Jump to content


Photo

Creating Models From Scratch

models user created

  • Please log in to reply
33 replies to this topic

#1 sudiptasingh

sudiptasingh

    Just Startin'

  • Members
  • Pip
  • 11 posts

Posted 03 September 2013 - 04:59 PM

Why not open up the HD model software so that users can create their own guitars, bodies, pickups... In short Workbench on steroids. I hear the griping about Strats 1.9 and 2.0 - why not model EXACTLY the sound you want ? :-) I understand if there is IP Line 6 wants to protect - but surely "open sourcing" the HD models would be a great thing! I am sure most users wouldn't want to go to the level of sampling guitar waveforms and tweaking fft parameters - but enough might!
  • 0

#2 phil_m

phil_m

    Uber Guru

  • Line 6 Expert
  • 5512 posts
  • LocationMinneapolis, MN

Posted 03 September 2013 - 05:09 PM

I guess I'm not sure of what you mean by the "from scratch" part... Are you envisioning a bigger library of guitar bodies and pickups that people can choose from?


  • 0
Time is a train
Makes the future the past
Leaves you standing in the station
Your face pressed up against the glass

 


#3 sudiptasingh

sudiptasingh

    Just Startin'

  • Members
  • Pip
  • 11 posts

Posted 03 September 2013 - 05:19 PM

Well .. Yes - apart from the 29 guitars modeled, and the library of "components" (bodies, pickups, pots) - why not open up the software so users can create ( and paint) completely new guitar types ( pedal steel anyone?), components, etc. As a flight and train simmer it's always really cool to watch people building new planes, scenery , locomotives - not just tweaking from a admittedly large palette.
  • 0

#4 clay-man

clay-man

    Gear Head

  • Members
  • PipPipPip
  • 871 posts

Posted 03 September 2013 - 05:59 PM

The presets, bodies pots and pickups are coded in the firmware. A preset doesn't carry the algorithm codes and all, it just tells the variax what settings you're using, as in it picks the bodies pickups and pot values and applies that to the coding inside the variax's firmware to give you the sound you get.

 

A preset is nothing more than values for variables with the firmware to work with.

 

What your asking would possibly require a mass rewrite and is way too complicated.

Work with what you got. I agree there should be more bodies and pickups perhaps, but painting a guitar? Making guitar bodies? That's impossible unless you know how to model guitars in their programming yourself.

 

I made an SG preset out of a junior LP body and LP pickups rolled back towards the bridge. It sounds pretty spot on. I also made a Baritone Jaguar preset. I'm pretty sure you can come close to stuff if you just try.


  • 0

For a minute there, I lost myself.

Radiohead_bear-728286%5B1%5D.png 


#5 sudiptasingh

sudiptasingh

    Just Startin'

  • Members
  • Pip
  • 11 posts

Posted 03 September 2013 - 06:52 PM

I understand the point you're making. Please just consider that there is a vast amount of talent out on this planet that is very comfortable with programming and customizing code (even firmware) to extend its capabilities. Just look at custom ROMs people have developed for mobile phones, game emulators and so on. Opening up the software to this community would enable Line6 to tap into this pool and get additional functionality to add to their core platform or even generate revenue for developers building new add-ons. Think Apple/ MS Flight Simulator or X-plane etc. Not to mention getting more beta testers and a rich and thriving ecosystem around their core product.
  • 0

#6 sudiptasingh

sudiptasingh

    Just Startin'

  • Members
  • Pip
  • 11 posts

Posted 03 September 2013 - 06:56 PM

Just sayin.... :-)
  • 0

#7 Charlie_Watt

Charlie_Watt

    Power User

  • Members
  • PipPipPipPip
  • 1161 posts

Posted 03 September 2013 - 07:00 PM

You're smoking rope if you think that's going to happen.


  • 1

#8 clay-man

clay-man

    Gear Head

  • Members
  • PipPipPip
  • 871 posts

Posted 03 September 2013 - 07:40 PM

Yeah, basically what you're asking is if you can code the firmware yourself. What you're asking for requires a lot of memory from the presets. Like I said, that's way too complicated to do.

 

Workbench is supposed to be what you're asking, just at a feasible scale. If you're not satisfied with your tone then you should use an EQer.

Other than that, more bodies and pickups, but asking to adjust every little nuance is insane.


  • 0

For a minute there, I lost myself.

Radiohead_bear-728286%5B1%5D.png 


#9 ozbadman

ozbadman

    Power User

  • Members
  • PipPipPipPip
  • 1689 posts

Posted 03 September 2013 - 09:38 PM

I understand the point you're making. Please just consider that there is a vast amount of talent out on this planet that is very comfortable with programming and customizing code (even firmware) to extend its capabilities. Just look at custom ROMs people have developed for mobile phones, game emulators and so on. Opening up the software to this community would enable Line6 to tap into this pool and get additional functionality to add to their core platform or even generate revenue for developers building new add-ons. Think Apple/ MS Flight Simulator or X-plane etc. Not to mention getting more beta testers and a rich and thriving ecosystem around their core product.

 

This is technically already possible.

 

A JTV is just a computer in a guitar body. Its code is downloaded from a PC through the interface cable. If someone were sufficiently interested, they could look at the chips on the JTV, disassemble the existing JTV code, and write their own JTV firmware. Line 6 is not stopping them.

 

But it's never going to happen. Beyond the programming, you have to learn the domain: how to model different guitar bodies, pickup positions, process the piezos, etc. Line 6 has had professionals working on it full time for over a decade, and they've done a remarkably good job.


  • 1

#10 clay-man

clay-man

    Gear Head

  • Members
  • PipPipPip
  • 871 posts

Posted 03 September 2013 - 10:35 PM

This is technically already possible.

 

A JTV is just a computer in a guitar body. Its code is downloaded from a PC through the interface cable. If someone were sufficiently interested, they could look at the chips on the JTV, disassemble the existing JTV code, and write their own JTV firmware. Line 6 is not stopping them.

 

But it's never going to happen. Beyond the programming, you have to learn the domain: how to model different guitar bodies, pickup positions, process the piezos, etc. Line 6 has had professionals working on it full time for over a decade, and they've done a remarkably good job.

 

I was going to say that you could hack the firmware, but that would be insane to mess with. You'd need to reverse engineer everything.

The problem with this is that you don't have the source code obviously. 

Hacking the firmware could possibly brick your JTV too. It's not really feasible unless you're that insane.


  • 0

For a minute there, I lost myself.

Radiohead_bear-728286%5B1%5D.png 


#11 ozbadman

ozbadman

    Power User

  • Members
  • PipPipPipPip
  • 1689 posts

Posted 03 September 2013 - 10:44 PM

I was going to say that you could hack the firmware, but that would be insane to mess with. You'd need to reverse engineer everything.

The problem with this is that you don't have the source code obviously. 

Hacking the firmware could possibly brick your JTV too. It's not really feasible unless you're that insane.

 

I couldn't agree more. I was just pointing out the level of complexity, and the fact that if anyone was crazy enough to want to do it, they already could without Line 6 changing anything. Line 6 didn't start with Source Code originally either.


  • 0

#12 clay-man

clay-man

    Gear Head

  • Members
  • PipPipPip
  • 871 posts

Posted 03 September 2013 - 11:40 PM

I couldn't agree more. I was just pointing out the level of complexity, and the fact that if anyone was crazy enough to want to do it, they already could without Line 6 changing anything. Line 6 didn't start with Source Code originally either.

 

Well the source code is the file that is editable to the programmers, not yet exported as the firmware file. That means it can still be open with the editing software and tweaked with it. Once you compile the source code it's not readable by the editing software, but is used to actually run as a process, in this case the Variax's DSP.

 

I'm sure they used something to write up the algorithms/convolutions to model the guitars and other stuff. It would be pretty hard and ridiculous to code in raw computing language (hexadecimal/binary)


  • 0

For a minute there, I lost myself.

Radiohead_bear-728286%5B1%5D.png 


#13 ozbadman

ozbadman

    Power User

  • Members
  • PipPipPipPip
  • 1689 posts

Posted 03 September 2013 - 11:50 PM

Well the source code is the file that is editable to the programmers, not yet exported as the firmware file. That means it can still be open with the editing software and tweaked with it. Once you compile the source code it's not readable by the editing software, but is used to actually run as a process, in this case the Variax's DSP.

 

I'm sure they used something to write up the algorithms/convolutions to model the guitars and other stuff. It would be pretty hard and ridiculous to code in raw computing language (hexadecimal/binary)

 

I think you missed my point. When Line 6 started this process, they started like the rest of us: 0 lines of Source Code (ignoring startup libraries, etc.). There's nothing stopping someone else from doing the same thing. (FWIW, I am a former embedded software engineer, working in assembler and higher level languages).


  • 0

#14 clay-man

clay-man

    Gear Head

  • Members
  • PipPipPip
  • 871 posts

Posted 03 September 2013 - 11:59 PM

I think you missed my point. When Line 6 started this process, they started like the rest of us: 0 lines of Source Code (ignoring startup libraries, etc.). There's nothing stopping someone else from doing the same thing. (FWIW, I am a former embedded software engineer, working in assembler and higher level languages).

 

You mean a custom firmware from scratch?


  • 0

For a minute there, I lost myself.

Radiohead_bear-728286%5B1%5D.png 


#15 ozbadman

ozbadman

    Power User

  • Members
  • PipPipPipPip
  • 1689 posts

Posted 04 September 2013 - 12:01 AM

You mean a custom firmware from scratch?

 

I'm saying that's how Line 6 started, so yeah, if someone wanted to replicate what Line 6 had done, that's what they could do.


  • 0

#16 clay-man

clay-man

    Gear Head

  • Members
  • PipPipPip
  • 871 posts

Posted 04 September 2013 - 01:29 AM

I'm saying that's how Line 6 started, so yeah, if someone wanted to replicate what Line 6 had done, that's what they could do.

 

Well, I know, but you'd have to know how the hardware works and all, something Line 6 knows because they're the one that made the guitar. 

 

This person is asking for additional stuff though, and making new firmware from scratch would throw out all the features of the real firmware out the window.


  • 0

For a minute there, I lost myself.

Radiohead_bear-728286%5B1%5D.png 


#17 ozbadman

ozbadman

    Power User

  • Members
  • PipPipPipPip
  • 1689 posts

Posted 04 September 2013 - 01:48 AM

Well, I know, but you'd have to know how the hardware works and all, something Line 6 knows because they're the one that made the guitar. 

 

This person is asking for additional stuff though, and making new firmware from scratch would throw out all the features of the real firmware out the window.

 

That's my point. You can't just "open up the modelling" like is being asked for. You could open up certain parameters, but that's effectively what they're doing by providing different bodies, pickups, angles, positions, etc. Even adding an api would require the end-user to understand the problem domain to a much higher level than your average guitarist wants to. You could possibly add customisation of elements like pickup inductance, resistance, body resonances, etc., but that really depends on how their current algorithms work. If they're not component based, all that goes out the window.

 

The point is, they already provide the ability to vary the things they think are useful to an end-user. They could possibly provide other things, but what they are is algorithm specific and only Line 6 knows what they would be. Beyond that, everything rapidly gets much harder, and if someone really wants to do it, they could do so right now. But, we seem to be going around in circles somewhat.

 

Me, I'm happy they provide the ability to cutomize the tones at all. Pretty amazing really.


  • 1

#18 jpmull

jpmull

    Just Startin'

  • Members
  • Pip
  • 40 posts

Posted 04 September 2013 - 03:38 AM

Maybe they could add functionality similar to the Kemper Profiling amp?

In my opinion, I have more than enough options to play with, given the original models and the flexibility Workbemch offers. I would rather spend time learning how to play better than have a zillion options to tweak or sort through a million models created by other users. Not that it's not fun to experiment, but I only have so much time in a day!
  • 0

#19 ReverendLove

ReverendLove

    Iknowathingortwo

  • Members
  • PipPip
  • 124 posts
  • LocationCologne, Germany

Posted 04 September 2013 - 04:02 AM

How should that work?
Every single part, the bodies, the pickups even the pots and caps are represented as mathematical formulas and algorithms. That is called physical modeling. So: From scratch means that you build your own mathematical rules such as for an active pick up. That isn't anything what a consumer can do unless he is a firmware programmer and at the same time an engineer for electronics, mechanics and acoustics.

Yes. The only solution would be to hack the code of the firmware.

 


  • 1

#20 clay-man

clay-man

    Gear Head

  • Members
  • PipPipPip
  • 871 posts

Posted 04 September 2013 - 04:08 AM

Maybe they could add functionality similar to the Kemper Profiling amp?

In my opinion, I have more than enough options to play with, given the original models and the flexibility Workbemch offers. I would rather spend time learning how to play better than have a zillion options to tweak or sort through a million models created by other users. Not that it's not fun to experiment, but I only have so much time in a day!

 

I've thought about them doing that, but I think modeling a guitar is more complicated than modeling an amp, and there's reasons.

Modeling an amp is pretty simple since you can send a raw signal through the amp to play, then record how the amp sounds.

 

With a guitar, you can't feed a pick up a "raw signal". I'm sure there's some ways to get the frequency response of a pickup, but it's not as easy as an amp.

A regular guitarist doesn't have the tools at their disposal to possibly profile a guitar correctly.

With an amp, you just need a cable and a mic to record how the amp sounds.


  • 0

For a minute there, I lost myself.

Radiohead_bear-728286%5B1%5D.png 


#21 Charlie_Watt

Charlie_Watt

    Power User

  • Members
  • PipPipPipPip
  • 1161 posts

Posted 04 September 2013 - 05:25 AM

You guys have no idea how complex these models are or how many man hours went into the Variax design.  They made many very complex measurements to capture the data needed to model these guitars.  Good luck trying to replicate that!  Workbench gives you lots of options to tweak the models.  Use them!  It's ok to suggest new models but don't expect that to happen either.


  • 0

#22 sudiptasingh

sudiptasingh

    Just Startin'

  • Members
  • Pip
  • 11 posts

Posted 04 September 2013 - 06:41 AM

Just to clear up some points:

(a) I am not an advocate of hacking or reverse engineering - that always creates problems and issues (not supported, hacks break on new versions etc.) let alone the fact that it is extremely painful.
 

(B) I am advocating an API based approach - a set of class libraries and documentation, that outline the L6 HD architecture and API, and points at which programmers can access/leverage/add/extend functionality in a way that is supported by the L6 development team.
 

© This API is NOT for the Workbench User. This is for developers that are interested in extending the capabilities of (for example) Workbench (new Models, new Components etc.) and the core firmware. The majority of users should not know or care about how all this goodness is created - they should just be able to add them into Workbench, load them onto their Variax, Fire them up and Rock on!!

 

I hate to keep using the MS FlightSimulator analogy (because it is not exactly apropos) - but the base product ships with only X number of planes, airports, and scenery, but the developers have made available their tools, the API and the architecture of the product so that talented developers can develop for e.g. a full 737 simulation (a/c and systems), create custom sceneries/airports, add their own objects to the airport libraries (fuel trucks, runway markings, gauges etc.) This community continues to innovate even when the original product is no longer being developed....

 

If not - so be it - no worries!.... just pointing out some possibilities...


  • 0

#23 silverhead

silverhead

    Uber Guru

  • Line 6 Expert
  • 11449 posts
  • LocationOttawa, Ontario, Canada

Posted 04 September 2013 - 06:52 AM

@sudiptasingh:  ... and what's in it for Line 6? .... Remember, they're in business to make a profit. How does this generate revenue for them?


  • 0

Life is what happens to you while you're busy making other plans.
.... John Lennon

 

 


#24 sudiptasingh

sudiptasingh

    Just Startin'

  • Members
  • Pip
  • 11 posts

Posted 04 September 2013 - 07:02 AM

@sudiptasingh:  ... and what's in it for Line 6? .... Remember, they're in business to make a profit. How does this generate revenue for them?

1. More people innovating on their platform = more features on it = better competitive positioning = more sales

2. Revenue can also be indirectly linked to brand awareness and buzz... = more community - more buzz

3. Shortened beta product cycles (more testers) = quicker time to market = revenue impact

4. Increased customer satisfaction = ... more repeat business? More referrals? ...

5. Better product quality???? Architecture improvements etc. from many minds looking at the software...

 

Think Android vs. Nokia..... :-)


  • 0

#25 silverhead

silverhead

    Uber Guru

  • Line 6 Expert
  • 11449 posts
  • LocationOttawa, Ontario, Canada

Posted 04 September 2013 - 07:11 AM

The key factor here is the size of the participating audience. Android is a very general app platform, with the potential for many millions of users worldwide and virtually unlimited scope for applications. I don't think anyone buys Android for a single app.

 

Are there really that many guitar players who find the existing capabilities of JTV and Workbench sufficiently limiting that they would invest time and effort in this? Obviously you would. I wonder how many others like you there are?


  • 0

Life is what happens to you while you're busy making other plans.
.... John Lennon

 

 


#26 phil_m

phil_m

    Uber Guru

  • Line 6 Expert
  • 5512 posts
  • LocationMinneapolis, MN

Posted 04 September 2013 - 07:53 AM

I think I'd sooner expect monkeys to fly out my butt than I'd expect Line 6 to go down an open source road. In other words, it ain't gonna happen...


  • 1
Time is a train
Makes the future the past
Leaves you standing in the station
Your face pressed up against the glass

 


#27 TheRealZap

TheRealZap

    Uber Guru

  • Line 6 Expert
  • 15519 posts
  • LocationClemmons, NC USA

Posted 04 September 2013 - 07:56 AM

big fat smelly monkeys.

monkey-flying-small.gif

I think I'd sooner expect monkeys to fly out my butt than I'd expect Line 6 to go down an open source road. In other words, it ain't gonna happen...


  • 0

#28 sudiptasingh

sudiptasingh

    Just Startin'

  • Members
  • Pip
  • 11 posts

Posted 04 September 2013 - 11:20 AM

Love the monkey - but API based - not Open Source.... and developers not guitar players... all right - nuff said.


  • 0

#29 daedae

daedae

    Just Startin'

  • Members
  • Pip
  • 32 posts

Posted 04 September 2013 - 12:55 PM

The key factor here is the size of the participating audience. Android is a very general app platform, with the potential for many millions of users worldwide and virtually unlimited scope for applications. I don't think anyone buys Android for a single app.

 

Are there really that many guitar players who find the existing capabilities of JTV and Workbench sufficiently limiting that they would invest time and effort in this? Obviously you would. I wonder how many others like you there are?

It's interesting that I came across this thread today because I was just thinking about the same thing recently -- wondering if anybody had tried to reverse engineer the firmware.  As a developer (but not with any real experience in the embedded systems world), I would find it interesting, but of course all of the issues with the math of physical modeling that have been pointed out would be a major limiting factor.  And so I guess that's the real issue for a version with Line6's support: while it's true that it could expand the technology base and show some customer goodwill, the amount of time and money it would take to develop and test a usable API and make it sufficiently resilient against bricking would probably be better spent with them just internally creating models.  (They probably have an API for internal use, but "internal, domain expert API" is likely to be significantly different from "external, general user API.")


  • 1

#30 alsithi

alsithi

    Iknowathingortwo

  • Members
  • PipPip
  • 101 posts

Posted 04 September 2013 - 12:57 PM

I would have thought ultimately the quality of the tone is largely down to the skill and imagination of the guitarist?


  • 0

#31 TheRealZap

TheRealZap

    Uber Guru

  • Line 6 Expert
  • 15519 posts
  • LocationClemmons, NC USA

Posted 04 September 2013 - 01:11 PM

Crap! i'm screwed :D

 

I would have thought ultimately the quality of the tone is largely down to the skill and imagination of the guitarist?


  • 0

#32 Charlie_Watt

Charlie_Watt

    Power User

  • Members
  • PipPipPipPip
  • 1161 posts

Posted 04 September 2013 - 02:13 PM

Me Too! :)


  • 0

#33 clay-man

clay-man

    Gear Head

  • Members
  • PipPipPip
  • 871 posts

Posted 04 September 2013 - 02:51 PM

I still think making your own guitar from scratch on the Variax is ridiculous no matter how you spin it.

You do not have the same tools as Line 6 to create pickup and body convolutions/algorithms correctly.

 

Like I said, maybe adding a bit more depth in what you can edit would be alright, but that depends if the current coding will allow that without having to rewrite everything again.

 

I say just add more bodies and pickups. That's honestly all we really need. SG, Jaguar, Jazzmaster, stuff like that.

It all depends on how much memory is left on the JTV to be used though.


  • 0

For a minute there, I lost myself.

Radiohead_bear-728286%5B1%5D.png 


#34 ozbadman

ozbadman

    Power User

  • Members
  • PipPipPipPip
  • 1689 posts

Posted 04 September 2013 - 05:21 PM

I still think making your own guitar from scratch on the Variax is ridiculous no matter how you spin it.

You do not have the same tools as Line 6 to create pickup and body convolutions/algorithms correctly.

 

Like I said, maybe adding a bit more depth in what you can edit would be alright, but that depends if the current coding will allow that without having to rewrite everything again.

 

I say just add more bodies and pickups. That's honestly all we really need. SG, Jaguar, Jazzmaster, stuff like that.

It all depends on how much memory is left on the JTV to be used though.

 

amen


  • 0





Also tagged with one or more of these keywords: models, user created

0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users