Impedance curve in amp section?

Impedance curve is a winner for me. For a first step, I try to match it to whatever cab I'm using. But often I'll find myself fighting something, getting close with the right IR's but not quite there. That's when I find that trying out a different curve altogether that is in the same neighborhood (i.e. a different 4x12 cab), or just adjusting the height, slope, or resonant frequency of the impedance curve takes care of the issue in a way that EQ doesn't.
 
Is there a chance you disabled the update impedance curve on legacy patch load global setting, and were noticing legacy patches that didn't update to their new default?
Aha!

Yes, I'm pretty sure that's correct... I don't think I updated that setting.

I'll confirm later.
 
This curve is specific to a cab and speaker, this calls for philosophical/methodological question regarding amp modeling block:
1. Could the impedance curve considered part of the cab and speaker modeling as well?
2. I wonder if the IR already captures some of the impedance curve information, does that overlaps with this part of modeling...?
3. If one can capture speaker and cab characteristics well enough, would it make sense to model the cab and speaker out-right and make it an alternative to IR player instead?

The current Cab blocks does IR, smoothing, proximity, mic pre-amp modeling, which seems have a lot of overlapping with what's captured inside the IR already. Maybe it makes sense to separate cab from IR player...
Let me try to answer you:
Since the impedance is a characteristic of the captured cab, it may be considered part of the cab & speaker modeling. What speaks against this?
  1. The amplifier "reacts" to the impedance it sees. Since Impedance is U(f)/I(f) representation, you can see that on higher frequencies there is more current necessary from the amplifier to "generate" a certain voltage at the speaker terminals. Since this affects the ouput stage of the amplifier, it will couple back to nonlinear processes (depending on the amount of feedback used in the amp).
  2. A speaker gets hot. I mean, really hot. This depends on the amount of current flowing through a virtual lumped resistance (the voice-coil wire). Thus, the impedance curve gets shifted. Signal depending. Level depending. A slow linear shift, over time (linear time-variant system). And, as said in 1), this will affect the speaker. If I assume correctly, this effect is modeled as "Speaker compression" in the amp block.
  3. Speaker compliance. The speaker is highly nonlinear, with some dominant nonlinearities (mostly force factor and compliance). Depending on the displacement, more force is needed to generate more displacement. And displacement is necessarily for those low end. Uh, a nonlinear process. And it's coupled back (through the force factor) via a virtual impedance to the amplifier. See 1)
What does this all mean?

Putting nonlinear modeling in the cab block would introduce an invisible feedback loop to the amp block. That means, parameters set on the cab block affect the performance of the amp block (since you are changing something on the cab block will affect the modeling in the amplifier block).
For most people just knowing IRs (linear, feed-forward), this seems counter intuitive ("Hey, why is one block affecting another? It's digital, put it away!"), or people would start complaining about other things. Also, though you can measure the loudspeaker nonlinearities (and model them), you need more parameters, and the measurement equipment for doing so is costly. And then, you also need a way feeding the information back to the amplifier modeling.
From a developers point of view, it also makes more sense to keep the number of modeled nonlinear blocks as low as possible, since most nonlinear systems tend to get instable, and bugfixing can be tedious.

Hope that helps.
 
I generally set the curve to match the IR I'm using at the time.

I mainly play to FRFR using a York Audio Mesa 2x12 IR and a real Mesa 2x12 through a Matrix GT1000 and set the curve to the 2X12 Recto.
 
I almost always tweak this, the low end could sometimes be boomy, and I need to shrink the low end peak a bit. And for warm jazz tones, I often tilt down the high end. This is one of the most effective tools to deal with boomy IRs...
I'm with you, it's a purpose-built parametric eq, with a bunch of reality-based preset starting points, great tool. It don't always tweak the curve, but sometimes I definitely do exactly the things you said, tame boominess or excess highs. Lots of ways to adjust this box...
 
Using a real 212 cab so had to change back the 6 parameters on the speaker page to my cab settings whenever another amp is selected.
One 212 impedance curve matches imp measurement i did on my cab, only have to set LF and HF resonance after selecting that curve.
A global User Defined Speaker page (same layout/params as amp spkr page) for real cab users overriding the one in the amp model,
that would be very convenient. Or a global option to lock the impedance curve, regardless amp model selected.
But look at what we got already, par output eq was a nice surprise, 6CA7 etc etc, so I’m far from complaining
 
Great question.
I imagine this is something 85% of guys don't even consider. May not even notice is there in any detail. Add me to that list. I shouldn't have been ignoring this. :pensive:

This is the load the amp model is "seeing" so naturally it's going to open up a whole other level of amp/curve/cab permutations when building presets.
 
I set the cirve to match the cab block as close as possible - seems logical.

For my real cabs setup (2 ported Mesa112s (C90s) powered by Matrix GT1000fx) I've tried to calculate the LF curve parameters according to the formulas fractal posted in a related thread here and the specs info posted by Mesa/Celestion, and guesstimates for some of the other parameters - ie, for cab, I chose the 112 v30 to match my Mesa 112s but that may not be correct as mime are ported so... I've fiddled around with these by ear but for the deeper parameters I prefer to "tweak by logic" than by ear.
 
I've found the Texas Star 2x12 works well with my preferred IRs... But sometimes others work better with different amp models.

Any time I try new IRs I'll try out some of the different curves to see what sounds best.

Sometimes a different curve can do wonders especially in the bass response.

Also, some of the amps default to reactive load, which isn't likely "authentic".
if you import an old preset via Fractal sometimes you get th eract
Interesting situation, if I convert Axe-FX 2 preset to a 3 preset, the default seems to be reactive load.

I will tag @AlGrenadine so he can investigate.
Yes, I found the same.
 
I LOVE the speaker impedance curves. Spent a LOT of time with them in making the Live Gold presets, just trying various speaker impedance options out (not just default) with various Cab IRs, until it felt and sounded right like amp-in-a-room to me. They can be just the thing that wakes up the amp if you are struggling with dialing it in.

These really help help get the bottom end "just right" when using/auditioning different IRs. Huge leap forward when these preset curves came along for the speaker page!

Kudos to Cliff for making this part of the amp block.
 
I LOVE the speaker impedance curves. Spent a LOT of time with them in making the Live Gold presets, just trying various speaker impedance options out (not just default) with various Cab IRs, until it felt and sounded right like amp-in-a-room to me. They can be just the thing that wakes up the amp if you are struggling with dialing it in.

These really help help get the bottom end "just right" when using/auditioning different IRs. Huge leap forward when these preset curves came along for the speaker page!

Kudos to Cliff for making this part of the amp block.

Totally agree. Not only for the bottom end, but for the high end where the bite can be too much for some cabinet, not enough for another. Changes the "feel" or response of the amp in those areas in a different way than just adding or removing the frequency.

Also the resonance parameter: in some cases, turning it up or down a notch can make or break the IR's interaction with an amp.
 
Hello Everyone,

Which is the best impedance curve to select for a Carvin Legacy amp block? the default is pointing towards 4x12 USA lead, where as the suited IRs are all 2x12 leggy's, so how come it is pointing to a 4x12 Mesa model?

Anyone has any more information? As far as I remember the imp curve was added (FW 9 onwards I assume) where as the legacy model is very old and has never been updated.

Havent seen or heard any comparisons for the Vai model with the original legacy model like we usually see comparison with other models, hence cant understand how close is the modelling compared to real legacy amp.

Can anyone help to shed some light?
 
Hello Everyone,

Which is the best impedance curve to select for a Carvin Legacy amp block? the default is pointing towards 4x12 USA lead, where as the suited IRs are all 2x12 leggy's, so how come it is pointing to a 4x12 Mesa model?

Anyone has any more information? As far as I remember the imp curve was added (FW 9 onwards I assume) where as the legacy model is very old and has never been updated.

Havent seen or heard any comparisons for the Vai model with the original legacy model like we usually see comparison with other models, hence cant understand how close is the modelling compared to real legacy amp.

Can anyone help to shed some light?

You can try 1 X 12 V30 or the 2x12 Lead 80 and see what you think...
 
Back
Top Bottom