Active Room Treatment (ART) by Dirac

[Dovydas]

From my own test i guess it works something like this.

Minus value mostly effects the decay, more minus = lower decay.
But the volume the sub will be playing is roughly the same.

Correct, at leats it was the same in my experience as well. I had to reduce SW support in mains group and in SW group because of huge dip. It helped with a dip but decay times became slightly worse. I need to play with subs location or add a third one to use more support.

 

[fzst]

From my own test i guess it works something like this.

Minus value mostly effects the decay, more minus = lower decay.
But the volume the sub will be playing is roughly the same.

Correct, at leats it was the same in my experience as well. I had to reduce SW support in mains group and in SW group because of huge dip. It helped with a dip but decay times became slightly worse. I need to play with subs location or add a third one to use more support.

Interesting, so a higher support level doesn't really mean that the sub/speaker contributes more to the main bass signal but rather puts out more "cancelling frequencies", which could potentially even hurt the FR? I definitely have to try -24dB or lower than -18dB in my setup as well to see if it has any measurable or audible difference. I suspect with my 4 subs, increasing support to -24dB probably wouldn't have an impact on the FR but who knows...

 

[lessthanjoey]

Interesting, so a higher support level doesn't really mean that the sub/speaker contributes more to the main bass signal but rather puts out more "cancelling frequencies", which could potentially even hurt the FR? I definitely have to try -24dB or lower than -18dB in my setup as well to see if it has any measurable or audible difference. I suspect with my 4 subs, increasing support to -24dB probably wouldn't have an impact on the FR but who knows...

It contributes more both constructively and destructively. The cancelation in this case sounds like an artifact but I agree with the earlier statement that I wouldn't take these numbers as any kind of golden. Dirac's defaults are reasonable but if you can measure with REW then you can experiment and adjust from there.

 

[smygolf]

I do not think it would impact the FR to much, after all DLBC nearly gets the exact same result.

 

[Oddball]

Interesting, so a higher support level doesn't really mean that the sub/speaker contributes more to the main bass signal but rather puts out more "cancelling frequencies", which could potentially even hurt the FR? I definitely have to try -24dB or lower than -18dB in my setup as well to see if it has any measurable or audible difference. I suspect with my 4 subs, increasing support to -24dB probably wouldn't have an impact on the FR but who knows...

As noted the level of support will go both ways - contributions and cancelations. I have tried many different combinations but at the end FQ response comes in pretty much the same - with +/- 1dB on average which is really not audible, at least for me. Decay is also relatively consistent, but it does get up in the 20hz range with more aggressive curves. How much we can hear vs. feel in 20hz range is questionable. I do have a +1.5dB bump in one of the curves at 20hz and still have a deliberation if I like it or don't.

As I noted before, ART has made me rethink some of the things I have done in the past. It just turns out so right even with the lightest touch (my room and my system - so caution of that) that I would for sure not be able to make it much better ironing out the graphs or making any really smart adjustments. I did try though.

I do see some issues with 2 channel presentation and agree that support levels should be limited in that case. Not for the subs, but for the other supporting speakers. For the most part 2 good towers and some subs should pull that off easily.

 

[kawauso]

As noted the level of support will go both ways - contributions and cancelations. I have tried many different combinations but at the end FQ response comes in pretty much the same - with +/- 1dB on average which is really not audible, at least for me. Decay is also relatively consistent, but it does get up in the 20hz range with more aggressive curves. How much we can hear vs. feel in 20hz range is questionable. I do have a +1.5dB bump in one of the curves at 20hz and still have a deliberation if I like it or don't.

As I noted before, ART has made me rethink some of the things I have done in the past. It just turns out so right even with the lightest touch (my room and my system - so caution of that) that I would for sure not be able to make it much better ironing out the graphs or making any really smart adjustments. I did try though.

I do see some issues with 2 channel presentation and agree that support levels should be limited in that case. Not for the subs, but for the other supporting speakers. For the most part 2 good towers and some subs should pull that off easily.

Yeah, changing the support levels doesn’t seem to make a big difference — even when I measure in REW, the changes are pretty subtle.
And honestly, even if there is some variation, it’s hard to tell how (or if) it actually translates into audible differences.
I’m quite happy with how my current setup sounds, but I saw someone on AVS mentioning a specific “good” value for support level, so it just made me curious.

 

[Oddball]

You found a rare jewel - so we should all thank you for that. Likes up for kawauso. She needs at least 10 of them for this find.

Well bumping this up as no 10 likes. Not my post but don't feel embarrassed to bump it up post. This is a significant find and you will not find it in D&M manuals. So it is a "hack" that is really important to understand the value of your system. If there is so little people that appreciate it, will be indicative of its value after all. But can't say that anyone with a flat panel would want to pass on this. Two centers aligned might be just what you were craving for.

 

[Fidji]

This could be relevant to tactile transducers shaking the room in certain single digit frequencies but otherwise I really doubt many of us here in ASR (30 years old or older) are still able to hear anything below 25Hz. Low bass filters are very expensive in terms of computation and delay. Audyssey uses 1024 filter taps and DLBC uses 2048. Despite lots of tricks like polyphase decimation etc., their min bass frequency resolution is around 5Hz. Receivers need to buffer 4K video for the filter length so there's a limitation to number of FIR taps that can be practically used as well. I personally don't find it too irresponsible that studio mixers are adding a high pass filter around 20-25Hz. You certainly have a great system but most people would just experience boomy resonances lingering over the mids and highs with such content and would also have to limit volume levels to avoid extra distortion.

I think that running to lowest common denominator is not the correct thing to do. [talking about whether it ok to cut off bass in movies]. Building on this logic, you could argue that we should cut all the material at 50Hz, as anyway most of the people have issues below this frequency. I don’t really mind what is happening to Netflix releases, but physical media and Kaleidascape are already niche thing, targeting enthusiasts market, so they should offer some added value compared to streaming.

Usually biggest mess is in 50-100Hz area anyway, frequencies below first modal are benign - no standing waves or resonances, maybe some furniture rattling. And you can EQ them very simply, usually with some low shelf. Addressing lowest frequencies is not so complex as it might seem, also there is much higher tolerance to FR ripples, GD etc.

Also RP22 from CEDIA is requiring FR at 115dB to go into teens, if you want to go beyond “ The minimum level of performance necessary to convey basic artistic intent.” And regarding audibility of lowest frequencies - just go to Ascendo showroom and you will understand the difference between the system playing capab;e fo 20Hz and single digits.

BTW - I would strongly recommend to every HT fan to download and print out RP22 - gold mine for understanding how to get performance out of HT, what is important and what not so much.

On D&M processors, you can actually create movie-specific house curves and save/load them via USB, but it’s a bit inconvenient in practice.
Even if a movie has reduced bass, it’s difficult to restore it just by creating a curve that exceeds the +12 dB limit within Dirac.
Even with something like a miniDSP, you’d need to apply filters to all speakers, otherwise the MIMO effect would be weakened.
Since you mentioned you’re not interested in Storm, I’d say you might want to consider another Trinnov for your second HT — it already does exactly what you’re looking for.

My situation - we will be moving to new house and I want to have 2 theaters - one top class, dedicated room, all bells and whistles, and second one in living room, kind of good enough to enjoy Netflix etc. I have the project already - in-wall DIY speakers, front wall will be Infinite Baffle, 115 in TV etc. What is important - my wife is THE Basshead of our family, this is why I ask about BEQ.

And i have total budget set up, as I can re-use or trade in everything from my current gear. And difference between Trinnov and AV10 is 2 24 in subs to our main HT. But let’s. see, maybe I will just smash 10dB low shelf on the subs directly and be done with it.

 

[Oddball]

I think that running to lowest common denominator is not the correct thing to do. [talking about whether it ok to cut off bass in movies]. Building on this logic, you could argue that we should cut all the material at 50Hz, as anyway most of the people have issues below this frequency. I don’t really mind what is happening to Netflix releases, but physical media and Kaleidascape are already niche thing, targeting enthusiasts market, so they should offer some added value compared to streaming.

Usually biggest mess is in 50-100Hz area anyway, frequencies below first modal are benign - no standing waves or resonances, maybe some furniture rattling. And you can EQ them very simply, usually with some low shelf. Addressing lowest frequencies is not so complex as it might seem, also there is much higher tolerance to FR ripples, GD etc.

Also RP22 from CEDIA is requiring FR at 115dB to go into teens, if you want to go beyond “ The minimum level of performance necessary to convey basic artistic intent.” And regarding audibility of lowest frequencies - just go to Ascendo showroom and you will understand the difference between the system playing capab;e fo 20Hz and single digits.

BTW - I would strongly recommend to every HT fan to download and print out RP22 - gold mine for understanding how to get performance out of HT, what is important and what not so much.

View attachment 489937



My situation - we will be moving to new house and I want to have 2 theaters - one top class, dedicated room, all bells and whistles, and second one in living room, kind of good enough to enjoy Netflix etc. I have the project already - in-wall DIY speakers, front wall will be Infinite Baffle, 115 in TV etc. What is important - my wife is THE Basshead of our family, this is why I ask about BEQ.

And i have total budget set up, as I can re-use or trade in everything from my current gear. And difference between Trinnov and AV10 is 2 24 in subs to our main HT. But let’s. see, maybe I will just smash 10dB low shelf on the subs directly and be done with it.

Well don't just get down with it on behalf of the specs. Looks like budget is there so my advice is not to go with second best. You would for sure enjoy much of the convenience features with AV 10 but then can't really take it to a racing track against Trinnov. But then Marantz has ART so could be less intrusive to implement.

As to 115" TVs, that is indeed a sensitive area. Last year models were great, this years not so much, although a bit cheaper - that would be TCL that is released in that size. There is Hisense though but might be 110". There are bigger ones but then they cost 2x the Trinnov. The only caution is not to go with the Hisense 2024 RGB mini led - that is a really bad unit that was never supposed to be released, but it did for one reason or another. 2026 will be a really good vintage for the TV's, or at least I hope so.

 

[kawauso]

Well bumping this up as no 10 likes. Not my post but don't feel embarrassed to bump it up post. This is a significant find and you will not find it in D&M manuals. So it is a "hack" that is really important to understand the value of your system. If there is so little people that appreciate it, will be indicative of its value after all. But can't say that anyone with a flat panel would want to pass on this. Two centers aligned might be just what you were craving for.

There are some assignment limitations for the center and height 4 terminals, but you can still achieve flexible routing by using any unused power amp channels.
In this setup, both center speakers will play at the same time, but if you save them as Preset 1 and Preset 2, you can easily switch between different center speakers.
It’s a lesser-known feature — I personally use it to run different height speakers for Auro and Dolby setups.

 

[Oddball]

There are some assignment limitations for the center and height 4 terminals, but you can still achieve flexible routing by using any unused power amp channels.
In this setup, both center speakers will play at the same time, but if you save them as Preset 1 and Preset 2, you can easily switch between different center speakers.
It’s a lesser-known feature — I personally use it to run different center speakers for Auro and Dolby setups.
View attachment 489984

Well that is another great post. Many thanks for being so inquisitive and persistent. Not that I could say that for other members. We do tend to fall into the indulgence or our own setups and not looking for more.

BTW can imagine doing the upper center on top of my TV, buti it just might happen one day when I feel like it. Will need a shelf due to tilt, but all doable and possible. Upper center will be a bit weaker that the lower one, but still can imagine them working together well.

 

[kawauso]

Well that is another great post. Many thanks for being so inquisitive and persistent. Not that I could say that for other members. We do tend to fall into the indulgence or our own setups and not looking for more.

BTW can imagine doing the upper center on top of my TV, buti it just might happen one day when I feel like it. Will need a shelf due to tilt, but all doable and possible. Upper center will be a bit weaker that the lower one, but still can imagine them working together well.

Ever feel like switching between two center speakers?
Running both at once seems a bit odd to me, but having a higher one for party mode or casual listening sounds like a fun idea actually.

By the way, my JBL HB1500 is arriving today — I’ll be running four subs in total.
Not sure if I’ll actually feel the difference, but it sure feels good anyway

 

[scotchrocks]

Just adding a few REW measurements for reference in case anyone's curious. These are FR (0-20KHz) and waterfall (10-200 Hz) charts from my center channel with Dirac Off and Dirac ART On (with default Dirac curve). Smoothing is 1/6th for FR.

I haven't tweaked anything recently but don't have any complaints really to address. Maybe adding a fourth sub or rear channels is next.

 

[OCA]

I think that running to lowest common denominator is not the correct thing to do. [talking about whether it ok to cut off bass in movies]. Building on this logic, you could argue that we should cut all the material at 50Hz, as anyway most of the people have issues below this frequency. I don’t really mind what is happening to Netflix releases, but physical media and Kaleidascape are already niche thing, targeting enthusiasts market, so they should offer some added value compared to streaming.

Usually biggest mess is in 50-100Hz area anyway, frequencies below first modal are benign - no standing waves or resonances, maybe some furniture rattling. And you can EQ them very simply, usually with some low shelf. Addressing lowest frequencies is not so complex as it might seem, also there is much higher tolerance to FR ripples, GD etc.

Also RP22 from CEDIA is requiring FR at 115dB to go into teens, if you want to go beyond “ The minimum level of performance necessary to convey basic artistic intent.” And regarding audibility of lowest frequencies - just go to Ascendo showroom and you will understand the difference between the system playing capab;e fo 20Hz and single digits.

BTW - I would strongly recommend to every HT fan to download and print out RP22 - gold mine for understanding how to get performance out of HT, what is important and what not so much.

View attachment 489937



My situation - we will be moving to new house and I want to have 2 theaters - one top class, dedicated room, all bells and whistles, and second one in living room, kind of good enough to enjoy Netflix etc. I have the project already - in-wall DIY speakers, front wall will be Infinite Baffle, 115 in TV etc. What is important - my wife is THE Basshead of our family, this is why I ask about BEQ.

And i have total budget set up, as I can re-use or trade in everything from my current gear. And difference between Trinnov and AV10 is 2 24 in subs to our main HT. But let’s. see, maybe I will just smash 10dB low shelf on the subs directly and be done with it.

I can’t really comment on bass preferences, but personally, living in a building with close neighbors has made me allergic to rattling furniture. Whenever I turn up the volume, it becomes a constant battle between enjoying a movie and keeping the peace next door.

That said, uniform bass response across seats isn’t the holy grail. Once you’ve got that dialed in, there’s still a lot more to great sound. For me, true Atmos immersion hinges far more on precise speaker timing than anything else. And since it’s impossible to time-align every speaker perfectly for every seat, compromises are inevitable and nothing kills the surround experience faster than poorly time aligned speakers. There’s much more to good sound than just uniform frequency response IMO.

 

[kawauso]

I installed the HB1500. It’s really good.

 

[Oddball]

I can’t really comment on bass preferences, but personally, living in a building with close neighbors has made me allergic to rattling furniture. Whenever I turn up the volume, it becomes a constant battle between enjoying a movie and keeping the peace next door.

That said, uniform bass response across seats isn’t the holy grail. Once you’ve got that dialed in, there’s still a lot more to great sound. For me, true Atmos immersion hinges far more on precise speaker timing than anything else. And since it’s impossible to time-align every speaker perfectly for every seat, compromises are inevitable and nothing kills the surround experience faster than poorly time aligned speakers. There’s much more to good sound than just uniform frequency response IMO.

I am in similar position - close neighbors, but then I can't really rattle below 100o dB at 25hz. Otherwise they are fine, or just being too sensitive.

 

[peng]

Just want to add my two cents on OCA's info on the number of "Taps" for Audyssey and Dirac. I have great respect for his contributions as always, but his "1024 " taps for Audyssey is something that I am very skeptical of in terms of accuray for at least too reasons, a) we don't know exactly the basis of the calculations especially when comparing two different platforms, and b) not sure about Dirac yet but in my extensive search and read online, there has been no credible information, and the reported range by various forum members are huge, likely because of again, the caveat, method of calculations, definitions used, and their sources. So, I don't recommend anyway to pay much attention to that specific criteria etc. if comparing within the same product, such as XT vs XT32, minidsp vs minidsp, and Dirac vs Dirac et.

 

[Oddball]

I fully agree. This is not really approximation of what you would hear in your room, but it is still a valid technical point.

I was long in the Audy camp, probably one of the last survivors, but then...not there anymore.

 

[OCA]

Just want to add my two cents on OCA's info on the number of "Taps" for Audyssey and Dirac. I have great respect for his contributions as always, but his "1024 " taps for Audyssey is something that I am very skeptical of in terms of accuray for at least too reasons, a) we don't know exactly the basis of the calculations especially when comparing two different platforms, and b) not sure about Dirac yet but in my extensive search and read online, there has been no credible information, and the reported range by various forum members are huge, likely because of again, the caveat, method of calculations, definitions used, and their sources. So, I don't recommend anyway to pay much attention to that specific criteria etc. if comparing within the same product, such as XT vs XT32, minidsp vs minidsp, and Dirac vs Dirac et.

I’m sorry to break the news, but it’s not even 1024 taps, it’s 1003 for speakers and only 687 for subwoofers. You must have skipped A1 Evo during your “extensive research” which is odd considering it’s free, widely used, and pretty well-known among audiophiles. These aren’t guesses and the numbers are prompted at the start of optimization based on the model:

You’d think the person who wrote the actual filter transfer tool for these receivers might get at least a brief benefit of the doubt but I gave up long time ago

PS Dirac sends 2048 filter taps to the receiver but I never checked how many of them are not zero (not many would be my guess) as I never found their calibration algorithm worth investigating.

 

[Oddball]

I’m sorry to break the news, but it’s not even 1024 taps, it’s 1003 for speakers and only 687 for subwoofers. You must have skipped A1 Evo during your “extensive research” which is odd considering it’s free, widely used, and pretty well-known among audiophiles. These aren’t guesses and the numbers are prompted at the start of optimization based on the model:

View attachment 490173

You’d think the person who wrote the actual filter transfer tool for these receivers might get at least a brief benefit of the doubt but I gave up long time ago

PS Dirac sends 2048 filter taps to the receiver but I never checked how many of them are not zero (not many would be my guess) as I never found their calibration algorithm worth investigating.

We would like you to switch to Dirac. While Audy is your baseline, what they are doing in Dirac seems to be worth a go?