Active Room Treatment (ART) by Dirac

[Oddball]

I think Timoj is right as I've read the same thing from the big Dirac thread over at AVS. Not only that, I think there is a preference for Sub1's timing (LFE) to more or less match the center channel. Dirac hasn't explained why the center channel is the lucky one but at least it is a basis as opposed to none.

Both of you guys are wrong in substance. And please don't quote AVS as authority as things happening there are entirely unscientific and quite chaotic.

The timing sequences are correct though. You can choose the subs timing reference, and C seems to be most appropriate. If you don't think so then you have option to chose another reference. Then, ART will align Sub 1, aka LFE sub to that and work from there to align other subs.

This has absolutely nothing to do with how much will your LFE sub aka sub 1 contribute to the LFE or supporting channels vs other subs. There are plenty of graphs including Dirac filters for LFE to support that. All that one has to do is to go back and check.

I feel like people are not really reading through this thread - which might understandable as it is mighty long. But then I would have hoped that at least advanced members would have picked up basic notions and not introduce further confusion based on a side note that really means nothing in substance.

 

[R4pture]

"The selected speaker will serve as reference for the time target of the LFE channel. Thereby, LFE content will be optimised to propagate through the listening area as if it was played from the selected reference speaker."

+

"The selected LFE reference can be used as a tuning parameter, finding the reference speaker that best fits the desired LFE content propagation.

• A full-range speaker is preferred over a subwoofer as LFE reference, since the impulse response is then better defined across the listening area.
• Not selecting one of the subwoofers also potentially improves the LFE load distribution across all subwoofers. By having a speaker outside the set of subwoofers defining the time target, one particular subwoofer will not be contributing more than the others because of its response fully matching the time target." - From the official Dirac Art Setup Guide.

Thought I should post this for reference.

 

[Oddball]

"The selected speaker will serve as reference for the time target of the LFE channel. Thereby, LFE content will be optimised to propagate through the listening area as if it was played from the selected reference speaker."

+

"The selected LFE reference can be used as a tuning parameter, finding the reference speaker that best fits the desired LFE content propagation.

• A full-range speaker is preferred over a subwoofer as LFE reference, since the impulse response is then better defined across the listening area.
• Not selecting one of the subwoofers also potentially improves the LFE load distribution across all subwoofers. By having a speaker outside the set of subwoofers defining the time target, one particular subwoofer will not be contributing more than the others because of its response fully matching the time target." - From the official Dirac Art Setup Guide.

Thought I should post this for reference.

Thanks for posting this.

The previous default timing reference to C was L by default before they enabled C. I must admit that I never tried any of the subs as timing reference, so there could be something in what Dirac is saying. I would say that it is clear that if bed channel timing reference is chosen, looks like all will fall in place as it should.

 

[Jon AA]

Dirac hasn't explained why the center channel is the lucky one but at least it is a basis as opposed to none.

It makes sense as the LFE channel is supposed to "sound like" it's coming from the screen, so the center channel is a good choice if it has that effect. Various Dolby specs require the LFE sub(s) be located underneath the screen, close to the center.

 

[R4pture]

Thanks for posting this.

The previous default timing reference to C was L by default before they enabled C. I must admit that I never tried any of the subs as timing reference, so there could be something in what Dirac is saying. I would say that it is clear that if bed channel timing reference is chosen, looks like all will fall in place as it should.

Back in October 2025 I mistakenly set the LFE-Reference from C to Sub1. I felt it was the right move. After some reading I changed it back to C.

I was unable to spot any differences between Sub1 (right front corner) or C as LFE-Reference. I mostly focused on differences in pressure build-up between both my ears for each seat.
Aside from an unfortunate error in my brain/ears at around 35-40 Hz, there was no difference between both reference points. But I have 4 subs. Back when I had only one sub (Audyssey correction), I experienced pressure differences between my ears whenever I moved closer to a sidewall. Maybe someone with just one sub will hear/feel stronger differences changing the LFE-Reference when focusing just on a single local position (EDIT: I must say I am not sure if you can change that with different timings? Maybe it is mostly relevant for people with multiple subs?).

Still, I agree with you and Dirac. Higher frequencies should generally be more reliable as a timing reference across multiple measurement points than low frequencies (unless the LFE-Reference speaker show massive problems in context of all measurement points). Any non sub speaker should be better suited for LFE-Reference. At least in my theory.

 

[TimoJ]

With all respect TimoJ, this is a completely wrong answer.

I don’t think so. Only one speaker per channel (in this case LFE and sub1), other speakers/subs support it.

 

[KenMasters]

I don’t think so. Only one speaker per channel (in this case LFE and sub1), other speakers/subs support it.

There is only one LFE group but there can be more than one subwoofer. The subwoofer attached to sub output one gets the "LFE" designation - this is arbitrary, it doesn't mean subwoofer 1 is treated any differently than the others. They are all LFE.

Nothing about this changes when you separate one or more subwoofers from the LFE group. Moving a sub out of the group only serves to allow for the changing of its individual parameters if you feel it's performance characteristics diverge significantly enough to benefit from different settings to those of the rest of the group.

 

[TimoJ]

There is only one LFE group but there can be more than one subwoofer.

But only sub1 gets the full LFE signal, others support it.

 

[smygolf]

To add to the discussion on LFE, my experience in my system and others i have helped.

Sub 1 always play louder, between 5 - 12dB then the supporting subs.
In my case sub 1-2 sits in a baffle wall and have close to identical response but nr 2 is -10dB

IMO then sub 1 is indeed LFE and the rest are pure support

 

[KenMasters]

But only sub1 gets the full LFE signal, others support it.

They all get the same signal, applied filters differ based on measured response and interaction thereof.

To add to the discussion on LFE, my experience in my system and others i have helped.

Sub 1 always play louder, between 5 - 12dB then the supporting subs.
In my case sub 1-2 sits in a baffle wall and have close to identical response but nr 2 is -10dB

IMO then sub 1 is indeed LFE and the rest are pure support

That's just how things shook out in your room and the others you have seen. Not the case in my room, generally quite even, in fact it's sub two and three that have the highest peaks.

EDIT: I had a similar idea about the LFE grouping as you guys, but others in this thread challenged me on that. I then tried changing groupings and inputs and found it didn't make a lick of difference. All subs are LFE.

Think about it, why would ART fundamentally compromise bass performance by prioritising a single sub? How would that be optimal?

 

[smygolf]

They all get the same signal, applied filters differ based on measured response. and interaction thereof.


That's just how things shook out in your room and the others you have seen. Not the case in my room, generally quite even, in fact it's sub three that has the highest peaks.

EDIT: I thought the same as you guys at the start, but others in this thread challenged me on that. I then tried changing groupings and inputs and found it didn't make a lick of difference. All subs are LFE.

I think we have different views of what LFE are with ART really is.
If i let my LCR support the subs, then they also are LFE in a sense.
Same as the subs would be LCR via BM.

And for the record, i do not really care, ART works and i am happy

 

[KenMasters]

I think we have different views of what LFE are with ART really is.

LFE as far as the ART interface is concerned is a label applied to sub output 1 because that is logical and convenient for grouping purposes. All subwoofer outputs output the same signal and are then optimised to work in concert.

If i let my LCR support the subs, then they also are LFE in a sense.
Same as the subs would be LCR via BM

The signal sent to the LCRs is not the same as what is sent to the subwoofer outs. Support is something the speakers do on top of the signal they receive.

You can say that there is one chief LFE signal, but no sub is the chief LFE reproducer - they work together/support one another for optimise output.

It's Bass Control, but taken a step further through the interoperational optimisation of individual speakers and subs.

 

[OCA]

Just a heads up on MIMO behavior: every support speaker gets the full audio signal, so all support groups will see the complete LFE track. The designated LFE sub only handles the timing alignment of the combined signal. However, the filters applied to each speaker will differ. I believe ART's solver uses a "greedy" algorithm which works through available support speakers sequentially rather than optimizing everything at once. Because of this, you'll need to manually tweak support levels, adjust frequency cutoffs, or split channel groups to ensure your most capable drivers do the heavy lifting.

 

[TimoJ]

All subwoofer outputs output the same signal and are then optimised to work in concert.

Have you checked Dirac LFE-channel subwoofer filter responses? In my system, only Sub 1 has response up to 120 Hz, the others do not.
Previously, before the fixed 120 Hz filter was added, it was very obvious that only Sub 1 was handling the LFE channel.

 

[smygolf]

LFE as far as the ART interface is concerned is a label applied to sub output 1 because that is logical and convenient for grouping purposes. All subwoofer outputs output the same signal and are then optimised to work in concert.



The signal sent to the LCRs is not the same as what is sent to the subwoofer outs. Support is something the speakers do on top of the signal they receive.

You can say that their is one chief LFE signal, but no sub is the chief LFE reproducer - they work together/support one another for optimise output.

It's Bass Control, but taken a step further through the interoperational optimisation of individual speakers and subs.

How ART truly works only Dirac knows.
And Yes support lvl on subs are higher the support lvl on LCR.

Your take on Chief LFE sounds plausible for sure.

 

[KenMasters]

Have you checked Dirac LFE-channel subwoofer filter responses? In my system, only Sub 1 has response up to 120 Hz, the others do not.
Previously, before the fixed 120 Hz filter was added, it was very obvious that only Sub 1 was handling the LFE channel.

Filters and estimated response (Sub 1 filter highlighted):

Natural sub responses:

Post-Cal measurements of individual subs and combined response via LFE channel:

To me it looks like a considered distribution of work amongst the subs to smooth the LFE response.

 

[Oddball]

Filters and estimated response (Sub 1 filter highlighted):

View attachment 541029

Natural sub responses:

View attachment 541026

Post-Cal measurements of individual subs and combined response via LFE channel:

View attachment 541045

To me it looks like a considered distribution of work amongst the subs to smooth the LFE response.

Quite similar to my experience where subs 3 and 4 take some heavier duty than sub 1.

Using the terminology like one is LFE and others are support is confusing even if it comes from Dirac. In reality, ART will best distribute LFE signal between the supporting group. How will that end depends on the room and system, but it certainly does not mean that support subs are "waisted" and that you loose SPL with ART, which are the conclusions that some less knowledgeable people draw from this confusing premisse.

Again, can't believe that we need to discuss things that were solved in the past. As noted, have a feeling that only few people around here are actually following and understanding the thread.

 

[OCA]

looks like a considered distribution of work amongst the subs to smooth the LFE response.

Exactly, support subs get the 150Hz sharp LPF like all support groups but otherwise are just as active as Sub 1.

 

[RenPa]

But only sub1 gets the full LFE signal, others support it.

Could this prove it? In REW, select the LFE channel, turn off sub 1, and take a measurement. No sound should come out of the other subs, right? I answer to my question: don't think so because sounds will come from the spporting subs to sub 1...i think...

 

[Capsa]

If Dirac or the support-level user settings determine that subs 2/3/4 contribute more (or less) than sub 1 to the LFE signal, aren't they still just supports subs ?
"Subs 2/3/4 are support" and "the LFE signal is distributed in an optimized manner among all the subs" seem to mean the same thing, with ART.

But only sub1 gets the full LFE signal, others support it.

I think you would agree that if circumstances require it (such as suboptimal frequency response), supports subs can play LFE signal louder than sub 1.

It seems to me to be more of a misunderstanding of the terms than anything else this disagreement.