Axe III dyna-cabs VS Helix VS NeuralDSP

[Orvillain]

(Now it seems he’s saying you can’t record to different tracks?…)

Mult == summing multiple sources to a single recording - otherwise known as bussing or summing.

I'm fully aware that you can record multiple tracks from any of these modellers, and perform any phase unalignment you want in the DAW.

 

[SwirlyMaple]

Mult == summing multiple sources to a single recording - otherwise known as bussing or summing.

I'm fully aware that you can record multiple tracks from any of these modellers, and perform any phase unalignment you want in the DAW.

Indeed, but if you sum a pair of real mics, you’ll have the same inability to edit them after the fact.

When I replied, based on the context, I thought maybe you weren’t aware that you could still have the virtual mics recorded to separate tracks even if you used the phase delay parameter in the Helix dual cab block.

 

[Orvillain]

I thought maybe you weren’t aware

I'll do a video tomorrow of two mics on a cab, and what you get when you have one of them in a static position versus one of them being swept around. I'll do the same thing in Helix, Neural, and Axe3. Might be interesting.

 

[DrewJD82]

Mic up your studio monitors and put the mics out of phase. Problem solved.

 

[yeky83]

I'll do a video tomorrow of two mics on a cab, and what you get when you have one of them in a static position versus one of them being swept around. I'll do the same thing in Helix, Neural, and Axe3. Might be interesting.

I dunno, is sweeping the mic around representative of what you do in the studio? Do you have some intern boy in front of a blasting cab and yell at him from the control room to sweep this way or that? For most folks in the studio nowadays, isn’t it more like: place mics, listen; move one a bit, listen; adjust faders a bit, listen; repeat as necessary?

 

[Boudoir Guitar]

The simple answer for why the default isn't like reality is because then people would quickly realize how bad their cab block(s) with four different cabs, each quad mic'ed at wildly different positions/distances, would sound in "the real world", and how much better (and faster) it'd all be if they just used a single cab with 1 or 2 mics and learned to use a little EQ here and there when needed.

 

[JiveTurkey]

The simple answer for why the default isn't like reality is because then people would quickly realize how bad their cab block(a) with four different cabs, each quad mic'ed at wildly different positions/distances, would sound in "the real world", and how much better (and faster) it'd all be if they just used a single cab with 1 or 2 mics and learned to use a little EQ here and there when needed.

 

[SwirlyMaple]

If it's useful to anyone, here's a table you can use to match real-world phase offset using the delay parameter with Helix's dual cabs. (I'm guessing Fractal has a way of doing this too, so if you can specify an offset in milliseconds, this will work the same.)

If you just want a simple equation to calculate the phase offset, you can use (mic distance 1 - mic distance 2) * 0.074. Mic distances must be in inches, and the calculated value will be in milliseconds.

To use the table:

• Find one of your mic distances in the left column, and the other mic distance in the top row
• The value where they cross is the phase delay time (in milliseconds)
• Enter this value on the Delay slider for the cab which has the greatest mic distance

 

[James Freeman]

To add to what Swirly said;

Helix has 0.02ms delay steps which is one digital sample at 48kHz sampling rate, 1/48000= 0.0208ms.
Speed of sound is 343 m/s, so every 0.02ms steps equals to 7mm or 0.28" virtual distance.

I don't know how AxeFx can do 0.1mm or 0.004" distance steps at 48kHz sampling rate but that's awesome.

 

[SwirlyMaple]

I don't know how AxeFx can do 0.1mm or 0.004" distance steps at 48kHz sampling rate but that's awesome.

Gotta be some kind of interpolation. I could be wrong, but I suspect Fractal didn't capture discrete IRs with distance increments measured in sheets of paper

 

[Orvillain]

I dunno, is sweeping the mic around representative of what you do in the studio?

Yes, sometimes.

Do you have some intern boy in front of a blasting cab and yell at him from the control room to sweep this way or that?

No, I don't have enough money for that.

isn’t it more like: place mics, listen; move one a bit, listen; adjust faders a bit, listen; repeat as necessary?

Yeah sometimes, but generally I've got closed-back headphones on, and I'm sat in front of the cab moving the mic as I listen to it moving around - often in context with the bass guitar and kick drum too. Usually I'll pick a palm-mutey bit in the guitar. The idea is to find a spot that works in cohesion with the other low-frequency elements.

 

[James Freeman]

Even when phase aligned, blending 2 mics is going to cause the phase relationship to change (which is the purpose of blending mics in the first place).

Yep, the angle of the microphone and the cone itself is angled in relation to the grille cloth.
With Fredman's technique the proximity of the two mics affecting each other. (put a finger near the capsule head and listen).
And many other real world parameters that create phase misalignment when using two+ mics are removed by digitally aligning the samples.

How far down the realism rabbit hole we want to go?
It's almost impossible to digitally preserve all the tiny phase nuances that real micing has.

 

[Orvillain]

Gotta be some kind of interpolation. I could be wrong, but I suspect Fractal didn't capture discrete IRs with distance increments measured in sheets of paper

Cliff said in another thread that Helix capture an IR every 1.2inches and that the new Dyna-cabs have a finer degree of resolution.

Eric says they've got more data than this, but in their critical listening tests, nobody could tell that these IR's were being interpolated.

From my listening tests, he's correct. I don't think I can really tell that the Helix has lower resolution. But again... within a window.

 

[Orvillain]

Yep, the angle of the microphone and the cone itself is angled in relation to the grille cloth.
With Fredman's technique the proximity of the two mics affecting each other. (put a finger near the capsule head and listen).
And many other real world parameters that create phase misalignment when using two+ mics are removed by digitally aligning the samples.

How far down the realism rabbit hole we want to go?
It's almost impossible to digitally preserve all the tiny phase nuances that real micing has.

I was just saying to @MirrorProfiles last night, I'm not sure with all this normalization and IR alignment that the Fredman technique can even work inside these blocks. I intend to check it out today.

 

[James Freeman]

Gotta be some kind of interpolation. I could be wrong, but I suspect Fractal didn't capture discrete IRs with distance increments measured in sheets of paper

Oh, not the capture samples but digital samples.
I mean, how can you delay digital data by less than 1 quantized sample?

 

[James Freeman]

I'm not sure with all this normalization and IR alignment that the Fredman technique can even work inside these blocks.

It can't, fact.
There's too many tiny phase relationships between the two mics and cone.

We can sorta fake it, which works quite well in my experience.

 

[MirrorProfiles]

The simple answer for why the default isn't like reality is because then people would quickly realize how bad their cab block(s) with four different cabs, each quad mic'ed at wildly different positions/distances, would sound in "the real world", and how much better (and faster) it'd all be if they just used a single cab with 1 or 2 mics and learned to use a little EQ here and there when needed.

Bingo.

Like, the argument of “well they’re guitarists, not audio engineers” doesn’t hold up well when all the main modellig platforms have access to a pile of different EQ’s, compressors etc. There’s all kinds of potential for doing damage, but we’ll draw the line at IR’s that have the potential to be a bit delayed and this MIGHT cause a problem? There’s more than enough tools available to make a phasey mess.

The chart for aligning also sums my point up - this is what people should use if they want to manually align a distant mic to a close one OR it should be internally linked with an option to have it off.

 

[James Freeman]

I don't think I can really tell that the Helix has lower resolution.

Me neither, they chose just the right amount of samples to be enough.

 

[Foxmeister]

Cliff said in another thread that Helix capture an IR every 1.2inches and that the new Dyna-cabs have a finer degree of resolution.

Dyna-Cabs have a ¼ of the mic models that the Helix cabs have, so it seems to me that Fractal and Line 6 have simply taken different approaches to reduce the number of IRs required to support a cabinet with moveable mics.

 

[SwirlyMaple]

Oh, not the capture samples but digital samples.
I mean, how can you delay digital data by less than 1 quantized sample?

View attachment 6166

Easy. You do a realtime spline fit between the points, upsample to 96kHz, shift the waveform by 1 sheet of paper, and then downsample back to 48kHz. (/s)