Where would you prefer to install modeling pickup?

Where would you prefer to install modeling pickup?


  • Total voters
    23

Per

Member
[BACKGROUND]
This is a follow-up for https://forum.fractalaudio.com/threads/would-you-find-a-modeling-pickup-useful.150392/ that I posted last year. The background is: I'm working on building a modelling pickup, one that can sound like a P90, or a PAF, or a Tele bridge etc. My hope is to build something that does to pickup modelling what Fractal did to amp modelling; not the first, but maybe the first that makes you not miss the real thing.

The progress has (predictably) been slower than I hoped for... A key part of this project is to build a sensor that is sufficiently accurate. Magnetic pickups are hard to work with because they are wildly nonlinear and piezo under saddle transducers are not exactly the best either. Since last year I have been trying a few different types of sensors and I now have a couple of good candidates.
[/BACKGROUND]

If there existed a modelling pickup that sounds good enough that you wouldn't miss your real magnetic pickups, how would you prefer to install it in your guitar? (See the poll above.)

I have one sensor type idea that would be installed in the bridge pickup slot. The main benefit of this approach is that it is quite easy to install.

The other approach would come as a set of saddles, not unlike Graphtech GHOST. Here you would get to keep your bridge pickup, and it alters the looks of your guitar less. However it would be less easy to install, and likely not be compatible with guitars with special bridges (e.g. Evertune).
 
What do you mean? An alternative to the Roland GK-3?

In the case of Roland V-Guitar modelling, a GK-3 installed between the bridge and the bridge pickup gives slightly different results than the ghost piezo saddles. I have both and I prefer the GK-3 because the piezo saddles capture some subsonic frequencies.

There are extended discussions about this at vguitarforums.com
 
Yes, an alternative to Roland GK-3, but with built-in pickup modelling, so in that sense it will be more like a Variax. The main thing I'm trying to achieve is more accurate modelling than either of these systems.

The saddle variation of what I am trying to build would be installed like Ghost, but it would not actually be a piezo pickup and wouldn't sound like one. This poll is mainly about the form factor. It sounds like for you the form factor is less important than the sound? So given the choice you'd choose the one that sounds better, even if the difference is small?
 
Last edited:
Thank you for the clarification.

Option 1: "In the bridge pickup slot" would replace the existing bridge pickup, or would it be an additional pickup installed between the bridge and the bridge pickup (like the GK-3)?
 
It would replace the existing bridge pickup. I don't think I can fit that type of sensor in a GK-3 sized package. But I'm interested to know what your preference is.
 
If that pickup modeler could faithfully reproduce my existing pickup, I wouldn't mind replacing it with the pickup modeler
 
Last edited:
The other approach would come as a set of saddles, not unlike Graphtech GHOST. Here you would get to keep your bridge pickup, and it alters the looks of your guitar less. However it would be less easy to install, and likely not be compatible with guitars with special bridges (e.g. Evertune).
This is why I voted for the bridge pickup option. It seemed like an easier way to go. Good luck on your work!
 
Obviously a pickup swap is easier-to-do (often much easier to do) than messing around with wired-up saddles.

So, I think pretty much anyone would prefer replacing their existing bridge pickup with the Modelling Pickup, provided that:
  • One of the modelled sounds could replicate or (better yet) improve-upon the sound of the replaced pickup; and,
  • The modelling itself isn't noticeably better (for physics reasons, one assumes) when installed in the saddles.
But, if either of those two provisos are questionable, then more discussion/comparison is required.

Question: Can a Modelling Pickup installed in the bridge successfully model the sound of a neck pickup? How about an acoustic?

I'd be thrilled to own a guitar with ONLY a Sustainiac in the neck and ONLY a Modelling Pickup in the bridge, if by doing so I could pull up all my favorite acoustic, strat, tele, les paul, metal, etc., pickup sounds (from every possible pickup location and wiring) at will!
 
My problem with pickup replacements is that I almost exclusively use Parker guitars and there's not much meat on these bodies. So the pickup has to be thin enough to fit into the Parker cavity.
With that caveat, brige pickup replacement would be my choice.
 
Agree that it'd be easier to install in the bridge position.

My question would be whether it's going to be tweakable enough (or have enough models) to have greater variety than just "P90, PAF, Tele bridge..." There are so many different flavors of pickups that I don't think many people would just want generic pup sounds. Also wondering whether it could model the neck or middle pup from the bridge position, and also a variety of in-between positions.

And one last thought: how will it be controlled? Do we have to route new switches/dials, or can the 3 or 5-way do it? What if we want more than 3 or 5 sounds available? Or would there be an external device that does the actual processing?
 
Thanks for the responses so far, it's very helpful to see what your preferences are!

So, I think pretty much anyone would prefer replacing their existing bridge pickup with the Modelling Pickup, provided that:
  • One of the modelled sounds could replicate or (better yet) improve-upon the sound of the replaced pickup; and,
  • The modelling itself isn't noticeably better (for physics reasons, one assumes) when installed in the saddles.
But, if either of those two provisos are questionable, then more discussion/comparison is required.

Question: Can a Modelling Pickup installed in the bridge successfully model the sound of a neck pickup? How about an acoustic?

A modelling pickup that is placed at the bridge and configured to emulate a different pickup position is technically at a disadvantage compared to one that is very close to the saddle. I believe the difference is audible if not compensated but the effect might be small enough that it can be hidden by DSP. Basically, any pickup that is not placed at the saddle will be unable to detect some overtones. To give some sense of the severity of the effect: If the sensing is done 1 inch from the saddle the first overtone that loses more than 3dB is approximately the 19th.

I haven't thought much about acoustic emulation yet. I think the bridge placement (vs saddle) might be more detrimental to acoustic tones than it is for electrics, simply because acoustics have more high frequency content. So for quality acoustic simulation I think the saddle pickups could have the edge. It's possible that software can hide the problem well here too but I'm not sure.

The sensor designs for the bridge pickup and the saddle work completely differently. The saddle variant has higher resolution (needed because the string moves less there) but also smaller dynamic range (it would not be able to capture the larger vibrations/bends farther from the saddle). The economics are also rather different, the saddle version will cost significantly more up-front to develop (needs special machines to assemble etc) but will cost less per unit to produce, so there is a threshold of # units sold where below that the bridge version is more profitable and above it the saddle version is more profitable. The saddle version is cheap enough that it could possibly be sold to guitar manufacturers to include in their guitars, the bridge version isn't.
 
I could be completely off-base here, so please feel free to correct me. Would you be able/willing to produce a pickup modeler that plugs into the guitar that uses the ghost saddles? IF so, it would be less for you to produce and be cheaper for those of us with the ghost system already.
 
My question would be whether it's going to be tweakable enough (or have enough models) to have greater variety than just "P90, PAF, Tele bridge..." There are so many different flavors of pickups that I don't think many people would just want generic pup sounds.

The modelling technology itself will be capable of modelling pretty much any pickup. Its input is basically a 3D model of the magnetic materials in the pickup + material descriptions and the coil(s) together with a schematic of the guitar's wiring (including parasitics). The issue is more how this should be exposed to the user, see below.

Also wondering whether it could model the neck or middle pup from the bridge position, and also a variety of in-between positions.

It will be able to model different pickup positions and combinations of pickups in different positions. The placement emulation is essentially the same tech that Variax and Roland use.

And one last thought: how will it be controlled? Do we have to route new switches/dials, or can the 3 or 5-way do it? What if we want more than 3 or 5 sounds available? Or would there be an external device that does the actual processing?

This is a really good question :) My plan is something like this:

* The guitar can run off a battery and behave as a normal guitar. But the user can also connect the guitar to a pedal (with a TRS AES3 cable) that acts as a remote control (and also as charger/power supply). This pedal will have some knobs that lets you control the sound (choose between presets, pickup models, maybe something to control virtual tuning, maybe something to control how hot the pickup is, stuff like that), but will be limited to ~the UI complexity of a typical guitar pedal. The pedal will probably be MIDI controllable and could possibly send MIDI control messages triggered by knobs/switches on the guitar.

On the guitar:
  • The user can optionally replace one knob on the guitar with a rotary encoder that acts as a preset selector knob.
  • To the extent that you want to be able to control the modelled sounds from the guitar, pots and switches on the guitar need to be replaced with digital ones (these will be solderless and easy to install). They might come with dip switches that lets the user choose function (like "volume 1"/"volume 2"/"tone 1"/"tone 2"/"midi cc 1") for each pot.

This is probably what will be there in the initial version. An app with more detailed controls will probably come as a follow-up, following the philosophy of pedals like Source Audio which offer full basic functionality through knobs and advanced stuff via the app.
 
Last edited:
I could be completely off-base here, so please feel free to correct me. Would you be able/willing to produce a pickup modeler that plugs into the guitar that uses the ghost saddles? IF so, it would be less for you to produce and be cheaper for those of us with the ghost system already.

The problem (at least for what I am trying to do) with Ghost saddles is that they are piezo under-saddle pickups. The pickups I'm working on use a different type of sensor that is designed to be more accurate. If the input to this system is a piezo pickup I do not think I would be able to make something that sounds better (or even as good as) Variax. The whole reason I think I have a chance to do better is that I think it's possible to build a better sensor.

(I will also use DSP chips that have much more power than the ones in Variax, but if the input signal is not good there is only so much you can do. I don't believe Variax's sound quality is mainly limited by raw DSP power.)
 
Thanks for the responses so far, it's very helpful to see what your preferences are!



A modelling pickup that is placed at the bridge and configured to emulate a different pickup position is technically at a disadvantage compared to one that is very close to the saddle. I believe the difference is audible if not compensated but the effect might be small enough that it can be hidden by DSP. Basically, any pickup that is not placed at the saddle will be unable to detect some overtones. To give some sense of the severity of the effect: If the sensing is done 1 inch from the saddle the first overtone that loses more than 3dB is approximately the 19th.

I haven't thought much about acoustic emulation yet. I think the bridge placement (vs saddle) might be more detrimental to acoustic tones than it is for electrics, simply because acoustics have more high frequency content. So for quality acoustic simulation I think the saddle pickups could have the edge. It's possible that software can hide the problem well here too but I'm not sure.

The sensor designs for the bridge pickup and the saddle work completely differently. The saddle variant has higher resolution (needed because the string moves less there) but also smaller dynamic range (it would not be able to capture the larger vibrations/bends farther from the saddle). The economics are also rather different, the saddle version will cost significantly more up-front to develop (needs special machines to assemble etc) but will cost less per unit to produce, so there is a threshold of # units sold where below that the bridge version is more profitable and above it the saddle version is more profitable. The saddle version is cheap enough that it could possibly be sold to guitar manufacturers to include in their guitars, the bridge version isn't.

That must be the reason why they recommend to install the GK-3 as close as possible to the bridge.

I voted bridge pickup slot, but for reliability I wouldn't mind to have a saddle pickup
 
IIRC, accurate pickup modeling requires the model to change depending where you fret (the length of the string) and that this problem increases with the distance between pickup and bridge, along with many other not-so-obvious complexities. I don't think realistic pick up modeling will arrive for quite some time. I'd love to be proven wrong though, by any company.

I've spent countless hours trying to get both the Roland and Line 6 products to sound like my standard humbucker. Not even close. I use the Roland for Strat positions 2 or 4 on my dual humbucker guitar, but not much beyond that. It's passable for live use, but I'd never record with it.
 
Last edited:
  • Like
Reactions: Per
When using the GK3 you have to set some parameters at the Roland VG device, like the Guitar Scale or the distance from each individual saddle to the GK pickup. These measurements are important for an accurate pickup modeling.

Also, the GK3 shall not be installed more than 20 mm farther from the bridge. The closer to the bridge the better. I suspect that the regular bridge pickup position may not be optimal for pickup modeling, as it is too far from the bridge.
 
Last edited:
  • Like
Reactions: Per
One other problem:

The saddle-based piezos, on the Variax, are famous for creating strange responses when doing palm muting.

Metal players frequently complained that it didn't sound the same as a real pickup would, because with a real pickup, the pickup is detecting string vibrations at a point between the heel of the hand and the fretted note, whereas with saddle piezos, the pickup is detecting string vibrations from behind the heel of the hand (on the opposite side from where the note is fretted).

Maybe this has been technologically solved since the last time I investigated the topic (which was some years ago). But at that time, I know certain heavy-muting metal players found the Variax unusable.
 
Back
Top Bottom