect the Akai takes more effort to produce the sorts of expression that Yamaha achieves via a typical saxophone styled mouthpiece. The majority of people on the sax forum prefer the WX-5 to the 4000s:

http://www.saxontheweb.net/vbulletin/showthread.php?t=82416

Yet, here another top user finds the EWI 4000s more to his liking, after kmuch experience with both units:

http://derekwyatt.org/ewi_and_wx/ewi_and_wx.html
http://derekwyatt.org/blog/files/ewi_4ks_review.html

Norton Music has posted some tweaks for the Turbo ROM, as Matt Traum is an EVI player (not even an EWI player, much less a WX player) and optimised the VL70m replacement ROM using the EVI:

http://www.nortonmusic.com/vl70-m_turbo.html

His General Tips have also been updated since I last checked in 2006:

http://www.nortonmusic.com/wx5.html
_________________________
Eugenio Upright, Dingwall ABI, Yamaha BB1025X, 50th T-Bass, 60th P-Bass, Select J-Bass
Yamaha WX5, Hammond 44 Melodion & XK1c, Moog Voyager, DSI Prophet12 & Evolver

Top
#1985646 - 08/30/08 08:02 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Tusker]
Dan Phillips Offline
Senior Member

Registered: 02/12/05
Posts: 92
Loc: California, USA
 Originally Posted By: Tusker

If there is a market opportunity for Yamaha it's in re-introducing pm with a higher end intuitive interface. Something that provides good pm with near-rompler simplicity. The V-synth has not cracked that problem yet (imo) and no other synth has taken aim at it.


V-synth uses sample playback, something which I believe is additive resynthesis ("elastic audio"), and VA. An interesting synth to be sure, but not related to physical modeling.
_________________________
Dan Phillips
Product Manager, Korg R&D

Top
#1985654 - 08/30/08 09:01 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Dan Phillips]
Tusker Online   content
MP Hall of Fame Member

Registered: 08/02/00
Posts: 5997
Loc: Saddle Ridge
Dan,

My bad, I should have mentioned the "GT". The original V-synth coupled the elastic audio concept with a VA and a couple of other features. The newer V-Synth GT adds "AP synthesis", which is I what I am referring to.

Here is an excerpt from the Keyboard Magazine review describing AP synthesis ... unique to the V-synth GT ... which involves associating samples with "models":

AP mimics the way some acoustic instruments respond to a player’s technique, by shaping the attack, transitions between notes, tuning, and other qualities as you play.

Four of the five models for doing this are violin, erhu, sax, and flute. The flute model, for example, adjusts breath noise, slurring, multiple vibrato settings, and even slight detuning based on how staccato or legato you play. Hit a key hard, and the pitch will detune as though a player were “overblowing” a real flute. The models are just as good at simulating other bowed, blown, and reed instruments, and even better at creating yet-unheard hybrid sounds where waveforms that you want to sound synthetic also adopt some of the phrasing and articulation qualities of acoustic instruments. In fact, this is the specialty of the fifth model, which combines elements of the other four.

The models themselves don’t generate sound — they act upon a special set of sampled waveforms the GT uses exclusively for AP synthesis.


Sound on Sound describes it in this fashion:

Articulative Phrase Synthesis is the component of the V-Synth GT that is completely new. Described as neither sampling nor modelling as we currently understand it, APS simulates the way an instrument responds to human playing techniques.

I am not really sure how to describe this.

The technique reminds me of the Roland SH32, a synth that appeared to utilize samples, yet was more than simple sample playback, as it offered both PWM and hard sync.

Jerry
_________________________
"Musical ideas are prisoners, more than one might believe, of musical devices." Pierre Schaeffer

Top
#1986207 - 09/01/08 12:04 PM Re: Is the Yamaha VL1 still the best physical modeling synth [Re: Tusker]
Dan Phillips Offline
Senior Member

Registered: 02/12/05
Posts: 92
Loc: California, USA
 Originally Posted By: Tusker
The newer V-Synth GT adds "AP synthesis", which is I what I am referring to.


Yes, I guessed that you were referring to the GT.

Before going further, I'd like to stress that I am not commenting on either the musical effectiveness of different synthesis techniques, or on those of the V-Synth GT in particular. Each synthesis method has its own strengths, and the V-Synth is widely acclaimed.

By my reading, Roland does not say that the V-Synth uses physical modeling. My guess is that the usage of the term "model" in the quoted magazine articles may be the source of confusion. In this case, I think they mean something like the following definition from Wikipedia:

"Model: an abstraction or conceptual object used in the creation of a predictive formula"

In this case, the model describes the sounds one expects from an instrument in response to the musician playing in a certain way. For instance, if you play harder, the instrument might become louder and/or brighter; if you play very loud, you might start to hear a different timbre come in, such as an overblown wind instrument; if you release a piano note quickly, the note-off sound might be louder than if the note was held for several seconds. These models can also take more than one note into account. For instance, if you play quickly on a guitar, there might be more chance of string squeaks.

This might be called "performance modeling." Note that this type of model doesn't necessarily imply anything about the way that the sounds are produced. The synth could use samples, FM, analog, physical modeling, additive, etc.; from the perspective of the model, it doesn't matter.

As far as I can tell, the V-Synth uses Roland's elastic audio (probably a form of additive resynthesis; according to the Keyboard summary, the Roland processes "act upon a special set of sampled waveforms") to do this sort of modeling, including the insertion of note transitions, such as the sound of a wind or brass instrument as it changes pitch. Synful Orchestra, using additive synthesis, does something similar.

You could think of this as a top-down process: the model describes what the overall result should be in response to a given musical input.

Physical modeling synthesis, in comparison, is a from-the-ground-up process. It starts with creating a model of a physical instrument, using mathematical equations to describe the various components, such as the strings, bridge, and body of a guitar, along with the device that makes the string move (such as a pick); the reed, body, and bell of a sax, along with the breath input to the reed; etc.

In the current state of the art, many shortcuts and approximations are made along the way; "cheap hacks," as physical modeling pioneer Perry Cook used to call them gleefully. That said, the basic idea is that the resulting sound comes naturally from the way that the different parts of the model work and interact. Used effectively, this can result in organic-sounding note transitions, among other things - a different means to a similar result.

- Dan
_________________________
Dan Phillips
Product Manager, Korg R&D

Top
Page 2 of 3 < 1 2 3 >


Moderator:  Dave Bryce, Stephen Fortner