Audyssey's Next Generation of Room Correction (MultEQ-X)

[HarmonicTHD]

^^ So, trying out your tip for a slight roll-off from 2kHz up, I tried out a 1st order high shelf filter for Ref only, deselected the Harman curve altogther, and ticked Toole for both Ref and Flat. So Ref with the added roll-off at the high end. I have to say I like it better now.

While I realise I have entered the realm of psychoacoustics well and truly, I mean, that is what Toole was about in the first place right. Even with Dynamic EQ in play. Interesting.

Trying out, learning. But man if this did not take Audyssey to a completely another level.

Look here from the Toole/Olive publication (generally called Harman preference curve for speakers):

The roll off (starts at around 1kHz) is caused by physics, simply put, an anechoically flat speaker interacting with room (size, reflectiveness, speaker and listener positioning). As you can see average listener preference differs quite a bit especially in the bass region, so dont be afraid to experiement.

I approximate my personal preference curve by superimposing a general tilt (in my case 0.8db/Octave at 1kHz) with two BiQuad filters to approximate the "bass" bump) - I currently use the "green" one named Reference with up to ca. 4dB bass boost. Like this:

Again, this is not universally applicable. It is my preference in my room and with my system, but of course try it and adjust to your preference.

 

[phn]

Ah, tilt, not high shelf. Thank you.

 

[HarmonicTHD]

Ah, tilt, not high shelf. Thank you.

I am sure there are several ways to achieve the same result (approximation of Harman curve) in MultiEQX. Toole himself however, mostly talks about tilt when he describes "desirable" basic tone controls and gives an approximate range of this tilt (see his book or publications), so I used tilt.

 

[randomer]

Curious, what does MultiEQ-X provide on top of the App and Ratbuddesey? Debating whether to go for MultiEQ-X or DL for my Denon 3800. Thanks!

 

[Acerun]

Does it matter where the volume "dial" is set when doing your initial measurements for MQX? I don't see anything about it in the user guide.

 

[HarmonicTHD]

Does it matter where the volume "dial" is set when doing your initial measurements for MQX? I don't see anything about it in the user guide.

Doesn’t matter. Audyssey controlled by MQX sets it automatically.

 

[Acerun]

Doesn’t matter. Audyssey controlled by MQX sets it automatically.

Thank you

 

[SCG]

It's not available for my computer (Mac) and I can't seem to buy it from the Microsoft store using my emulated Win 11 install

I was going to install VMware Fusion on my Mac (with mac silicon) and then run Win 11, but after reading the info below from Audyssey and your post looks like it's a complete no go.

"Newer Mac computers are based on the Apple Mx processors, which is an ARM-based processor. Running MultEQ-X in Windows ARM natively on the Mac Mx is not supported. Mac computers can also run Windows 11 ARM in emulation, such as with VMWare Parallels/Fusion. This configuration is NOT supported because Windows 11 ARM would require x64 virtualization to run MultEQ-X, but multiple layers of emulation are not possible."

Going from 'Design Target Curves' to 'Filter Settings' pages (where the bulk of the initial iterative back & forth work takes place) is about 10 secs per screen on my older PC laptop but almost instantaneous on my desktop, so using my new Mac would be so much quicker; OR if Audyssey would let you to design your house-curves/filters on the measured results pages like Dirac does.

 

[middlemarch]

I don't understand that, because I run various x86 binaries in my Parallels Win 11 installation (I even paid Microsoft for a full Win 11 license). For example, MSO runs fine, so does Ratbuddesy, as well as a couple of 32 bit x86 games. But when I go to the Microsoft store it won't even allow me to download Multeq-x. Are MSO and the others all 32 bit apps? I'm not that familiar with the Windows world. I do have an older Intel I7 based Macbook Pro that can run Bootcamp. Maybe I should try that, but I only have a legal copy of Windows 8.1. Would that work with Multeq?

In the meantime I've gotten better with finger drawing curves using the $20 Ipad app. Not sure I'd gain a whole lot for the extra $200.

 

[SCG]

I'm new to running native Windows ARM 11 and x64 apps in a emulation environment on Mac silicon. Per Audyssey on Windows requirements: "MultEQ-X runs on Windows 10 systems, version 1809 (October 2018 Update, AKA “Redstone 5”) Build 17763 or newer. The software is x64 native. Windows 11 is also supported."

I do know that one of the main advantages of EQ-X; with the phone app (post calibration) a house curve doesn't actually raise the sub levels but actually lowers everything else so that's why most people have to go back and raise the trim levels on the subs, causing a potential mismatch level in the crossover region.

But with 'Disable Auto-leveling' on, EQ-X it will keep all other freqs the same level (post calibration) while allowing you to actually increase the sub level outputs (i.e. keep room gain via your house curve by applying less attenuation in that region or boosting it if possible).

With the BF sale at 25% off I gave it a try and it works much better (less trial and error) other than not being able to modify the house curves on the measurement page, but they say that feature is on the 'dev list'.

 

[middlemarch]

In answer to some of my questions, it appears minimum requirement is Windows 10. My old laptop's version of Bootcamp I believe only supports win 8.1, so that option appears out for me. However, I went to Ars Technica's review of Win 11 Arm, and here's a full paragraph from that review:

"Today, Microsoft is formally blessing Parallels as a way to run the Professional and Enterprise versions of Windows 11 on Apple Silicon Macs. The Arm version of Windows running under Parallels has some limitations—no support for DirectX 12 or newer OpenGL versions, no support for the Linux or Android subsystems, and a few missing security features. But it can run Arm-native Windows apps as well as 32- and 64-bit x86 apps thanks to Windows 11's x86-to-Arm code translation features. Pretty much anything that isn't a game should run tolerably well, given the speed of Apple's M1 and M2 chip families, though there are still weird edge cases for specific apps, and hardware that requires specialized drivers to work in Windows may not support the Arm version."

So it does not appear that a 64 bit X86 app should be a problem, unless it's one of the listed support features? I think I've done everything right, but no bueno.

 

[SCG]

Good find, I think the reason is what MS stated in the info you linked, "Experiences that depend on an additional layer of virtualization (nested virtualization) are not supported." which matches with Audyssey's statement, "This configuration is NOT supported because Windows 11 ARM would require x64 virtualization to run MultEQ-X, but multiple layers of emulation are not possible.

So the unsupported nested virtualization requirement is the issue. Would be nice if someone found a hack, as I use the Mac for Dirac, REW, and using EQ-X would be nice as that computer is faster, lighter, quieter, with a larger better screen than my older PC laptop. Or (if by now ~ 2024) Audyssey made a Mac silicon version, like Dirac does.

 

[middlemarch]

Yeah, except I'm not buying the "nested virtualization" excuse. Win 11 Arm is fully Arm native, it just uses Parallels to map access to Mac system resources. Win 11 provides all the (most of the) hooks a 64 bit x86 app needs to run, then translates the remaining app specific stuff to Arm code. So they say Win 11 Arm " would require x64 virtualization to run MultiEQ-x". Well, Win 11 does just that. Many (most?) other x64 x86 (hate doing that) have no trouble running in that environment. So my suspicion is there is some low level code or specific interface that would need to be tweaked to make it run fine, and they just have no interest in putting out the effort. But with the soon to be proliferation of other Arm based PC platforms as soon as next year I don't think the effort would be wasted. But not my nickel.

Ultimately I don't really care, I agree with you, in my opinion if audessey wants to play at the same level as Dirac, they'd best be working on a native Arm MacOS version. They're halfway there since the IOS app works well, and under the hood there's not a lot of difference between IOS and MacOS. And hey, if they can do a Unix version, how hard would a Linux version be?

 

[Theriverlethe]

Yes, use the mic which came with your AVR.

The UMIK can not be directly used as you can’t hook it up to the AVR nor tell the AVR it’s calibration. Use the UMIK with REW.

So a $200 PC-based software package can't use a microphone connected to the PC with a calibration file?

 

[RDoc]

So a $200 PC-based software package can't use a microphone connected to the PC with a calibration file?

No because that's not how it works. The PC app analyses the data recorded and then sent by the ARV, does whatever processing it needs in the PC, and sends the results back to the AVR. The system is designed to work with the calibration hardware and software built into the AVR.

If the sound was analysed by the app in the PC what would it do with it? The output of the app is only going to work with an AVR with the Audyssey firmware built into it and they all have calibration mics and associated hardware.

 

[Theriverlethe]

No because that's not how it works. The PC app analyses the data recorded and then sent by the ARV, does whatever processing it needs in the PC, and sends the results back to the AVR. The system is designed to work with the calibration hardware and software built into the AVR.

If the sound was analysed by the app in the PC what would it do with it? The output of the app is only going to work with an AVR with the Audyssey firmware built into it and they all have calibration mics and associated hardware.

It would do it with a general-purpose CPU like every other program on a computer. Are you suggesting Audyssey doesn't know their own algorithms?

 

[HarmonicTHD]

So a $200 PC-based software package can't use a microphone connected to the PC with a calibration file?

No. That’s not how it is designed. Read the manual.

Don’t buy it. Or complain to Audyssey.

 

[Theriverlethe]

No. That’s not how it is designed. Read the manual.

Don’t buy it. Or complain to Audyssey.

I have no intention of buying it when the competing Dirac client is free for Pioneer, Onkyo, NAD, etc. users and works as one would expect.

 

[Palmspar]

I use the normal MultEQ app with a denon x3400h receiver and 5.1.2 atmos setup.
Midrange compensation off for all speakers, and only room correction under 500hz in small rooms or under 300hz for bigger rooms.
It works great with movies and music for me.
Dyn eq and dyn volume is nice to have, dirac live has not something like that i mean.

 

[Acerun]

Anyone willing to comment on this measurement? Suggestions or +/- welcomed. Crossing at 60 obviously. Do I want the cross at 75db, is that right?

Other short thoughts, is that hump at 50/60 normal? I also see all of the measurements converging at 100, might I cross there instead?

It sounds good but I know only enough to be a little dangerous, I don't really know what I'm shooting for, honestly.