Active Room Treatment (ART) by Dirac

[smygolf]

ART can work in mysterious ways....

I hooked up the system as it was originaly,
sub 1: 2 x front subs in parallell
sub 2: standing on the rear wall a couple of meters behind MLP

Now the front subs have a sensitivity of 104dB - was curious to see how ART should behave.
Well the joke was on me, all the support from 20-35hz was on the rear sub and front sbs was barely working


Lesson learned...
I need to build a baffle wall for the rear wall and ad one more 4645C - sound as fun

 

[kawauso]

ART can work in mysterious ways....

I hooked up the system as it was originaly,
sub 1: 2 x front subs in parallell
sub 2: standing on the rear wall a couple of meters behind MLP

Now the front subs have a sensitivity of 104dB - was curious to see how ART should behave.
Well the joke was on me, all the support from 20-35hz was on the rear sub and front sbs was barely working


Lesson learned...
I need to build a baffle wall for the rear wall and ad one more 4645C - sound as fun

ART seems to have smartly prioritized the rear subwoofer over the front ones, effectively taking the low bass away from the front.
Good luck, smygolf, with adding another subwoofer at the rear.

 

[smygolf]

ART seems to have smartly prioritized the rear subwoofer over the front ones, effectively taking the low bass away from the front.
Good luck, smygolf, with adding another subwoofer at the rear.

Short decay has priority as it should

 

[JRowland]

my lowest volume is -10dB and moving up just makes things better.

I live on top of a mountain with no close neighbors but I share the house with my wife. I would if I could (and do when I can), that's awesome that you can.

 

[dlaloum]

This is an interesting comparison of default ART with modified ART, but it seems that you compare L before with R after. It would be interesting to see L+R both before and after. Any chance of seeing that?

Accidental - I was intending to share L across the board... will remedy later!

 

[dlaloum]

Edited update: progress from first trial to current... - all measurements are for L (got rid of the R accidental inclusion!)

First try is 5.1.4 (single sub) - everything pretty much default ART... not so impressive (clarity was already improved over DL however - already substantial decay control here!) - given the default has all support active, I assume it was using all 98 / 98 available filters - but I didn't check or record that paramters at this stage (I wonder how good this more basic configuration could be if I used the paramters more akin to the bottom one in this posting!)

Then my initial (semi default) ART setup, this time with 2 subs...(5.2.4) vs my 2nd try improved setup...

The first attempt was clearly suffering from a number of issues, where the 2nd attempt is much better sorted

Here are the actual settings with their associated charts:

Initial Setup configuration: groups chosen by Dirac, Levels set by Dirac, FSL manually adjusted for Group 1, Group 3, Group 6 on G1,2,3) - All FSH @ 150Hz
	Supported By Group (sup level, FSL)	Filters Used	98 / 98
	Group 1	Group 2	Group 3	Group 4	Group 5	Group 6
Group 1 L/R	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz
Group 2 C	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz
Group 3 SL, SR	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz
Group 4 FLH, FRH	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 29Hz
Group 5 SLH, SRH	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 29Hz
Group 6 SW1, SW2	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz

And here is where I started to take charge of the support groups, and switch some of them off - one noticeable change - is I went from using all 98 filters, down to only 51 filters (perhaps the default configuration as above, needed more than the 98 filters available??)

	2nd Setup configuration: groups chosen by Dirac, Levels set by Dirac, FSL manually adjusted for Group 1, Group 3, Group 6 Support allocations edited, All FSH @150Hz
	Supported By Group (sup level, FSL)	Filters used	51 / 98
	Group 1	Group 2	Group 3	Group 4	Group 5	Group 6
Group 1 L/R	-18, 30Hz		-18, 35Hz			-18, 24Hz
Group 2 C	-18, 30Hz		-18, 35Hz			-18, 24Hz
Group 3 SL, SR	-18, 30Hz		-18, 35Hz			-18, 24Hz
Group 4 FLH, FRH	-18, 30Hz			-18, 20Hz		-18, 29Hz
Group 5 SLH, SRH			-18, 35Hz		-18, 20Hz	-18, 29Hz
Group 6 SW1, SW2	-18, 30Hz		-18, 35Hz			-18, 24Hz

Well isn't that interesting !! - it still isn't perfect (what is!) - but this is a very nicely controlled environment - massively improved, and yet there is less support specified...
My best guess is that this configuration was within the filter capabilites, whereas the previous config, required more filters than I had available - hence it let through more flaws!!

Perhaps we should keep an eye on the filters used field... avoid any configuration that maxes out the filters (as we cannot tell whether it is just enough filters or not enough filters... ART doesn't tell us how many it wants... just how many it has used.

Still a bit messy in the 20Hz to 35Hz and 120Hz to 150Hz ranges... I have a 130Hz resonance which is pernicious.. and performance of all my speakers is constrained below 30Hz... although the front L/R and subs should be capable of 24Hz... the Dirac measurements sets their FSL at 28.6Hz... I think the peaks and troughs tend to trick Dirac's frequency limitation algorhythm - I would have hoped for better decay control in the 24Hz to 35Hz range.... perhaps more tweaking is warranted.

 

[dukeside]

In a perfect room, would these waterfalls just drop almost straight off or close to 100ms or something?

 

[dlaloum]

In a perfect room, would these waterfalls just drop almost straight off or close to 100ms or something?

Yes - in a true anechoic perfect space...

But these measurements are damn good... (thank you ART!)

 

[dukeside]

Do you have any photos of the room or any before/after? It would be neat to see that alongside the data.

 

[dlaloum]

Edited update: progress from first trial to current... - all measurements are for L (got rid of the R accidental inclusion!)

First try is 5.1.4 (single sub) - everything pretty much default ART... not so impressive (clarity was already improved over DL however - already substantial decay control here!) - given the default has all support active, I assume it was using all 98 / 98 available filters - but I didn't check or record that paramters at this stage (I wonder how good this more basic configuration could be if I used the paramters more akin to the bottom one in this posting!)

View attachment 498039

Then my initial (semi default) ART setup, this time with 2 subs...(5.2.4) vs my 2nd try improved setup...

The first attempt was clearly suffering from a number of issues, where the 2nd attempt is much better sorted

Here are the actual settings with their associated charts:

Initial Setup configuration: groups chosen by Dirac, Levels set by Dirac, FSL manually adjusted for Group 1, Group 3, Group 6 on G1,2,3) - All FSH @ 150Hz
	Supported By Group (sup level, FSL)	Filters Used	98 / 98
	Group 1	Group 2	Group 3	Group 4	Group 5	Group 6
Group 1 L/R	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz
Group 2 C	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz
Group 3 SL, SR	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz
Group 4 FLH, FRH	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 29Hz
Group 5 SLH, SRH	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 29Hz
Group 6 SW1, SW2	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz

View attachment 498040


And here is where I started to take charge of the support groups, and switch some of them off - one noticeable change - is I went from using all 98 filters, down to only 51 filters (perhaps the default configuration as above, needed more than the 98 filters available??)

	2nd Setup configuration: groups chosen by Dirac, Levels set by Dirac, FSL manually adjusted for Group 1, Group 3, Group 6 Support allocations edited, All FSH @150Hz
	Supported By Group (sup level, FSL)	Filters used	51 / 98
	Group 1	Group 2	Group 3	Group 4	Group 5	Group 6
Group 1 L/R	-18, 30Hz		-18, 35Hz			-18, 24Hz
Group 2 C	-18, 30Hz		-18, 35Hz			-18, 24Hz
Group 3 SL, SR	-18, 30Hz		-18, 35Hz			-18, 24Hz
Group 4 FLH, FRH	-18, 30Hz			-18, 20Hz		-18, 29Hz
Group 5 SLH, SRH			-18, 35Hz		-18, 20Hz	-18, 29Hz
Group 6 SW1, SW2	-18, 30Hz		-18, 35Hz			-18, 24Hz

View attachment 498041

Well isn't that interesting !! - it still isn't perfect (what is!) - but this is a very nicely controlled environment - massively improved, and yet there is less support specified...
My best guess is that this configuration was within the filter capabilites, whereas the previous config, required more filters than I had available - hence it let through more flaws!!

Perhaps we should keep an eye on the filters used field... avoid any configuration that maxes out the filters (as we cannot tell whether it is just enough filters or not enough filters... ART doesn't tell us how many it wants... just how many it has used.

Still a bit messy in the 20Hz to 35Hz and 120Hz to 150Hz ranges... I have a 130Hz resonance which is pernicious.. and performance of all my speakers is constrained below 30Hz... although the front L/R and subs should be capable of 24Hz... the Dirac measurements sets their FSL at 28.6Hz... I think the peaks and troughs tend to trick Dirac's frequency limitation algorhythm - I would have hoped for better decay control in the 24Hz to 35Hz range.... perhaps more tweaking is warranted.

More comments and another experiment:

I added the center channel back in for support of the Fronts and the subs - this is my new parameter set:

	Extend FSL of Group1 to 24Hz (see if 20 to 35Hz zone can be improved)
	Update to 2nd setup, bringing in C channel to support L & R Lightly
	Supported By Group (sup level, FSL)	Filters used	54 / 98
	Group 1	Group 2	Group 3	Group 4	Group 5	Group 6
Group 1 L/R	-18, 24Hz	-5, 60Hz	-18, 35Hz			-18, 24Hz
Group 2 C	-18, 24Hz	-18, 50Hz	-18, 35Hz			-18, 24Hz
Group 3 SL, SR	-18, 24Hz		-18, 35Hz			-18, 24Hz
Group 4 FLH, FRH	-18, 24Hz			-18, 20Hz		-18, 29Hz
Group 5 SLH, SRH			-18, 35Hz		-18, 20Hz	-18, 29Hz
Group 6 SW1, SW2	-18, 24Hz	-5, 60Hz	-18, 35Hz			-18, 24Hz

Better extension down around 26Hz now, but I gained an 80hz dip... (how? why? ) slight improvement up around 130 to 150Hz

Interesting to note that THD recorded peaks at around 6% THD 26hz, but drops off quickly as F rises, 35Hz is 1%, 80Hz = 0.9%

Looking at all this I feel like the demonstrable lack of decay control below 40Hz may best coincide with the limitations of the Surround speakers which have F/R down to 35Hz... so that is where their contribution phases out, but also where their distortion is rising - and therefore the ability to apply control is becoming more limited

I do wonder whether we should also be considering rising distortion as being a limiter to ART .... my hypothesis: - where the distortion is too high, it will definitely throw off the calculated response - resulting in higher decay

Perhaps the limit of a speakers contribution range should be gauged not only by SPL frequency response, but also by distortion contribution - and perhaps a speakers limit is when distoriton rises past 1.5%... - which for my setup seems to be 40Hz

 

[Dj7675]

More comments and another experiment:

I added the center channel back in for support of the Fronts and the subs - this is my new parameter set:

	Extend FSL of Group1 to 24Hz (see if 20 to 35Hz zone can be improved)
	Update to 2nd setup, bringing in C channel to support L & R Lightly
	Supported By Group (sup level, FSL)	Filters used	54 / 98
	Group 1	Group 2	Group 3	Group 4	Group 5	Group 6
Group 1 L/R	-18, 24Hz	-5, 60Hz	-18, 35Hz			-18, 24Hz
Group 2 C	-18, 24Hz	-18, 50Hz	-18, 35Hz			-18, 24Hz
Group 3 SL, SR	-18, 24Hz		-18, 35Hz			-18, 24Hz
Group 4 FLH, FRH	-18, 24Hz			-18, 20Hz		-18, 29Hz
Group 5 SLH, SRH			-18, 35Hz		-18, 20Hz	-18, 29Hz
Group 6 SW1, SW2	-18, 24Hz	-5, 60Hz	-18, 35Hz			-18, 24Hz

View attachment 498043

Better extension down around 26Hz now, but I gained an 80hz dip... (how? why? ) slight improvement up around 130 to 150Hz

Interesting to note that THD recorded peaks at around 6% THD 26hz, but drops off quickly as F rises, 35Hz is 1%, 80Hz = 0.9%

Looking at all this I feel like the demonstrable lack of decay control below 40Hz may best coincide with the limitations of the Surround speakers which have F/R down to 35Hz... so that is where their contribution phases out, but also where their distortion is rising - and therefore the ability to apply control is becoming more limited

View attachment 498044

I do wonder whether we should also be considering rising distortion as being a limiter to ART .... my hypothesis: - where the distortion is too high, it will definitely throw off the calculated response - resulting in higher decay

Perhaps the limit of a speakers contribution range should be gauged not only by SPL frequency response, but also by distortion contribution - and perhaps a speakers limit is when distoriton rises past 1.5%... - which for my setup seems to be 40Hz

Fun watching your experiments and posting detailed measurements and your settings. I had always thought that Dirac should “help” consumers with their low frequency support setttings. Many speakers will shoot up in distortion and at some point just shouldn’t be used at those lower frequencies. They could even have a default THD % that is acceptable and maybe editable on the user end in advanced settings. For those testing and measuring ART like you have done, also showing distortion is a great idea.

 

[ban25]

More comments and another experiment:

I added the center channel back in for support of the Fronts and the subs - this is my new parameter set:

	Extend FSL of Group1 to 24Hz (see if 20 to 35Hz zone can be improved)
	Update to 2nd setup, bringing in C channel to support L & R Lightly
	Supported By Group (sup level, FSL)	Filters used	54 / 98
	Group 1	Group 2	Group 3	Group 4	Group 5	Group 6
Group 1 L/R	-18, 24Hz	-5, 60Hz	-18, 35Hz			-18, 24Hz
Group 2 C	-18, 24Hz	-18, 50Hz	-18, 35Hz			-18, 24Hz
Group 3 SL, SR	-18, 24Hz		-18, 35Hz			-18, 24Hz
Group 4 FLH, FRH	-18, 24Hz			-18, 20Hz		-18, 29Hz
Group 5 SLH, SRH			-18, 35Hz		-18, 20Hz	-18, 29Hz
Group 6 SW1, SW2	-18, 24Hz	-5, 60Hz	-18, 35Hz			-18, 24Hz

View attachment 498043

Better extension down around 26Hz now, but I gained an 80hz dip... (how? why? ) slight improvement up around 130 to 150Hz

Interesting to note that THD recorded peaks at around 6% THD 26hz, but drops off quickly as F rises, 35Hz is 1%, 80Hz = 0.9%

Looking at all this I feel like the demonstrable lack of decay control below 40Hz may best coincide with the limitations of the Surround speakers which have F/R down to 35Hz... so that is where their contribution phases out, but also where their distortion is rising - and therefore the ability to apply control is becoming more limited

View attachment 498044

I do wonder whether we should also be considering rising distortion as being a limiter to ART .... my hypothesis: - where the distortion is too high, it will definitely throw off the calculated response - resulting in higher decay

Perhaps the limit of a speakers contribution range should be gauged not only by SPL frequency response, but also by distortion contribution - and perhaps a speakers limit is when distoriton rises past 1.5%... - which for my setup seems to be 40Hz

The Center channel is the most active in home theater scenarios, so I think it's probably best *not* to use it for support of other channels. In fact, several folks have seen best results using only subwoofers as support because they have the most output <150 Hz.

 

[dlaloum]

More learnings...

The full range surrounds (B&O Penta) with F reach down to 35Hz, are a ported design, so the distortion rises down around 35 to 40Hz (port related?)

Limiting their support to 40Hz cleans up the decay around the 35Hz to 40Hz range....
I have long hypothesised that ported designs will have limitations when used with ART... driven by port related effects (distortion)
Looks like the measurements/experiments are bearing this out!

All I did here, was raise the FSL for the S channel from 35Hz to 40Hz result is:

Note I am still "self supporting" the surround channel down to 35Hz - probably the reason for that raised decay just below 35Hz

Still have the 80Hz dip... have to go back to the measurments without that dip, and experiment to identify the cause

 

[kawauso]

Edited update: progress from first trial to current... - all measurements are for L (got rid of the R accidental inclusion!)

First try is 5.1.4 (single sub) - everything pretty much default ART... not so impressive (clarity was already improved over DL however - already substantial decay control here!) - given the default has all support active, I assume it was using all 98 / 98 available filters - but I didn't check or record that paramters at this stage (I wonder how good this more basic configuration could be if I used the paramters more akin to the bottom one in this posting!)

View attachment 498039

Then my initial (semi default) ART setup, this time with 2 subs...(5.2.4) vs my 2nd try improved setup...

The first attempt was clearly suffering from a number of issues, where the 2nd attempt is much better sorted

Here are the actual settings with their associated charts:

Initial Setup configuration: groups chosen by Dirac, Levels set by Dirac, FSL manually adjusted for Group 1, Group 3, Group 6 on G1,2,3) - All FSH @ 150Hz
	Supported By Group (sup level, FSL)	Filters Used	98 / 98
	Group 1	Group 2	Group 3	Group 4	Group 5	Group 6
Group 1 L/R	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz
Group 2 C	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz
Group 3 SL, SR	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz
Group 4 FLH, FRH	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 29Hz
Group 5 SLH, SRH	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 29Hz
Group 6 SW1, SW2	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz

View attachment 498040


And here is where I started to take charge of the support groups, and switch some of them off - one noticeable change - is I went from using all 98 filters, down to only 51 filters (perhaps the default configuration as above, needed more than the 98 filters available??)

	2nd Setup configuration: groups chosen by Dirac, Levels set by Dirac, FSL manually adjusted for Group 1, Group 3, Group 6 Support allocations edited, All FSH @150Hz
	Supported By Group (sup level, FSL)	Filters used	51 / 98
	Group 1	Group 2	Group 3	Group 4	Group 5	Group 6
Group 1 L/R	-18, 30Hz		-18, 35Hz			-18, 24Hz
Group 2 C	-18, 30Hz		-18, 35Hz			-18, 24Hz
Group 3 SL, SR	-18, 30Hz		-18, 35Hz			-18, 24Hz
Group 4 FLH, FRH	-18, 30Hz			-18, 20Hz		-18, 29Hz
Group 5 SLH, SRH			-18, 35Hz		-18, 20Hz	-18, 29Hz
Group 6 SW1, SW2	-18, 30Hz		-18, 35Hz			-18, 24Hz

View attachment 498041

Well isn't that interesting !! - it still isn't perfect (what is!) - but this is a very nicely controlled environment - massively improved, and yet there is less support specified...
My best guess is that this configuration was within the filter capabilites, whereas the previous config, required more filters than I had available - hence it let through more flaws!!

Perhaps we should keep an eye on the filters used field... avoid any configuration that maxes out the filters (as we cannot tell whether it is just enough filters or not enough filters... ART doesn't tell us how many it wants... just how many it has used.

Still a bit messy in the 20Hz to 35Hz and 120Hz to 150Hz ranges... I have a 130Hz resonance which is pernicious.. and performance of all my speakers is constrained below 30Hz... although the front L/R and subs should be capable of 24Hz... the Dirac measurements sets their FSL at 28.6Hz... I think the peaks and troughs tend to trick Dirac's frequency limitation algorhythm - I would have hoped for better decay control in the 24Hz to 35Hz range.... perhaps more tweaking is warranted.

In my 7.4.6 setup, I found that reducing the number of support speakers from the default 93/94 to 69, limiting support to only the same group and the subwoofers, resulted in better FR and decay at the MLP.

In my case, the default results were not particularly problematic, but this more conservative support configuration performed better than the default in all five of my rooms. The results are more than sufficient, so I am using this configuration.

In theory, having more support speakers should be more effective. However, it is unclear whether this behavior is due to a bug in ART or whether ART is prioritizing results away from the MLP. For now, I am making decisions based on the FR at the MLP.

dlaloum, if you have added a subwoofer and there is a performance difference between the subs, it may be effective to separate them into different subwoofer groups. For example:
• SW1: 20–150 Hz (Infra Bass ON)
• SW2: 25–150 Hz (Infra Bass OFF)

Could you also check the results with the second subwoofer turned off? Even with a single subwoofer, you should be able to confirm that the results are better than your initial default ART measurement. While you will likely use two subs for normal operation, a single sub can already produce very good results.

 

[dlaloum]

The Center channel is the most active in home theater scenarios, so I think it's probably best *not* to use it for support of other channels. In fact, several folks have seen best results using only subwoofers as support because they have the most output <150 Hz.

That's why my earlier config excluded C support - but I wanted to see what happened when I added in light C support (-5db)
I did a couple of changes alongside adding C support - extended the mains and subs down to 24Hz - and it did smooth and extend the 35Hz to 20Hz region - but also added in an 80Hz dip.... which in the overall balance of things, is probably a step backwards!

The biggest single improvement was adding the 2nd sub into the room (positioned at Rear Left, near the Surround Left speaker), SW1 is at Front Right next to the Main R.

So yes there is perhaps something to say for using only subs for support... on the other hand my mains and my subs use the exact same driver.... capabilities should be very close (with subs having greater internal volume being the primary difference)

The other notable improvement, was dropping down to fewer filters - running everything to everything support in a 5.2.4 setup probably requires more than the available 98 filters. (might work well on something like the Storm AVP, where more filters are available...)

I'm not done yet!

 

[dlaloum]

In my 7.4.6 setup, I found that reducing the number of support speakers from the default 93/94 to 69, limiting support to only the same group and the subwoofers, resulted in better FR and decay at the MLP.

In my case, the default results were not particularly problematic, but this more conservative support configuration performed better than the default in all five of my rooms. The results are more than sufficient, so I am using this configuration.

In theory, having more support speakers should be more effective. However, it is unclear whether this behavior is due to a bug in ART or whether ART is prioritizing results away from the MLP. For now, I am making decisions based on the FR at the MLP.

dlaloum, if you have added a subwoofer and there is a performance difference between the subs, it may be effective to separate them into different subwoofer groups. For example:
• SW1: 20–150 Hz (Infra Bass ON)
• SW2: 25–150 Hz (Infra Bass OFF)

Could you also check the results with the second subwoofer turned off? Even with a single subwoofer, you should be able to confirm that the results are better than your initial default ART measurement. While you will likely use two subs for normal operation, a single sub can already produce very good results.

My 2nd sub is an exactly model to the 1st sub... differences would be due to marginal output level difference (sub gain is an analog knob... hard to match precisely!) - and of course, positioning.

I'm not sure there is much point to having Infra on - as both my subs and my mains are at -14db @20Hz....

Where is that sneaky infra bass checkbox?

Does it make sense trying the SW2 off, given the measurements (and therefore all subsequent ART calculations) are based on it being in the setup... I would assume that it would require a new set of measurements... (on 2nd thought, I could go back to my 5.1.4 measurements.... hmmm...)

With identical subs - does it make sense seperating them? (when the difference is only their positions in the room?)

 

[kawauso]

My 2nd sub is an exactly model to the 1st sub... differences would be due to marginal output level difference (sub gain is an analog knob... hard to match precisely!) - and of course, positioning.

I'm not sure there is much point to having Infra on - as both my subs and my mains are at -14db @20Hz....

Where is that sneaky infra bass checkbox?

Does it make sense trying the SW2 off, given the measurements (and therefore all subsequent ART calculations) are based on it being in the setup... I would assume that it would require a new set of measurements... (on 2nd thought, I could go back to my 5.1.4 measurements.... hmmm...)

With identical subs - does it make sense seperating them? (when the difference is only their positions in the room?)

The Infra Bass checkbox, which is a bit hard to find, is located here. It only becomes available when you lower the bottom of the support range down to 20 Hz.
In dlaloum’s setup, unless the room gain is causing a significant extension in the very low frequencies, there does not seem to be much need to enable it.
In dlaloum’s environment, separating the subwoofers will probably not make much difference, but you can try turning off SW2. If you drag Group 6 with the mouse, a new Group 7 will be created and the sub will be separated. By not assigning Group 7 to any supported group, you can easily test the system with SW2 effectively turned off.

That's why my earlier config excluded C support - but I wanted to see what happened when I added in light C support (-5db)
I did a couple of changes alongside adding C support - extended the mains and subs down to 24Hz - and it did smooth and extend the 35Hz to 20Hz region - but also added in an 80Hz dip.... which in the overall balance of things, is probably a step backwards!

The biggest single improvement was adding the 2nd sub into the room (positioned at Rear Left, near the Surround Left speaker), SW1 is at Front Right next to the Main R.

So yes there is perhaps something to say for using only subs for support... on the other hand my mains and my subs use the exact same driver.... capabilities should be very close (with subs having greater internal volume being the primary difference)

The other notable improvement, was dropping down to fewer filters - running everything to everything support in a 5.2.4 setup probably requires more than the available 98 filters. (might work well on something like the Storm AVP, where more filters are available...)

I'm not done yet!

With my front speakers, the JBL S9900, using them in the 30–150 Hz range works fine. However, I found that using them in the 50–150 Hz range did not result in much audible difference, and since it provides more headroom, I am using the 50–150 Hz setting.
When set to 80–150 Hz, the sense of imaging felt slightly different, but if prioritizing headroom for the front speakers, this setting is not a bad option.

 

[KenMasters]

For those testing and measuring ART like you have done, also showing distortion is a great idea.

Good guidance, keeping an eye on distortion has been a big help in my decision making process.

That's why my earlier config excluded C support - but I wanted to see what happened when I added in light C support (-5db)...

Something I've noted, but may not be universally true, is that the scaling doesn't appear particularly nuanced - less a sloping scale, more a delineation of hierarchy. For example, if I set two subs both at -18 I get one result, if I set one as -24 and the other as -18 I get another result, but for values inbetween the result is much the same as -18 vs -24.

Which makes the three value guidance of Dirac a good template to follow I think.

Do you have any photos of the room or any before/after? It would be neat to see that alongside the data.

Basic PEQ

Dirac ART

And you can see my room in its entirety here

 

[smygolf]

More learnings...

The full range surrounds (B&O Penta) with F reach down to 35Hz, are a ported design, so the distortion rises down around 35 to 40Hz (port related?)

Limiting their support to 40Hz cleans up the decay around the 35Hz to 40Hz range....
I have long hypothesised that ported designs will have limitations when used with ART... driven by port related effects (distortion)
Looks like the measurements/experiments are bearing this out!

All I did here, was raise the FSL for the S channel from 35Hz to 40Hz result is:

View attachment 498046

Note I am still "self supporting" the surround channel down to 35Hz - probably the reason for that raised decay just below 35Hz

Still have the 80Hz dip... have to go back to the measurments without that dip, and experiment to identify the cause

Try to keep the scale the same when you compare
This one is between 50 - 90dB others 40-90dB

And obvious always have the same volume

Lastly, always check dB difference at peaks - highest to the lowest
Pic above and at 32hz - signal looks to be at 73dB and the decay lowest 52dB = 21dB difference
Thinking on that you have the default layout it rings a awful lot - more then 500ms

As a comparison, the ex below is not sweeped at the volume i normaly sweeps, it is a tad lower
Same margin as in your measurement from signal to window top - 15dB and 40dB scale

And now with 60dB scale - still 15dB headroom and from 15 - 150hz

 

[janbth]

Edited update: progress from first trial to current... - all measurements are for L (got rid of the R accidental inclusion!)

First try is 5.1.4 (single sub) - everything pretty much default ART... not so impressive (clarity was already improved over DL however - already substantial decay control here!) - given the default has all support active, I assume it was using all 98 / 98 available filters - but I didn't check or record that paramters at this stage (I wonder how good this more basic configuration could be if I used the paramters more akin to the bottom one in this posting!)

View attachment 498039

Then my initial (semi default) ART setup, this time with 2 subs...(5.2.4) vs my 2nd try improved setup...

The first attempt was clearly suffering from a number of issues, where the 2nd attempt is much better sorted

Here are the actual settings with their associated charts:

Initial Setup configuration: groups chosen by Dirac, Levels set by Dirac, FSL manually adjusted for Group 1, Group 3, Group 6 on G1,2,3) - All FSH @ 150Hz
	Supported By Group (sup level, FSL)	Filters Used	98 / 98
	Group 1	Group 2	Group 3	Group 4	Group 5	Group 6
Group 1 L/R	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz
Group 2 C	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz
Group 3 SL, SR	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz
Group 4 FLH, FRH	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 29Hz
Group 5 SLH, SRH	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 29Hz
Group 6 SW1, SW2	-18, 30Hz	-18, 50Hz	-18, 35Hz	-18, 20Hz	-18, 20Hz	-18, 24Hz

View attachment 498040


And here is where I started to take charge of the support groups, and switch some of them off - one noticeable change - is I went from using all 98 filters, down to only 51 filters (perhaps the default configuration as above, needed more than the 98 filters available??)

	2nd Setup configuration: groups chosen by Dirac, Levels set by Dirac, FSL manually adjusted for Group 1, Group 3, Group 6 Support allocations edited, All FSH @150Hz
	Supported By Group (sup level, FSL)	Filters used	51 / 98
	Group 1	Group 2	Group 3	Group 4	Group 5	Group 6
Group 1 L/R	-18, 30Hz		-18, 35Hz			-18, 24Hz
Group 2 C	-18, 30Hz		-18, 35Hz			-18, 24Hz
Group 3 SL, SR	-18, 30Hz		-18, 35Hz			-18, 24Hz
Group 4 FLH, FRH	-18, 30Hz			-18, 20Hz		-18, 29Hz
Group 5 SLH, SRH			-18, 35Hz		-18, 20Hz	-18, 29Hz
Group 6 SW1, SW2	-18, 30Hz		-18, 35Hz			-18, 24Hz

View attachment 498041

Well isn't that interesting !! - it still isn't perfect (what is!) - but this is a very nicely controlled environment - massively improved, and yet there is less support specified...
My best guess is that this configuration was within the filter capabilites, whereas the previous config, required more filters than I had available - hence it let through more flaws!!

Perhaps we should keep an eye on the filters used field... avoid any configuration that maxes out the filters (as we cannot tell whether it is just enough filters or not enough filters... ART doesn't tell us how many it wants... just how many it has used.

Still a bit messy in the 20Hz to 35Hz and 120Hz to 150Hz ranges... I have a 130Hz resonance which is pernicious.. and performance of all my speakers is constrained below 30Hz... although the front L/R and subs should be capable of 24Hz... the Dirac measurements sets their FSL at 28.6Hz... I think the peaks and troughs tend to trick Dirac's frequency limitation algorhythm - I would have hoped for better decay control in the 24Hz to 35Hz range.... perhaps more tweaking is warranted.

That is an impressive improvement - thanks for taking the time to do this! I would still be interested in seeing L+R, though (L and R at the same time). Do you happen to have the measurements around?