Adventures in Fractal's Dual Rectifiers

[Orvillain]

While those curves clearly track really well overall, the general character of the amp's curve seems much more 'irregular' than the Axe... Is this telling us anything? Is that manifesting as something in realspace, ie. is this the 'smoothness' people claim to hear from the Axe vs a real amp?

I think that is to do with the real properties of a physically moving speaker isn't it?

 

[MirrorProfiles]

Okay, so here's what I found. Our OG Dual Rectifier has a 10A MV taper. The model uses 20A which is the same taper as the Recto2 models (https://store.mesaboogie.com/products/pot-593792-270s-1m-594-shaft-20-taper.html). I don't remember if I did this so that the models would be consistent or if it's simply a mistake.

So for a given MV setting on the real amp you will want to set the MV lower on the model. I'm undecided as to whether I'll change the model as that will mess up presets.

Other than that the model is in very good agreement with the amp.

This checks out with what I was hearing, and lines up to how I’ve had to dial in real amps (the trick with 3 channel rectos was keeping the MV REALLY low).

Totally understand leaving it, although Recto’s are an amp where the majority of the useful taper is at the quieter end. If it’s going to be inaccurate I’d probably veer more toward having more resolution at low settings than less…..

I’ve been slammed today but I’m going to do some more tests with reactive loads as well as cabs, and at higher MV settings on the real amp. I have no doubt the Fractal model will match with either of those, it was really just about those conditions with a real cab and the poweramp having lots of headroom.

Very much appreciate you diving back into that amp and checking, as well as relaying all the cool findings. It’s a very awesome feeling to know a company cares about this stuff (as well as listening to us dumbass customers)

EDIT: Actually the more I think about this, I think it makes sense to correct it. Anyone who doesn't like the new tone can easily bump the MV up. I don't see much benefit in keeping the incorrect taper if we have another Recto Red channel with that specific taper. It will make Recto 1 more accurate, it will give more resolution in a useful area, and it offers more variation in behaviour to Recto2. Not fussed either way really, but I think matching the real amp adds more accuracy AND a better range of control=everyone wins

 

[dean701]

Totally understand leaving it, although Recto’s are an amp where the majority of the useful taper is at the quieter end. If it’s going to be inaccurate I’d probably veer more toward having more resolution at low settings than less…..

I'm not sure that's quite a truthful statement. Given the graphs that Cliff provided in comparison to his amp, the model seems very accurate.

 

[MirrorProfiles]

I'm not sure that's quite a truthful statement. Given the graphs that Cliff provided in comparison to his amp, the model seems very accurate.

I’m not saying it isn’t accurate- Cliff himself said the tapers don’t match and the model needs a lower value than the real amp to match positions. I’m just agreeing with that - the lower half is where the character of the amps changes most drastically and typically where it sounds best

 

[MirrorProfiles]

Been slammed today so only just catching up to this thread....

Firstly here is a comparison with the exact same settings and a React IR rather than a Mesa Cab:



and here is Recto1 Red from yesterday:



and the real amp amp with the Mesa cab:



So, when using the React IR, the top end is much closer and kind of in the ballpark. Obviously some differences in the low end due to the load. This is kind of why I wanted to demonstrate with a cab (and why I'm very thankful Cliff posted the graph demonstrating with the cab). FWIW I compared load boxes to a Mesa cab load on a Rectifier recently, so you can hear how they affect things here too:



So we have established:

- Recto1's MV taper is 20A, the real amp is 10A (will verify what is inside my own amp)
- The Recto models have otherwise been verified by Cliff as being correct, and map very closely on a graph.

I know that the amps can be made to sound similar if I use a load with less aggressive highs (like the React IR). Ill do some examples shortly, but I have a feeling if I turn my amps MV up slightly, the top end will smooth out and it will be easy to match.

To establish how best to match the Recto model to my amp:

- it could be caused by some kind of circuit inconsistency between my amp and Cliffs. Could be condition of valves, voltages, bias, pot tolerances, values drifting, Mesa changing things, different revisions of the amp etc.
- it could be down to the Fractal model needing an even lower value setting than the real amp because of the tapers being different. For reference here is a graph showing 20A vs 10A (20A ramps up faster, 10A has more resolution at lower settings)

- the impedance curve of my particular Mesa cabs vary from the Fractal ones
- user error from me being a dumbass (there is a very large margin for error)

I'm still curious to know whether I'm able to closely match the very top end of the signal to a cab loaded amp - I know it might be an outlier use case (it requires comparing against an amp loaded by a Mesa 4x12 AND a low enough MV setting). I'm just going to test a few more things, mostly to scratch my own curiousity. Either way, I'm very happy with the Fractal model

 

[FractalAudio]

- it could be down to the Fractal model needing an even lower value setting than the real amp because of the tapers being different.

This is exactly what it is. If you set the amp to, say, 5, you need to set the model to 3. At least until we change the taper in the model (if we decide to do that).

A 10A taper will have 10% of it's resistance when the knob is set to noon. A 20A taper (like in the model) will have TWICE the volume at noon. So you'll overdrive the virtual power amp more and have less high end.

 

[MirrorProfiles]

This is exactly what it is. If you set the amp to, say, 5, you need to set the model to 3. At least until we change the taper in the model (if we decide to do that).

If I set the Fractal MV to very low values, its still hard to get that sizzle the same at the very top without using advanced parameters or more extreme settings. Much easier to match when using a load box.

Recto1 Red even passes volume at 0 so its possible to dial in with more range than the real thing.

I just did a pass with the real amp MV at 10am. Fractal matches the amp very closely with similar settings there. Even me just knowing what conditions things might differ is very useful for me when using the amp and model, even if it means I need to tweak things slightly. Been mixing since 10am this morning so my ears are a bit shot (apologies if its not close enough). Even still, heres the amp at 10am MV and Recto1 Red (set lower).





and for good measure Helix with MV at 3.3 (any slider to pot conversions want to validate that?)



and same amp settings with React IR (this is why real cabs as the load matters for the most authentic tone):




I don't believe its possible to get a decent sound with the Red channel with the MV above noon, its just so bloaty and farty at that point. Between 10am and noon is dicey territory. Around 9am-10am is where its at IMO. Team 10A here for sure, genuinely think that'll be a nice one to have corrected.

 

[Mongillo19]

If I set the Fractal MV to very low values, its still hard to get that sizzle the same at the very top without using advanced parameters or more extreme settings. Much easier to match when using a load box.

Recto1 Red even passes volume at 0 so its possible to dial in with more range than the real thing.

I just did a pass with the real amp MV at 10am. Fractal matches the amp very closely with similar settings there. Even me just knowing what conditions things might differ is very useful for me when using the amp and model, even if it means I need to tweak things slightly. Been mixing since 10am this morning so my ears are a bit shot (apologies if its not close enough). Even still, heres the amp at 10am MV and Recto1 Red (set lower).





I don't believe its possible to get a decent sound with the Red channel with the MV above noon, its just so bloaty and farty at that point. Between 10am and noon is dicey territory. Around 9am-10am is where its at IMO. Team 10A here for sure, genuinely think that'll be a nice one to have corrected.

To me they sound close but also sound like an IR isn't loaded really emphasize that sizzle

I think we all appreciate the deep dive tho!

 

[la szum]

Appreciate all the work you guys have put into this.

Next round is on me. Orvillain can get the 5 or 6 rounds after that.

 

[MirrorProfiles]

LOL, uploaded the versions pre IR... will update

 

[Mongillo19]

LOL, uploaded the versions pre IR... will update

Glad I wasn't losing my hearing I was like "yeah his hearing has to be fried from mixing"

 

[MirrorProfiles]

Glad I wasn't losing my hearing I was like "yeah his hearing has to be fried from mixing"

2nd part is still true

 

[SeeD]

I love how the real amp sounds! Nasty and spitty

 

[Orvillain]

Orvillain can get the 5 or 6 rounds after that.

Get yer wallet out ya slut, I brought graphs!

 

[SwirlyMaple]

If I set the Fractal MV to very low values, its still hard to get that sizzle the same at the very top without using advanced parameters or more extreme settings. Much easier to match when using a load box.

The "Speaker Compliance" knob on the Speaker tab makes a huge difference here. At 0%, a lot more of those high frequencies are present. Default is 50%, which darkens the tone considerably. (I'm not sure if it's expected behavior for it to have this large of an effect, but that's what's happening.)

I'm not finding that dropping the MV really has that much of an impact on those high frequencies (at least not compared to your real amp), but that Speaker Compliance setting does. I think this is why @2112 's clips were so much brighter with mid knob settings -- he was using a load box and dialed all those speaker parameters down to 0.

 

[MirrorProfiles]

The "Speaker Compliance" knob on the Speaker tab makes a huge difference here. At 0%, a lot more of those high frequencies are present. Default is 50%, which darkens the tone considerably.

I'm not finding that dropping the MV really has that much of an impact on those high frequencies (at least not compared to your real amp), but that Speaker Compliance setting does. I think this is why @2112 's clips were so brighter -- he was using a load box and dialed all those speaker parameter down to 0.

Ah thank you - that’s very interesting, and I’m definitely going to try this. Probably a good area for me to read up on more.

Would love some kind of resource where exact values are mapped out for certain settings (like the NFB value to turn a 2204 closer to a 2203), or to simulate common mods or circuit variations between amps.

 

[SeeD]

Just for my own fun I made a Quad Cortex capture of the NAM profile ( DI RECTO) you made and shared @MirrorProfiles

1: I used your Gitar DI to reamp the NAM profile I made a capture of and used a QC cab block


2: Your real Recto AMP DI, reamped with the same QC cab block


My capture of the real amp NAM file, is missing a bit of omff in the low end, but the hi end is there I think. I wonder if things would be more real sounding if we could capture the real amp, and not use a NAM proflie to make a QC capture..that was a mouthful...lol

 

[dean701]

Given the following, I thought compliance was more for low frequencies than high end but I know nothing.

 

[FractalAudio]

The "Speaker Compliance" knob on the Speaker tab makes a huge difference here. At 0%, a lot more of those high frequencies are present. Default is 50%, which darkens the tone considerably. (I'm not sure if it's expected behavior for it to have this large of an effect, but that's what's happening.)

I'm not finding that dropping the MV really has that much of an impact on those high frequencies (at least not compared to your real amp), but that Speaker Compliance setting does. I think this is why @2112 's clips were so much brighter with mid knob settings -- he was using a load box and dialed all those speaker parameters down to 0.

Yes, an IR isn't the same as a real speaker. Speaker Compliance models the dynamic change in speaker impedance. When a speaker moves its impedance changes. The low-frequency resonance shifts and the inductance decreases thereby reducing the high frequencies. We model all this. You can adjust the intensity of the speaker impedance change via this control.

If you take a DI off the output of an amp and put it through and IR it will be brighter than the actual speaker because the IR is static but the speaker is dynamic.

There's all sorts of other stuff we do re. speaker effects that a static IR doesn't capture (speaker compression, speaker distortion, "thump", etc.). If you want to compare to a static IR you need to turn all this stuff off but then it doesn't sound as good IMO.

People get hung up on the amp and forget that the speaker is part of the equation. We even model voice coil heating and the effect it has on the amp. We even model the change in radiation impedance of the speaker.

 

[SwirlyMaple]

Yes, an IR isn't the same as a real speaker. Speaker Compliance models the dynamic change in speaker impedance. When a speaker moves its impedance changes. The low-frequency resonance shifts and the inductance decreases thereby reducing the high frequencies. We model all this. You can adjust the intensity of the speaker impedance change via this control.

If you take a DI off the output of an amp and put it through and IR it will be brighter than the actual speaker because the IR is static but the speaker is dynamic.

There's all sorts of other stuff we do re. speaker effects that a static IR doesn't capture (speaker compression, speaker distortion, "thump", etc.). If you want to compare to a static IR you need to turn all this stuff off but then it doesn't sound as good IMO.

People get hung up on the amp and forget that the speaker is part of the equation. We even model voice coil heating and the effect it has on the amp. We even model the change in radiation impedance of the speaker.

Thanks, I really enjoy reading all these details on how the modeling processes work.

I have tweaked the speaker settings on other amp models previously, to get a feel for the audible changes they produce, and it seems with the Dual Recto, its impedance response and unique freq voicing makes it more sensitive to the Compliance value than some of the other amps.

Question: should we be expecting to hear much difference when the Output mode of the speaker tab is changed from "FRFR" to SS Pwr Amp + Cab? Does that just inhibit Speaker Drive and Speaker Compression, but preserve the other settings?