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

[Steve Dallas]

I can get them looking good individually, but after EQ'ing both individually, the combined response of them both together is never flat, is this normal?

Yes. It usually takes 2 or 3 iterations to get it just right. You are EQing a combination of direct sound and reflections and related phase shifts, so some experimentation is required.

 

[montyliam]

I had a quick go of doing them individually, the black line is EQ together, the blue line is individual. They both look pretty much the same to me?

 

[montyliam]

Briefly listening to the EQ both vs individual, I can hear the image is slightly more focused in the center with individual EQ on the speakers.

Current setup - Genelec 8030'c (LCR), *pending* Genelec 8020d (Surrounds), DIY B&C sub, Denon AVR4400H

 

[Chromatischism]

You'll want different filters for each one set to the same target.

 

[montyliam]

You'll want different filters for each one set to the same target.

That’s what I ended up doing, the result is marginally better.

 

[pmar4015]

CAlso, I've been reading lots of mixed opinions about whether to EQ the LR channels together or separately. My room is fairly symmetrical, and the LR show pretty much the same response independently, so I'm thinking it’s better just to EQ them together?

Have you tried using Frequency Dependent Windowing (FDW) in REW? I have a very rudimentary understanding of the FDW feature but I have heard it helps in avoiding correction of ‘reflections” thus allowing the user to only correct for the ‘direct’ sound signals (which is I believe generally preferred)

 

[Chromatischism]

Depends on the nature of the reflection. If it's bothersome, and you don't correct it, you aren't doing room correction.

 

[montyliam]

Have you tried using Frequency Dependent Windowing (FDW) in REW? I have a very rudimentary understanding of the FDW feature but I have heard it helps in avoiding correction of ‘reflections” thus allowing the user to only correct for the ‘direct’ sound signals (which is I believe generally preferred)

I haven't, but most of the EQ I applied was below 1k. I used a total of 7 PEQ filters for L/R and 6 for Center channel. The first reflection points for LCR are treated side and ceiling, as well as absorption behind me.

 

[tjcinnamon]

With MultEQ-X, is there anyway to disable the filters Audyessy calibrates and just use manual filters generated from REW, I can't figure out how to configure it do this?

Yeah, it’s the way I do it. I got fantastic results using REW.

How I did it was:
1) run MQX with the Audyssey mic
2) remove all of the filters on my LR
3) move that calibration over to the AVR (save as originalMeasurment)
4) take measurements with MultiEQ on with the recently uploaded preset in REW (if you use DEQ then leave that on)
5) flatten curve to taste in REW and export the EQ
6) open MQX and the originalMeasurment file , exclude the LR channels (all measurement positions
7) import the Left and Right REW files (you may have to click Individual measurements)

Now when you measure in REW you’ll get almost the exact predicted results.

So the key is having to measure with the MultiEQ (and possibly DEQ) on then modify that. It’s confusing but it works very well

 

[montyliam]

Yeah, it’s the way I do it. I got fantastic results using REW.

How I did it was:
1) run MQX with the Audyssey mic
2) remove all of the filters on my LR
3) move that calibration over to the AVR (save as originalMeasurment)
4) take measurements with MultiEQ on with the recently uploaded preset in REW (if you use DEQ then leave that on)
5) flatten curve to taste in REW and export the EQ
6) open MQX and the originalMeasurment file , exclude the LR channels (all measurement positions
7) import the Left and Right REW files (you may have to click Individual measurements)

Now when you measure in REW you’ll get almost the exact predicted results.

So the key is having to measure with the MultiEQ (and possibly DEQ) on then modify that. It’s confusing but it works very well

I used the same process. I realised that at least one measurement needs to be done with Audyessy to be able to use MultEQX properly, but that this measurement can then be excluded allowing you to only use REW EQ filters, worked great for me too!

 

[montyliam]

And as others suggested, I ended up with a slope of 1db/octave for the mains which sounded much better than flat!

 

[GalZohar]

Better leave DEQ off if you want to flatten the response. Then after you are done, you can decide if you like DEQ on or not. If you flatten the response with DEQ on, you won't be able to change the master volume, not be able to turn DEQ off without breaking everything, nor will you be able to use DEQ for its original purpose.

 

[pmar4015]

Better leave DEQ off if you want to flatten the response. Then after you are done, you can decide if you like DEQ on or not. If you flatten the response with DEQ on, you won't be able to change the master volume, not be able to turn DEQ off without breaking everything, nor will you be able to use DEQ for its original purpose.

Agreed. I was thinking exactly the same thing. I dont see how you can eq to a curve that is by definition “dynamic” unless, as you say, you never change the master volume

 

[montyliam]

Agreed. I was thinking exactly the same thing. I dont see how you can eq to a curve that is by definition “dynamic” unless, as you say, you never change the master volume

What does DEQ do exactly? I've currently got it disabled!

 

[Zedly]

DEQ is a form of equal-loudness adjustment. The idea is that, at lower volumes, a larger boost of the bass is required to make the bass have the same perceptual loudness. DEQ does exactly this: as you adjust the volume on your receiver, it boosts the bass frequencies accordingly. The settings in the DEQ menu determine how aggressively the bass is boosted. Here is a graph from my system that shows how the bass is adjusted at different volume levels.

 

[krabapple]

Equal loudness compensation requires not just low frequency, but also high frequency boost...which is also apparent in your graph. Think of DEQ as a much more sophisticated version of the 'loudness' button on old receivers.

DEQ also has an offset feature that allows the user to set at what master volume level DEQ compensation starts to kick in. This exists because different sources eg movies vs pop music vs 'classical', are variously mastered 'loud' and don't need to be boosted so 'soon'.. So if you listen to mostly classical (not loud), you set it one way, if mostly modern pop (LOUD) set in another . It's a bit of a muddle, unfortunately...in an idea world, there would be 'loudness' standards across production media.

Last but not least, Audyssey brains determined (with not much literature support) that the rear channels also need overall level compensation at lower volumes, so DEQ boosts them a few dB, which a fair number of Audyssey users hate. (You can manually lower them back if so, though it's not a perfect fix)

 

[Newman]

What does DEQ do exactly? I've currently got it disabled!

If you mainly use the volume control to boost quiet source material or pull back overly loud source material, aiming to always listen at a consistent average loudness in your room, then DEQ will do something that you very much don’t want!

 

[montyliam]

If you mainly use the volume control to boost quiet source material or pull back overly loud source material, aiming to always listen at a consistent average loudness in your room, then DEQ will do something that you very much don’t want!

I listen mainly closer to reference levels watching films!

 

[krabapple]

If you mainly use the volume control to boost quiet source material or pull back overly loud source material, aiming to always listen at a consistent average loudness in your room, then DEQ will do something that you very much don’t want!

Right. Use Audyssey DV (Dynamic Volume) for that.

 

[pmar4015]

I found the following video extremely helpful in respect to understanding the various Audyssey features/modes available