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

[Chromatischism]

You can see it with Audyssey enabled.

You can see a similar thing in action with my S400's. The rolloff happens at 33 Hz, but the room picks them up and carries them further. This was testing different crossovers; just look at the black line.

 

[anotherhobby]

I hope this post answers all the speculation about the graph I posted. Man this place loves to speculate on graphs, LOL!

First, the system:

• Denon AVR-X4300H
• Fronts: Focal Aria 936
• Subs: 2 JL Audio e112's with a miniDSP 2x4
• Rest of the speakers: doesn't really matter, not covered here

Next, some context. This is a large open space with 9' ceilings and fully open to two other floors. All the surfaces are hard, including the wood floors, and all of the furniture is leather. It's a beautiful space that I absolutely love, but it's not ideal for audio, that's for sure. I'm incredibly resistant to additional room treatment here, although I have some absorbers behind the main speakers. I have two other heavily treated rooms in the house, so it's not like I don't know they work. There are two seats, and only two seats for music/movies/whatever. My wife and I have no kids (that ship sailed a long time ago) so this room is just ours. All corrections were made to balance both seats, with a small selfish bias towards my set. Below is a pic of the space. The TV and center channel are on a MantelMount that drops down 31" so the TV is at eye level when we watch stuff. My seat is on the right, my wife sits on the left.

The frequency response below 20 Hz in the earlier post I made is just room gain. There is zero EQ boosting as you can see from the raw measurement comparison below. Everything in EQ on the subs is a cut. I only used 5 bands of EQ per sub in MSO because I wanted MSO to handle only the bigger differences individually, and then Audyssey would correct the rest as a unit. The volume knob on both JL's is at the 0 mark, where it is also notched and the manual recommends:

When rotated fully counter clockwise, the E-Sub’s output will be fully muted. When at the “0” or straight up position, the level is at reference gain. When turned fully clockwise, the E-Sub’s level is at its maximum sensitivity (loudest). Set the subwoofer output level to “0” or its middle position.

The way this is currently set up, the JL subs have zero problems at "reference level" of 0 dB on my Denon watching something like an Avengers movie. Since there is zero EQ boost on the subs, and they are paired with Focal Aria 936's, that shouldn't be surprising.

Below are measurements of the completely uncorrected response of subs+mains (bypassed miniDSP, plugged subs in direct) compared to 5 EQ channels from MSO per sub (on a miniDSP) plus Audyssey, plus several bands of parametric EQ in MultEQ-X that I added to clean up the response at the only two seats I care about. If I did not EQ the subs first with MSO, Audyssey would have had me cut the volume on the subs by quite a bit. Instead I let MSO use up that headroom for EQ.

Just for fun, here is same MLP frequency response compared to when I only had 1 JL e112 sub and audyssey made me turn the sub volume down by 25%.

The bass response I'm getting with this setup is fantastic, powerful, and incredibly smooth. The next graph is comparing the same bass response above at my seat to the bass response at my wife's chair. I put a lot of effort into trying to maximize the response at both listening seats. I could boost some of my left over dips, but many of them will send her peaks into nasty audible level. I also did my manual EQ adjustments using spacial measurements around where our heads are to get a good balance. I wasn't correcting for head in a vice single point, but the measurements are all single point.

Next let's look at the full response. I still have more work to do. I was still correcting this past weekend when I had to stop, and I won't get a chance to dive back in deep for two more weeks. That 1400 Hz hump we both hear still needs attention, as do some other spots.

Below is what Audyssey corrected it to on it's own, before I tweaked it to the above. It rolled off the bass, it left the middle sucked out quite a bit, and rolled off the high's more than I like. My position here is the light green, and my wife's is blue:

Here is my seat, base Audyssey correction on top of MSO vs my additional EQ:

And the same comparison for my wife's seat:

Overall, I'm quite happy that MultEQ-X allows for additional parametric EQ. It let's you adjust things much more specifically for your situation. I was able to carefully improve both listening positions noticeably for both seats at the same time.

 

[Chromatischism]

Below is what Audyssey corrected it to on it's own, before I tweaked it to the above. It rolled off the bass, it left the middle sucked out quite a bit, and rolled off the high's more than I like. My position here is the light green, and my wife's is blue:

In my experience, Audyssey does not aim to roll off any bass like that. Looking again at your sub specs shows that JL has them at -10 dB by 17 Hz. Therefore, your response is normal for small sealed subs plus a little room gain. Audyssey will not boost them without Dynamic EQ or modifying the target curve, so trading headroom for flat bass is your decision to make (in my experience it's usually a good decision for sound quality, as long as you don't need super loud).

I guarantee that with your final response you are getting a lot of compression from 30 Hz and below at -0 MV during heavy LFE scenes. Such scenes will call for 115 dB output capability that simply isn't there until you get to two 15" vented subs. Two 12" vented can get there but only in a small room with lots of room gain. How do I know? I've moved from some of the better performing 12" sealed subs to a pair of vented 15". I know how they measure and how they sound, and the difference during these scenes was profound. I went 5 years not knowing I was getting compression. It wasn't audible distortion; they just ran out of gas. In that scenario, I didn't know what I was missing. But, during reasonable playback levels you should be happy with the sound.

The rollercoaster from 100-300 Hz is typical of SBIR.

 

[anotherhobby]

In my experience, Audyssey does not aim to roll off any bass like that. Looking again at your sub specs shows that JL has them at -10 dB by 17 Hz. Therefore, your response is normal for small sealed subs plus a little room gain. Audyssey will not boost them without Dynamic EQ or modifying the target curve, so trading headroom for flat bass is your decision to make (in my experience it's usually a good decision for sound quality, as long as you don't need super loud).

I guarantee that with your final response you are getting a lot of compression from 30 Hz and below at -0 MV during heavy LFE scenes. Such scenes will call for 115 dB output capability that simply isn't there until you get to two 15" vented subs. Two 12" vented can get there but only in a small room with lots of room gain. How do I know? I've moved from some of the better performing 12" sealed subs to a pair of vented 15". I know how they measure and how they sound, and the difference during these scenes was profound. I went 5 years not knowing I was getting compression. It wasn't audible distortion; they just ran out of gas. In that scenario, I didn't know what I was missing. But, during reasonable playback levels you should be happy with the sound.

The rollercoaster from 100-300 Hz is typical of SBIR.

Yeah, not trying to win any contests. I live in the city in a 28' wide house on a 40' wide lot, and my neighbors are the same. My subs are physically less than 20' from their living room. I barely have any right to listen to them as loud as they go now. Louder is absolutely 100% not a goal.

Maybe when I have some time in a couple of weeks I'll run some sub only sweeps and keep working the volume up. I'm sure you're right about compression at 115 dB. It'd be interesting to see where they start to run out of gas.

 

[Chromatischism]

My pair of Rythmik F12's could muster 102 dB at 20 Hz and 6 dB or so (I forget exactly) more by 50 Hz. That means the maximum volume level on the AVR without compression is around -13. And that's with a completely flat response, which almost no one runs. At a normal sub level of around +6 dB above the treble, that's -19 when you hit max output on the LFE in heavy scenes. Basically, you can turn it up, but the subs won't get any louder. That's why it takes dual vented 12-15" to play closer to 0 MV. With that said, a safe and comfortable movie level for me is -15 to -20, but you bet my subs are boosted.

 

[anotherhobby]

My pair of Rythmik F12's could muster 102 dB at 20 Hz and 6 dB or so (I forget exactly) more by 50 Hz. That means the maximum volume level on the AVR without compression is around -13. And that's with a completely flat response, which almost no one runs. At a normal sub level of around +6 dB above the treble, that's -19 when you hit max output on the LFE in heavy scenes. Basically, you can turn it up, but the subs won't get any louder. That's why it takes dual vented 12-15" to play closer to 0 MV. With that said, a safe and comfortable movie level for me is -15 to -20, but you bet my subs are boosted.

I can't test for subwoofer compression, as the UMIK-1 has too much sensitivity to go much over 95 dB. You can supposedly open up the UMIK-1 and adjust dip switches, but the ring on mine won't turn, and I'm done trying before I break something. A single e112 has been measured (open field) to be completely linear to 100dB at 2M, and then up to 105 dB at before compression becomes impactful. I have no idea what that translates into with room gain and also running two of them, but then with EQ on top... Who knows. I'm not worried about it, but if you want to worry for me about my bass at 115 dB go ahead.

 

[Chromatischism]

I can't test for subwoofer compression, as the UMIK-1 has too much sensitivity to go much over 95 dB. You can supposedly open up the UMIK-1 and adjust dip switches, but the ring on mine won't turn, and I'm done trying before I break something. A single e112 has been measured (open field) to be completely linear to 100dB at 2M, and then up to 105 dB at before compression becomes impactful. I have no idea what that translates into with room gain and also running two of them, but then with EQ on top... Who knows. I'm not worried about it, but if you want to worry for me about my bass at 115 dB go ahead.

Don't worry, I'm not. The reason I mention it is because you talked about playing at reference level (when calibrated to 0 MV, 105 dB peaks on speaker channels, 115 dB peaks on LFE). People understimate what it takes to actually produce it.

 

[chych7]

Interesting, I see the E112 measurements on audioholics and am comparing it to my own dual HSU ULS-15mk2s (measurements). The HSUs are a few dB stronger than the E112 based on the CEA measurements, but it seems that it has a steeper roll-off, so I don't get much at 10 Hz (drops off at 16 Hz for me). I have a closed HT room (19' x 12' x 9') too, so that should give me some gain.

 

[GalZohar]

Contrast that with my mains and center (Prime Towers and Center), which get detected as 2nd order at 30Hz (Large) and 40Hz. I run these with 40 and 60Hz crossovers, which gave me the best integration. Then I tried changing them to 4th order as well and it seemed to make bass in those channels sound more punchy. I'm not really sure why that would be, since those rolloffs would be below the crossover points I'm assigning. But that is what has given me the best sound.

If you set them to 2nd order rolloff at the crossover frequency (ignoring their measured rolloff), do you get any better results?
Do you have measurements of the effects of any/each of those changes?

Anthem Arc for example seems to do a 12db/oct in the target curve for both speaker and subwoofer, and then another 12db/oct from the crossover itself, providing symmetrical 24db/oct slopes (or at least that's what I managed to understand, don't have one to measure myself). Seems like Denon expects 12db/oct rolloff from Audyssey so crosses the speakers over with 12db/oct, while the subwoofer is not expected to roll off anywhere near the crossover frequency and thus gets a 24db/oct crossover slope. Not really sure what Dirac receivers do.

 

[Dj7675]

I hope this post answers all the speculation about the graph I posted. Man this place loves to speculate on graphs, LOL!

First, the system:

• Denon AVR-X4300H
• Fronts: Focal Aria 936
• Subs: 2 JL Audio e112's with a miniDSP 2x4
• Rest of the speakers: doesn't really matter, not covered here

Next, some context. This is a large open space with 9' ceilings and fully open to two other floors. All the surfaces are hard, including the wood floors, and all of the furniture is leather. It's a beautiful space that I absolutely love, but it's not ideal for audio, that's for sure. I'm incredibly resistant to additional room treatment here, although I have some absorbers behind the main speakers. I have two other heavily treated rooms in the house, so it's not like I don't know they work. There are two seats, and only two seats for music/movies/whatever. My wife and I have no kids (that ship sailed a long time ago) so this room is just ours. All corrections were made to balance both seats, with a small selfish bias towards my set. Below is a pic of the space. The TV and center channel are on a MantelMount that drops down 31" so the TV is at eye level when we watch stuff. My seat is on the right, my wife sits on the left.

View attachment 191519

The frequency response below 20 Hz in the earlier post I made is just room gain. There is zero EQ boosting as you can see from the raw measurement comparison below. Everything in EQ on the subs is a cut. I only used 5 bands of EQ per sub in MSO because I wanted MSO to handle only the bigger differences individually, and then Audyssey would correct the rest as a unit. The volume knob on both JL's is at the 0 mark, where it is also notched and the manual recommends:

The way this is currently set up, the JL subs have zero problems at "reference level" of 0 dB on my Denon watching something like an Avengers movie. Since there is zero EQ boost on the subs, and they are paired with Focal Aria 936's, that shouldn't be surprising.

Below are measurements of the completely uncorrected response of subs+mains (bypassed miniDSP, plugged subs in direct) compared to 5 EQ channels from MSO per sub (on a miniDSP) plus Audyssey, plus several bands of parametric EQ in MultEQ-X that I added to clean up the response at the only two seats I care about. If I did not EQ the subs first with MSO, Audyssey would have had me cut the volume on the subs by quite a bit. Instead I let MSO use up that headroom for EQ.

View attachment 191516

Just for fun, here is same MLP frequency response compared to when I only had 1 JL e112 sub and audyssey made me turn the sub volume down by 25%.

View attachment 191526

The bass response I'm getting with this setup is fantastic, powerful, and incredibly smooth. The next graph is comparing the same bass response above at my seat to the bass response at my wife's chair. I put a lot of effort into trying to maximize the response at both listening seats. I could boost some of my left over dips, but many of them will send her peaks into nasty audible level. I also did my manual EQ adjustments using spacial measurements around where our heads are to get a good balance. I wasn't correcting for head in a vice single point, but the measurements are all single point.

View attachment 191517
Next let's look at the full response. I still have more work to do. I was still correcting this past weekend when I had to stop, and I won't get a chance to dive back in deep for two more weeks. That 1400 Hz hump we both hear still needs attention, as do some other spots.

View attachment 191518

Below is what Audyssey corrected it to on it's own, before I tweaked it to the above. It rolled off the bass, it left the middle sucked out quite a bit, and rolled off the high's more than I like. My position here is the light green, and my wife's is blue:

View attachment 191522

Here is my seat, base Audyssey correction on top of MSO vs my additional EQ:

View attachment 191523

And the same comparison for my wife's seat:

View attachment 191524


Overall, I'm quite happy that MultEQ-X allows for additional parametric EQ. It let's you adjust things much more specifically for your situation. I was able to carefully improve both listening positions noticeably for both seats at the same time.

Thanks for taking the time to post your results of MSO+Audyssey. In particular I liked the measurements where each of you sit. I could have missed it, but what were your results with just MSO? Is there any issues with running MSO + Audyssey (ie having 2 correction systems) both applying filters? Thanks again

 

[Reverend Slim]

If you set them to 2nd order rolloff at the crossover frequency (ignoring their measured rolloff), do you get any better results?
Do you have measurements of the effects of any/each of those changes?

Anthem Arc for example seems to do a 12db/oct in the target curve for both speaker and subwoofer, and then another 12db/oct from the crossover itself, providing symmetrical 24db/oct slopes (or at least that's what I managed to understand, don't have one to measure myself). Seems like Denon expects 12db/oct rolloff from Audyssey so crosses the speakers over with 12db/oct, while the subwoofer is not expected to roll off anywhere near the crossover frequency and thus gets a 24db/oct crossover slope. Not really sure what Dirac receivers do.

No, as I don't have the gear to do REW, I don't have measurements. I'm actually NOT ignoring their measured rolloff so much as correcting Audyssey's poor detection of it. You can clearly see from the measurements screen that several of the speakers that it determined should be 2nd order actually more closely align to 4th order. For instance, my SVS Prime Satellites roll off almost exactly in line with 4th order yet Audyssey sees them as 2nd order. All I'm doing there is correcting things so it more closely matches the natural rolloff, which I do on my Satellites and Elevations based on how they measure. Basically, I'm setting the order of the cutoff then going to the next screen where it shows the filters to see how the rolloff complies with what was originally measured.

In all cases, what I've done to assess these changes is to load one version to reference and one to flat so I can quickly A/B while playing content to see which sounds best.

 

[anotherhobby]

Thanks for taking the time to post your results of MSO+Audyssey. In particular I liked the measurements where each of you sit. I could have missed it, but what were your results with just MSO? Is there any issues with running MSO + Audyssey (ie having 2 correction systems) both applying filters? Thanks again

There is no problem with layering Audyssey over MSO filters, and I think that's the best way to do it, as opposed to running MSO on top of Audyssey. The logic being that I'd rather correct the subs independently first, before applying a shared curve (Audyssey) on top of both.

As for using vs not using MSO for dual subs with Audyssey, the biggest impact was with my wife's seat. Both seats improved with MSO, but without MSO being applied before Audyssey, her seat sucked. She didn't have any bad peaks, but had several dips and a big broad suck out from 40-70 Hz. I only used 5 bands of EQ per sub with MSO.

I can't find before/after MSO only graphs, but maybe this will help. Here are both of our seats with 2 subs and Audyssey without MSO (my seat is the pink line, hers is blue):

Here are both of our seats (blue is my seat, green is her seat) with MSO first, then Audyssey, and no additional filtering on top at all, other than disabling the low frequency roll off curve on the subs in MultEQ-X:

It's NOT EVEN CLOSE. At least in my room, running MSO first is dramatically better than Audyssey on it's own. The only cost to do this was the $100 for the miniDSP 2x4. It's a bargain in my mind.

 

[ernestcarl]

reference level (when calibrated to 0 MV, 105 dB peaks on speaker channels, 115 dB peaks on LFE)

Consider that calibration can be done at a much lower level -- in such cases 0 dBFS signal sent to the LFE channel would not necessarily allow for 115dB maximum peaks. Using the term "reference" does not necessarily mean one has to abide by absolute THX standards -- well, since there's nothing really stopping anyone from adjusting their reference level to a lower (relative reference) SPL if desired.

@anotherhobby

Can you show us what your -20 dBFS or -30 dBFS sweeps are with AVR volume is set to max? From those traces, we should be able to determine what your "theoretical" maximum SPL output would be at 0 dBFS.

 

[Chromatischism]

If you set them to 2nd order rolloff at the crossover frequency (ignoring their measured rolloff), do you get any better results?
Do you have measurements of the effects of any/each of those changes?

Anthem Arc for example seems to do a 12db/oct in the target curve for both speaker and subwoofer, and then another 12db/oct from the crossover itself, providing symmetrical 24db/oct slopes (or at least that's what I managed to understand, don't have one to measure myself). Seems like Denon expects 12db/oct rolloff from Audyssey so crosses the speakers over with 12db/oct, while the subwoofer is not expected to roll off anywhere near the crossover frequency and thus gets a 24db/oct crossover slope. Not really sure what Dirac receivers do.

It is the Denon AVR itself doing the crossover filtering, but on the speaker side, 12 dB / oct comes from the receiver, while 12 dB would come from a sealed speaker, and 24 dB from a vented one, for a total of 24-36 depending on your speaker. On the subwoofer side, it is 12 and 12 (someone can correct if I'm wrong) for a total of 24. That's why THX recommends sealed speakers to get an equal 24 and 24, but I digress (and I don't think it's critical).

Some receivers do more, such as from NAD, where they do 24 dB / octave on the sub channel, according to Markus.

 

[Chromatischism]

Consider that calibration can be done at a much lower level -- in such cases 0 dBFS signal sent to the LFE channel would not necessarily allow for 115dB maximum peaks. Using the term "reference" does not necessarily mean one has to abide by absolute THX standards -- well, since there's nothing really stopping anyone from adjusting their reference level to a lower (relative reference) SPL if desired.

It applies whenever a film mixer puts a full-scale sound in that channel and the end user has their AVR calibrated to THX reference level.

THX reference level explained

What is 'THX Reference Level'? Reference level is a calibrated volume setting used for both movie production (in dubbing stages and post production houses) and reproduction (in screening rooms and theaters). The human hearing system is non-linear, especially in the bass, so having a consistent...

www.acousticfrontiers.com

 

[ernestcarl]

I don’t use it. Others may, of course. I know that a 0dBFS signal sent to my LFE channel by accident will not break my sub because I have specifically calibrated my levels taking the SPL limitation of speakers and my own usual listening volumes in mind.

How many people actually adhere to it?

 

[anotherhobby]

I don’t use it. Others may, of course. I know that a 0dBFS signal sent to my LFE channel by accident will not break my sub because I have specifically calibrated my levels taking the SPL limitation of speakers and my own usual listening volumes in mind.

How many people actually adhere to it?

My Denon AVR calibrates the volume exactly as Chromatischism explains. I appreciate that it calibrates the volume this way, because it has meaning, and provides a "reference" to the volume number.

I generally watch actual movies between -4 to -2 dB, and rarely but sometimes 0 dB. Reference level is almost always just a tiny bit louder than I like. Most Netflix/Amazon shows are -12 to -8 dB. HBO doesn't boost as much, so usually -8 to -4 dB. Most modern recordings for music I like at -15 dB for loud listening, while some recordings with more dynamic range I like at -12 to -10 db. Daft Punk's Random Access Memories is a good example of that. My wife is currently downstairs listening to Gillian Welch at -19 dB.

The point is, anybody else calibrated to reference level at 0 dB can set those exact same volume levels I just mentioned and get the same SPL as me. They can relate to all of the things I said, even if they have different preferences. I like that. It's scientific.

 

[ernestcarl]

My Denon AVR calibrates the volume exactly as Chromatischism explains. I appreciate that it calibrates the volume this way, because it has meaning, and provides a "reference" to the volume number.

I generally watch actual movies between -4 to -2 dB, and rarely but sometimes 0 dB. Reference level is almost always just a tiny bit louder than I like. Most Netflix/Amazon shows are -12 to -8 dB. HBO doesn't boost as much, so usually -8 to -4 dB. Most modern recordings for music I like at -15 dB for loud listening, while some recordings with more dynamic range I like at -12 to -10 db. Daft Punk's Random Access Memories is a good example of that. My wife is currently downstairs listening to Gillian Welch at -19 dB.

The point is, anybody else calibrated to reference level at 0 dB can set those exact same volume levels I just mentioned and get the same SPL as me. They can relate to all of the things I said, even if they have different preferences. I like that. It's scientific.

Thank you for clarifying what you use. I am quite curious though if you ever measured your sub at various volume increments — e.g. -20, -10, -5, -2 — and know when they begin to compress?

ex.

Red trace is just the 0 dBFS divided by the -20 dBFS (i.e. 'reference') measurement.

 

[Chromatischism]

I don’t use it. Others may, of course. I know that a 0dBFS signal sent to my LFE channel by accident will not break my sub because I have specifically calibrated my levels taking the SPL limitation of speakers and my own usual listening volumes in mind.

How many people actually adhere to it?

Well, this is an Audyssey thread. All Audyssey AVRs calibrate your system to 0 MV = THX Reference.

 

[ernestcarl]

Well, this is an Audyssey thread. All Audyssey AVRs calibrate your system to 0 MV = THX Reference.

Is there a protection mechanism by which one can set a certain known volume limit to prevent accidental damage?