Electric Mistress Simulation

DLC86

Fractal Fanatic
Here we go with my attempt to simulate my electric mistress within the axe fx III.
Sadly the addition of the new filter parameters in the flanger block wasn't enough to match the frequency response of both the dry and the wet path of the real pedal, my previous measurements were wrong due to a faulty cable.
These are the correct measurements (red is the wet signal, blue is the dry):

Mistress Dry+Wet.jpg

I matched those by using a combination of filters and a peq and here's the result:

Mistress Dry+Wet+Sim.jpg

I also matched the minimum delay time, the maximum depth and the amount of harmonic distortion by looking at an RTA and feeding the pedal with various test signals (pink noise, sine waves, etc.)

Can you spot the difference? (I mixed the noise coming from the pedal even in the sim sample, so as to not make you guess too easily)



Despite matching the frequency response I still hear a difference in the low end, I'll investigate more on that to find out what's the cause.
Also, the waveform of the LFO is a bit different and needs more work.

Attached you can also find the patch, try it out and let me know what y'all think ;)
 

Attachments

  • Electric Mistress DLC.syx
    48.2 KB · Views: 55
They both sound good - and honestly on my laptop speakers here I don't hear a difference! Would need to put on some headphones or hook up better speakers.
 
They both sound good - and honestly on my laptop speakers here I don't hear a difference! Would need to put on some headphones or hook up better speakers.
Yeah the difference is mainly on the low end and I doubt laptop speakers are able to even reproduce it.

Anyway, I thought about this difference today and I had an idea about what could be causing it.
I measured the dry path frequency response by disconnecting the wet output but I didn't disconnect the feedback path and, even though the feedback pot turned fully CCW shouldn't let any signal go back to the op amp, it might still influence it somehow.
As I have a bit of time I'll verify this.

BTW, I'll reveal which sample is which tomorrow
 
I think it sounds great. Mind you, I do the “Gilmour” thing and dial the mix way back, ala how he ran his through a parallel blender to balance out the level and reduce the mix intensity. For a subtle Wall era sense of modulation it sounds really spot on
 
I think it sounds great. Mind you, I do the “Gilmour” thing and dial the mix way back, ala how he ran his through a parallel blender to balance out the level and reduce the mix intensity. For a subtle Wall era sense of modulation it sounds really spot on
Yes, I do that too. Years ago I actually installed a wet level control in my real mistress to accomplish that same thing.
 
I do hear something on the low end as well. Patch 2 sounded like it’s got more sauce on it or maybe it’s just the sweep is starting on a different part of the phase cycle? I’ve got an older mistress with the hardwired power cable. Let me know if there is anything I can do to help. It is really nice to be able to have these sounds on tap!
 
I do hear something on the low end as well. Patch 2 sounded like it’s got more sauce on it or maybe it’s just the sweep is starting on a different part of the phase cycle?
I doubt that cuz I matched the min and max frequencies of the sweep pretty closely.
To my ears it's #1 that has more low end, but at the same time it is less affected by the flanging probably due to a different dry/wet ratio.

I’ve got an older mistress with the hardwired power cable. Let me know if there is anything I can do to help. It is really nice to be able to have these sounds on tap
If you're willing to dissect it you could try the same measurements I did to create your own version, cuz I'm sure it'll sound pretty different than mine, but other than that I don't know what else you could do to help.
Maybe just compare my patch to your real mistress? :)

These are approximately the corresponding knob ranges in the axe for my pedal:

Rate: 0.078Hz to 10Hz
Depth (Range): 8% to 40%
Feedback: 0 to -75%
 
I’d bet you could take two EM’s and they would likely have as much, if not more, tonal variation than these two samples

when I’m listening in a mix context, I don’t think I’d be able to tell it apart from the real pedal, but then again, I also couldn’t tell it apart from my old Hartman, or the ElecLady I replaced it with.

different in isolation, yes, but if I never looked at my rack I wouldn’t know if someone switched the pedals out.

to me that is the definition of a good emulation, and this one to me sounds spot on, like an EM. Maybe not one exact EM, but you hear it and think “that is an EM flanger”, doesn’t sound like anything else really.
 
The first one is the Axe-Fx because it has more bass. Measuring the frequency response by disconnecting the resistors will lead to erroneous measurements because the wet path loads down the dry path and vice-versa. Disconnecting the wet path resistor will lead to about 4 dB more bass response than there actually is.
 
Another thing is that the "net" frequency response through either path should be mostly flat. The EM uses the classic pre-/de-emphasis technique. There's an op-amp near the input that boosts the high frequencies. The RC network at the output is the inverse response. This is done to improve noise.

I did a quick simulation and the dry path only has about 3 dB boost at around 80 Hz.

After studying the schematic a bit the way it works is this:
  • Both paths have a pre-emphasis of about 20 dB centered at 2.5 kHz.
  • The wet path is high pass filtered at about 100 Hz via the 82 nF cap.
  • There's a bass boost in the dry path so that when the wet and dry are added the net frequency response is (nearly) flat. This accomplished by using a different mixer resistor value (13K) than the wet path.
  • A final de-emphasis cancels the pre-emphasis.

The effect is less pronounced in the bass frequencies. This is accomplished by filtering the wet path and then compensating the dry path to obtain a net flat response. Same thing is done in the Dimension D Chorus.
 
The first one is the Axe-Fx because it has more bass. Measuring the frequency response by disconnecting the resistors will lead to erroneous measurements because the wet path loads down the dry path and vice-versa. Disconnecting the wet path resistor will lead to about 4 dB more bass response than there actually is.
Correct, and that's exactly what I found out yesterday after I completely disconnected the sad1024 from the circuit (input, output and feedback path).

Mistress dry BBD-disconnected.jpg

Didn't have time to remeasure the wet path but I don't think it would change much between disconnecting the 13k mixer resistor or the 220nF cap behind it, the dry path is basically just those two components.


Another thing is that the "net" frequency response through either path should be mostly flat. The EM uses the classic pre-/de-emphasis technique. There's an op-amp near the input that boosts the high frequencies. The RC network at the output is the inverse response. This is done to improve noise.

I did a quick simulation and the dry path only has about 3 dB boost at around 80 Hz.

After studying the schematic a bit the way it works is this:
  • Both paths have a pre-emphasis of about 20 dB centered at 2.5 kHz.
  • The wet path is high pass filtered at about 100 Hz via the 82 nF cap.
  • There's a bass boost in the dry path so that when the wet and dry are added the net frequency response is (nearly) flat. This accomplished by using a different mixer resistor value (13K) than the wet path.
  • A final de-emphasis cancels the pre-emphasis.

The effect is less pronounced in the bass frequencies. This is accomplished by filtering the wet path and then compensating the dry path to obtain a net flat response. Same thing is done in the Dimension D Chorus.
Wow, thanks for taking the time to run the simulation and for giving me this detailed explanation, a lot of food for thought.

I'll try to correct my preset with these findings in mind, for now I just wish you a merry Christmas. 🥂
 
Do some of the Electric Mistress versions use a hypertriangle LFO ? That might be the missing ingredient to emulating more flangers.
 
Do some of the Electric Mistress versions use a hypertriangle LFO ? That might be the missing ingredient to emulating more flangers.
Yes the non-deluxe mistress seems to have an hyper triangular LFO, but the log waveform in the flanger block is pretty similar.
I didn't have the opportunity to plot the one on my pedal but by ear I can tell the only small difference is that the real mistress seems to stay a tiny bit longer at the extremes of the sweep.
 
Back
Top Bottom