Active Room Treatment (ART) by Dirac

[dlaloum]

100% agree on this one.

ART works miracles all the way from quiet stereo listening to high octane movie watching in a huge variety of room sizes and layouts. You just need enough cone area, power handling in either speakers or subs to reach the SPL and (currently limited) infrasonic extension of choice.

Other limitation is your and your partners willingness to have all that equipment "polluting" the living room.

Changing speaker output profile as SPL rises, would result in unpredictable ART results... the quality of the results is to some degree dependent on linear distortion free output at all SPL's used - if it varies from the measure/calculated response - then the end result will vary from it too!

On the other hand, what SPL do you need? - the Peak headroom required for a movie is 20db above reference - so if you use 75db as your reference point (as I do) - then your peak level will be circa 95db.

If you use cinema spl levels - that shifts the reference point to 85db and peak SPL to 105db

I might like to have 105db - as safety headroom, but I am unlikely to use the top 10db

Music often has wider dynamic range - especially well recorded acoustic instruments... doesn't seem as loud but the peaks can be deceptive.

Still 115db is seriously LOUD.

 

[Frank207be]

Changing speaker output profile as SPL rises, would result in unpredictable ART results... the quality of the results is to some degree dependent on linear distortion free output at all SPL's used - if it varies from the measure/calculated response - then the end result will vary from it too!

On the other hand, what SPL do you need? - the Peak headroom required for a movie is 20db above reference - so if you use 75db as your reference point (as I do) - then your peak level will be circa 95db.

If you use cinema spl levels - that shifts the reference point to 85db and peak SPL to 105db

I might like to have 105db - as safety headroom, but I am unlikely to use the top 10db

Music often has wider dynamic range - especially well recorded acoustic instruments... doesn't seem as loud but the peaks can be deceptive.

Still 115db is seriously LOUD.

115dB from 20Hz to 20kHz is too loud for me too as I don't want to be deaf at the ago of 60. I listen to my movies with a normal speach level.

But for obtaining tons of clean headroom I've done my math too:

My 4 DIY subwoofers are sealed for the best sound quality.
Yes together these subs could do 140dB at 70Hz but I like bass extension and sealed means that their sensitivity at 20Hz drops 12dB against their peak output at 70Hz so I have a theoretical SPL of 115dB at 10Hz and 125dB at 20Hz without room gain.
And how much headroom does ART need for a -18dB support level deep down from 20Hz? 12dB?
I like 5-10dB boost with movies. This offsets the loss of "resonance" energy and adds kick without bloat.

So nothing unpredictable here with ART: as soon as I switch from DLBC to ART I get kick instead of rumble and not a single sign of stress or distortion. All in all 7 grand for this result is not that expensive IMHO.

 

[Oddball]

Of course 115dB is insane level of noise - but that happens very occasionally in the soundtrack. Average for reference level would be more like 95dB which is still very loud and extended exposure to that noise level not good for hearing. I am trying to keep 85dB average, which would be 105dB peaks. Numbers are obviously lower for beds - this is max which is for subs.

 

[mmares056]

Of course 115dB is insane level of noise - but that happens very occasionally in the soundtrack. Average for reference level would be more like 95dB which is still very loud and extended exposure to that noise level not good for hearing. I am trying to keep 85dB average, which would be 105dB peaks. Numbers are obviously lower for beds - this is max which is for subs.

Yeah I wouldn't be able to listen at 115dB levels. I try to keep the levels in the 85dB-88dB average range which gives me 105dB-109dB peaks in my open concept living room. These levels equates to -13 MV on physical media. I used to calibrate to 78dB Pre ART but with ART my calibration is 85dB at 0 MV. Post ART, even -24dB MV on physical media is pretty loud. With my ART calibration I never listen louder than -22 MV.

 

[seashell]

Just tried ART today on an AV10, just letting it do its default thing. Result was basically no low base. So I looked closer and the subwoofer group was not assigned as support to any group, including itself. Since ART disables the crossover setting I assume this means it basically just felt it was best not to use my subs at all. Is that the correct interpretation?

ETA: I played a bass heavy song in stereo and indeed the subs are silent. So it seems my interpretation of how ART works is correct and somehow Dirac thinks the default bass management should be a 50 Hz high pass filter on all the mains and no subs. Brilliant. I now realize this also means that using 4 subs to support 15 mains will use 60 of the 96 filters available on the AV10. So every channel can basically have one other support pair besides the subs.

 

[KenMasters]

Just tried ART today on an AV10, just letting it do its default thing. Result was basically no low base. So I looked closer and the subwoofer group was not assigned as support to any group, including itself. Since ART disables the crossover setting I assume this means it basically just felt it was best not to use my subs at all. Is that the correct interpretation?

ETA: I played a bass heavy song in stereo and indeed the subs are silent. So it seems my interpretation of how ART works is correct and somehow Dirac thinks the default bass management should be a 50 Hz high pass filter on all the mains and no subs. Brilliant. I now realize this also means that using 4 subs to support 15 mains will use 60 of the 96 filters available on the AV10. So every channel can basically have one other support pair besides the subs.

You've probably run into the same bug I did. You could try taking new measurements and working from that basis, in my case that returned the functionality of the subs. Howver I still had issues and ultimately I had to perform a Factory Restore for ART to function properly.

I would imagine I could have skipped retaking the measurements if I'd performed the restore from the get go. If your AV10 is not set to automatically update, check if there aren't any updates pending, maybe you can avoid the hassle.

 

[seashell]

You've probably run into the same bug I did. You could try taking new measurements and working from that basis, in my case that returned the functionality of the subs. Howver I still had issues and ultimately I had to perform a Factory Restore for ART to function properly.

I would imagine I could have skipped retaking the measurements if I'd performed the restore from the get go. If your AV10 is not set to automatically update, check if there aren't any updates pending, maybe you can avoid the hassle.

Just to make sure I understand, you're saying that after the factory restore you now get audio from the subs even from ART groups that do not include the subwoofers as support for those groups? And the low end of self support is effectively the crossover?

 

[KenMasters]

Just to make sure I understand, you're saying that after the factory restore you now get audio from the subs even from ART groups that do not include the subwoofers as support for those groups? And the low end of self support is effectively the crossover?

My situation was twofold.

Trying ART for the first time with a pre-existing set of measurements, my subs did not engage at all, no matter the configuration. Once I took a fresh set of measurements, the subs operated when playing content.

However the bass reproduction still didn't sound right, and I later found that no matter what parameters I adjusted in ART, no change was made to the system.

I don't know what went wrong, maybe my the firmware on my AVR wasn't up to date, maybe DLBC and ART got their wires crossed, I don't know. All I know is that since removing and reinstalling my X6800H's firmware, ART has functioned perfectly.

 

[dlaloum]

My situation was twofold.

Trying ART for the first time with a pre-existing set of measurements, my subs did not engage at all, no matter the configuration. Once I took a fresh set of measurements, the subs operated when playing content.

However the bass reproduction still didn't sound right, and I later found that no matter what parameters I adjusted in ART, no change was made to the system.

I don't know what went wrong, maybe my the firmware on my AVR wasn't up to date, maybe DLBC and ART got their wires crossed, I don't know. All I know is that since removing and reinstalling my X6800H's firmware, ART has functioned perfectly.

1st generation product woes....

 

[seashell]

My situation was twofold.

Trying ART for the first time with a pre-existing set of measurements, my subs did not engage at all, no matter the configuration. Once I took a fresh set of measurements, the subs operated when playing content.

However the bass reproduction still didn't sound right, and I later found that no matter what parameters I adjusted in ART, no change was made to the system.

I don't know what went wrong, maybe my the firmware on my AVR wasn't up to date, maybe DLBC and ART got their wires crossed, I don't know. All I know is that since removing and reinstalling my X6800H's firmware, ART has functioned perfectly.

I tried the factory restore and used my existing measurements. Same results. Took quick new measurements, this time Dirac suggested adding the subs as support speakers to all the mains. I had already manually adjusted the good measurements before to add the subs to the mains, which did activate the subs. But that does use up 4 filters per main plus 3 for the 4 subs supporting each other. 63 of 96 filters just for basic bass management.

I'm not saying per channel sub filters can't do a better job than equalized subs and a crossover, but it's pretty obvious to me the DSP on the AV10 is pretty anemic for ART in a 15.4 setup. If this is the way it works there is basically no processing power left for support from main to main. If each paired group supports itself that's another 14 filters (the center has no partner). So now we're at 77 of 96 filters. Any pair supporting another pair is 4 more filters. So there's not much going to happen.

Pretty incremental gain, but I suppose ART's cost is relatively small compared to the AV10.

 

[dlaloum]

I tried the factory restore and used my existing measurements. Same results. Took quick new measurements, this time Dirac suggested adding the subs as support speakers to all the mains. I had already manually adjusted the good measurements before to add the subs to the mains, which did activate the subs. But that does use up 4 filters per main plus 3 for the 4 subs supporting each other. 63 of 96 filters just for basic bass management.

I'm not saying per channel sub filters can't do a better job than equalized subs and a crossover, but it's pretty obvious to me the DSP on the AV10 is pretty anemic for ART in a 15.4 setup. If this is the way it works there is basically no processing power left for support from main to main. If each paired group supports itself that's another 14 filters (the center has no partner). So now we're at 77 of 96 filters. Any pair supporting another pair is 4 more filters. So there's not much going to happen.

Pretty incremental gain, but I suppose ART's cost is relatively small compared to the AV10.

Yes you need to keep an eye on your setups appetite for filters (and avoid using 98... as it may in fact be asking for more than 98 - with poor results as a consequence)

Having said that - you can have 7 speakers involved in support without going over 60 filters - and the results are exemplary (honestly quite stunning... and I have posted my waterfall charts earlier in the thread)

Add more speakers (more subs, surrounds, heights) and the number of filters needed can escalate quite quickly - yes there is a strict filter budget, and 98 filters is the limit for all members of the D&M family.... (they all share the same DSP processor).

With care you should be able to optimise a 7.2.4 setup without any issue (and yes, you won't be able to configure it to have every speaker supporting every other speaker - or rather, you can set it up that way, but results will not be great as it will run out of filters!)

There are options other than D&M with larger filter numbers (ie: more powerful DSP's) eg: Monoprice HPT1, StormAudio

And I expect we will see more coming onto the market this year.

Also more powerful DSP chipsets may be the excuse for D&M's next generation... (the Xx900 ? )

I'm not complaining - my 5.2.4 setup is sounding better than it ever has - using circa 60 filters, excluding heights from support duties (although they themselves are supported)

 

[TimoJ]

Also more powerful DSP chipsets may be the excuse for D&M's next generation... (the Xx900 ? )

Isn't it odd (and funny) that the HTP-1, designed in 2019, can do it and Marantz's top model can't?

 

[seashell]

Isn't it odd (and funny) that the HTP-1, designed in 2019, can do it and Marantz's top model can't?

Sad really. But the HTP-1 only does 16 channels total including subs right? So basically 7.4.4 or 5.4.6 max.

And does it officially support ART yet? Not just for beta testers.

 

[smygolf]

Sad really. But the HTP-1 only does 16 channels total including subs right? So basically 7.4.4 or 5.4.6 max.

And does it officially support ART yet? Not just for beta testers.

 

[TimoJ]

But the HTP-1 only does 16 channels total including subs right? So basically 7.4.4 or 5.4.6 max.

And does it officially support ART yet? Not just for beta testers.

16 total channels, of them 5 subs max. ART firmware is available as a public beta that anyone can install and Dirac already sells ART licenses for the HTP-1.

 

[tifune]

16 total channels, of them 5 subs max. ART firmware is available as a public beta that anyone can install and Dirac already sells ART licenses for the HTP-1.

In this thread it's been mentioned that each processor has a different # of ART filters available for use. Do you know, how many does the HTP1 have and how can i determine the # in use?

 

[TimoJ]

In this thread it's been mentioned that each processor has a different # of ART filters available for use. Do you know, how many does the HTP1 have and how can i determine the # in use?

The HTP-1 has no filter/support channel limit. The number of used "filters" is not visible. Internal parameters show that my 7.3.6 system has 240 "max cross terms" which I believe is the same thing that is show on D/M units as 98 (or sometimes 96 or 97).

 

[Oddball]

The HTP-1 has no filter/support channel limit. The number of used "filters" is not visible. Internal parameters show that my 7.3.6 system has 240 "max cross terms" which I believe is the same thing that is show on D/M units as 98 (or sometimes 96 or 97).

Hopefully HTP-1 implementation will raise some eyebrows with D&M management and they expand the filter support. Not that I need it but excess never hurts.

As you note, DSP chip is obviously not the limitation. Difficult to say what kind of issues D&M is facing as their firmware covers a really large number of products while HTP-1 is a bespoke product.

 

[Oddball]

I tried the factory restore and used my existing measurements. Same results. Took quick new measurements, this time Dirac suggested adding the subs as support speakers to all the mains. I had already manually adjusted the good measurements before to add the subs to the mains, which did activate the subs. But that does use up 4 filters per main plus 3 for the 4 subs supporting each other. 63 of 96 filters just for basic bass management.

I'm not saying per channel sub filters can't do a better job than equalized subs and a crossover, but it's pretty obvious to me the DSP on the AV10 is pretty anemic for ART in a 15.4 setup. If this is the way it works there is basically no processing power left for support from main to main. If each paired group supports itself that's another 14 filters (the center has no partner). So now we're at 77 of 96 filters. Any pair supporting another pair is 4 more filters. So there's not much going to happen.

Pretty incremental gain, but I suppose ART's cost is relatively small compared to the AV10.

I must admit I got a bit lost in your posts so apologies if did not get something right.

I use 94/96 filters on AV10 for 9.4.4 system. 4 subs as one group and then 9 bed channels grouped per pair (center separately) and 4 atmos also grouped as pairs. Subs support themselves as they are grouped together. Subs also support all other groups (so they are in addition to support noted below).

Subs get supported by 3 other groups. Center gets supported by 2 groups and does not support any group. Mains get also supported by 2 groups. Not with my Mac right now but also think that FW get support from LR. Atmos just support each other in 2 groups. This is from memory, but it is substantially right, see comment below.

Out of the whole scheme, the most meaningful support is cross support to and from subs. The rest is not really that essential. Subs are 20-150hz, bed channels are 50-150hz and Atmos are 90-150hz.

Bass levels without any curves are indeed pitiful for most material (a bit of a bass head speak), so I have 6 presets from +2dB to +12dB shelf on low end. +12dB is earth-shaking on most material. My subs are by Dirac default between 40-60% of their output.

More filters to get some additional support for Atmos would not hurt, but also don't know if they would be really that beneficial.

 

[Dj7675]

Yes you need to keep an eye on your setups appetite for filters (and avoid using 98... as it may in fact be asking for more than 98 - with poor results as a consequence)

Having said that - you can have 7 speakers involved in support without going over 60 filters - and the results are exemplary (honestly quite stunning... and I have posted my waterfall charts earlier in the thread)

Add more speakers (more subs, surrounds, heights) and the number of filters needed can escalate quite quickly - yes there is a strict filter budget, and 98 filters is the limit for all members of the D&M family.... (they all share the same DSP processor).

With care you should be able to optimise a 7.2.4 setup without any issue (and yes, you won't be able to configure it to have every speaker supporting every other speaker - or rather, you can set it up that way, but results will not be great as it will run out of filters!)

There are options other than D&M with larger filter numbers (ie: more powerful DSP's) eg: Monoprice HPT1, StormAudio

And I expect we will see more coming onto the market this year.

Also more powerful DSP chipsets may be the excuse for D&M's next generation... (the Xx900 ? )

I'm not complaining - my 5.2.4 setup is sounding better than it ever has - using circa 60 filters, excluding heights from support duties (although they themselves are supported)

I think we need to be careful to extrapolate your room, speakers, configuration to making a statement that if you use max or near max filters, you will have poor results (either objective or subjective). There seems to be just as many that have great results and are happy at or near the max. With so many variables regarding speaker capabilities, locations of speakers, rooms, and personal preference it all seems quite difficult to make definitive statements one way or the other.
Thanks for taking the time to try different things with your system, measure, and report your findings.
One thing that might be interesting, with any of your configurations is to see the affect of spl on ART performance. If/when you do your next set of measurements, maybe take post measurements at 3 different SPL’s. It would be interesting if the results stay the same or if, as some speakers get closer to their limits, results get worse at higher spl’s. It would make sense, that the louder you listen, the more conservative you need to be on setting the lower limits of frequency support.