Favorite vintage ADA cabinet simulator

[vansinn]

Ah yeah, I see you points..  just to clarify, I do not own a microcab.
I follow this discussion because I miss a touch of the SM57 proximity in my MP-2 emulations, without which they tend to be a Bit on the lifeless side of things, so just might go about adding some limited version of the Thumpf into those filters..

[Casey_Butt]

I ended up bringing back the 'thump' but not so much that it gets ridiculous... about 1 dB down from stock at the highest setting.  After a few hours tweaking it, I'd say it's right at the upper end of what an SM57 could possibly get in terms of thump when right up against the grille cloth.  I don't see the point in any more 'thump' than that because it simply isn't realistic.

Yet, at the same time, it is a very useful feature.  It gives a realism to the cab emulation that many emulators don't have.  With the 'thump' down too far it does seem 'lifeless' but, more importantly, also a little 'fake'.  The proximity effect emulated by the 'thump' control does go a long way to making your ear think it's hearing a mic'ed cab.

But it's a delicate balance... too much and it muddies the sound (yet muddy at least in a realistic way :))).

It's a relatively simple thing... just a few extra dB fairly narrowly focused around 125 Hz and it starts to sound 'cab-y'.

[vansinn]

Isn't 125 Hz a Bit on the low side?
It was my impression that the thumpf should be related to the resonance of the an adult chest cavity, which to my understanding is round about 180 Hz..

As an example, up'ing the 170 Hz EQ point in the MP-2 does buy me a bit more realism, whereas 100 Hz makes it sound swampy, and 280 Hz adds more of a body-fullness.

Also, can we really detect the difference of the filter set at standard vs the subtle change of just ~1 dB?

[Casey_Butt]

Good point about chest cavity resonance.  The SM57 tends to peak higher - at around 200Hz - as well.  But the 'thump' on the uCab is centred at 125Hz because that's approximately the resonant point of a Marshall 4x12 with almost any 12" Celestions.  You can see it on the frequency response of just about any such cab.  Older 4x12's with earlier Celestions tended to resonate a bit higher, as do 2x12's and 1x12's.

You can see it quite dramatically on the graph below.  These are frequency sweeps I took myself of two 4x12's.  The black line is a Marshall 1960A with Celestion G12-65 and G12T-75 speakers in an X pattern - SM57 on-axis up to the grille cloth at the centre of the dust cap of a G12T-75; the red line is a Peavey JSX 4x12 with four UK Celestion G12M-25's, under the same mic'ing conditions.



As you can see, the resonant peak at roughly 125Hz is quite pronounced on the G12-65/G12T-75 cab and is also present on the G12M-25 cab, but not quite as 'thumpy' and shifted to a little higher frequency (though not by much and really more like an absence of the stronger peak at ~125Hz).  I think that's exactly what would be expected of G12M's compared to the higher powered and 'punchier' G12T-75.

But that peak also tends to reduce and smooth out as you move the mic back and when you mix multiple mics.  So if you want an accurate cab emulation you need to be able to adjust that resonant peak accordingly.

The 'Room EQ' on the MP-2 is something similar to a cab resonance control.  It sweeps from about 95Hz to 160Hz, but it is more broadly focused and less peaky than a single close-mic'ed 4x12 would be... it's more what you would see with more distant mic'ing and mutliple mics.  Lots of analog cab emulators do it that way, with no distinct cab resonant peak to speak of or a very broad one, including most direct boxes and all the Tech 21's SansAmps.  It's also a simpler circuit to emulate the broader, smoother peak.

You'd be surprised the difference a dB can make at that 125Hz peak.  For instance, the stock uCab's 'thump' control has a range of about 3.88 dB.  At '0' it sounds thin and weak and at '10' it sounds overbearing and muddy.  I cut it down to about a 1.5 dB range when I did some mods, but that didn't seem noticeable enough.  So now I have it at about 2.65 dB and it seems perfect - enough to get a good close-mic'ed type 'thump' but not so much that it's unrealistic.  I also have a resonance control, so I can shift the resonant point around from ~125 to 240Hz... and an 'open/bass' switch, so it can emulate an open or ported cab.  That covers pretty much all the basses.

[vansinn]

Lovely info!
Interestingly, when setting up my MP-2 EQ'ing for a more traditional rock/metal sound (but not for modern djenty stuff), I tend to end up with the EQ points/levels somewhat akin to the graphics above.

Examining such graphics somewhat vindicates me in my idea that what I feel is mostly missing in a filter approach, is indeed how the upper frequency part, jagging up'n'down quite finely, merely needs some sort of comb filtering.
Oh, and the fact that the more simple filters do not have the slope steepness on either side flank to accurately mimic the roll-off of actual speakers, thus still producing too much higher frequency fizz. The low side flank can be simulated by combining the Low and 100 Hz EQ's.

We're getting there. Small moves, Ellie, small moves..

[Casey_Butt]

What the uCabs and Ampulator have in common is that they include those comb filter notches to a much greater extent than most other analog cab sims.  That gives them a more realistic single, close-mic'ed cab flavour. 

Roughly 80 Hz to 1.2 kHz seems to be the most crucial range in terms of those frequency notches and peaks.  The higher ranges don't seem to matter nearly as much... as long as the high roll-off is right (which rolls off like a 2nd or 3rd order filter... the stock uCab is only 1st order).

If you think about it, that would be sort of obvious though, because the frequency range of a standard-tuned electric guitar spans 82 Hz (low E) to 1.175 kHz (high D - 22nd fret on the high-E string).  The lower ranges, say 82 Hz to 500 Hz, is where the bulk of the content is going to be.  That's where the rhythm work and chording is, and the significant harmonics of those notes are all in range of the speakers/cab.  Any frequency content the guitar can produce above 1.175 kHz are harmonics of decreasing strength... and if you're playing that high on the neck it is inevitably 'simple' single notes, not complex chords at the 22nd fret. :D  Not saying that the notches up there completely don't matter, but they are harder to notice and most analog cab emulators don't seem to bother with them at all.

Also, the further you move the mic away from the cab, the more those high-frequency notches smooth out.  Add in multiple mics and the response can get smoother again.

[vansinn]

Well.. I can see your points, but it seems to me you're analysing the electric guitar fully clean.
Once distortions are applied, the rate of fall-off of the harmonics changes quite dramatically, as distortion will add quite a lot of additional higher harmonics.
Hence my suggestions on using comb filtering in the upper mids to [guitar] treble region.
BTW, I'm well aware of the comb filter in the ucab, just never calculated it's frequency range.

Now, it is entirely possible that I'm in error on this, and if so, very simply because I might confuse my above statements with the fact that neither the ucab nor the MP-2 rolls off at the high end are any more than a 1st order function - like you stated.
As such, it's absolutely possible that I mostly just need a steeper slope on the high roll-off, check the results, and only then look into what else might be needed.

Another mechanism lending to your observations is that, when playing high up on the board, distorted trebles does sound less fizzy, which, even when the roll-off is a mere 1st order function, nevertheless does make the harmonics hit the slope much earlier than when playing further down.
As such, those higher notes automatically gets filtered off decently more smoothly, again supporting your comments.

Hmnn.. I'll go do the Om thing, poke the akashic in the 10th dimension, and ask my spiritual masters for guidance..

[Casey_Butt]

The point I was trying to make, and I think you said also, is that with high frequency notes the harmonics are simply too far apart and too high in frequency for the 1.2 ~ 5 kHz notches to impart much of an effect.

For instance, with a 1 kHz fundamental and even a severely distorted signal (say a pure square wave) the first significant harmonic will be at 3 kHz and the next at 5 kHz, which is past the roll-off of the low-pass filter.  So only the response of the cab (or emulation) at 1 kHz and 3 kHz will have an effect on the sound you hear.  Which is what you were saying.  At slightly lower frequencies the harmonics from ~ 1 kHz to 5 kHz will be in range, but they will also be spaced fairly widely apart and decrease in magnitude (of 1/n) with increasing order.

Lower frequencies, on the other hand, will have many more harmonics that will fall in range of the filters.  With a fundamental of 120 Hz, you could get harmonics at 240, 360, 480, 600, 720, 840, 960, 1.08k, 1.2k, etc.  So there is much more range for the notches to 'hit' the harmonics and impart their effect on the overall tone.  Plus, the lower frequencies are where you're going to be chording, and so playing several notes at a time which are themselves natural harmonics of each other (at least if you're in tune and playing the right notes :D).

It's just more unlikely that you'll 'hit' the high frequency notches with strong magnitude notes or harmonics, so they don't seem to have nearly as much an effect as the mid notches that can come into play much more significantly.

I modded my uCabs to have 2nd order high roll-offs.  I looked at dozens of cab responses and I just don't think 3rd order is really necessary... some cabs roll-off steeper, some more gradually, but most Marshall 4x12s are in the 2nd-3rd order range.  And it's much easier to filter a little more with EQ than it is to try to add sparkle back in when it's gone.

If you've got the fizzies I think the higher order filter is definitely in order... and it's an easy mod.

[vansinn]

We have reached consensus. I bow to thy observation skills.
But of course.. I missed observing the harmonics being spaced out further and further as we go up the range.

Good discussion ;)

[Casey_Butt]

Actually, I find the stock ADA cab emulations (uCabs and Ampulator at least... and the MP-2 as well from looking at the circuit) tend to be both too dark and also a bit fizzy.  I find extending the roll-off frequency a bit higher and increasing the filter to 2nd (maybe 3rd) order does the trick... gives you more brightness and clarity but also rolls of the higher fizzy frequencies more aggressively.

But I also think most people don't realise that an on-axis close-mic'ed cab with a distorted amp is really quite fizzy.  Most of the cab emulations I've looked at roll off the high-end sooner and more aggressively than any real cab I've mic'ed.  It seems they're trying to emulate the 'finished product' you get after EQ'ing, mixing and mastering - not really the raw, fizzy beast a close-mic'ed cab actually is.  An exception is the Tech 21 PSA.  That one seems to leave more of the high-end fizz in there.  Presumably with the understanding that the user will be studio savvy enough to realise that EQ is there to take care of that, if need be.

[Casey_Butt]

Actually, I have to correct something I said earlier about the range of the uCab 'Thump' control.  I had my numbers mixed up because I'm taking sweeps right now and I'm getting a range of 6.93 dB for the stock uCab.  I've modded mine to have a range of 5.73 dB.  The range was 2.83 dB when I felt it didn't have enough.

[MarshallJMP]

I must say this is a very interesting topic. 

[Harley Hexxe]

I've been following it, but it's getting in the realm of over-thinking it for my taste.

I'm getting good results from my MicroCabs, so I can't complain.

[rnolan]

I've always been happy with the MP2 cab sims, they work great into my studio setup.  They also do a great job on a MP1 (plug into the MP2 loop returns and use a zero'd patch).

[Casey_Butt]

In that case, this is what the MP-2 cab emulation is doing (green line is 4x12, blue is 2x12, Room EQ set at 12:00)...



That's based on a spice model, but it's accurate.

It's a smoother curve than the uCabs and Ampulator, which are more comb filter-ish.