Active Room Treatment (ART) by Dirac

[OCA]

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?

ART is definitely very interesting and a game changer for bass response and I am sure there's space for improvement where it hands over the calibration to DL but it will require a license to work and will only be useful to a few and even fewer will appreciate. I am working on using MIMO filters through Audyssey interface these days. I don't know why, I am not as enthusiastic as before lately. It's been weeks since I watched a movie from start to finish in my system.

 

[masterhw]

ART is definitely very interesting and a game changer for bass response and I am sure there's space for improvement where it hands over the calibration to DL but it will require a license to work and will only be useful to a few and even fewer will appreciate. I am working on using MIMO filters through Audyssey interface these days. I don't know why, I am not as enthusiastic as before lately. It's been weeks since I watched a movie from start to finish in my system.

I’ve greatly appreciated your work, my interest in ART is mostly academic right now and the fact that A1 Evo is even in the conversation is a phenomenal feat for what was the backmarker in room correction not long ago. It is always a joy when the hardware is capable of greater performance when more intelligently utilized - I firmly believe you’ve helped avoid some amount of e-waste, among your many contributions to the space.

On a personal note, I certainly encourage you to allow yourself some distance when you feel it natural. I know there is a chorus of overdramatic forum goers on AVS that make it sound like you’ve died when three days pass without an update. Let’s save that for the pope, not an overworked and under-appreciated hobbyist. I encourage you to wait until your spark returns and curiosity burns, this is when one’s best work is done. Until then, there are other joys in life, and we’ll all still be here arguing over ART support thresholds.

 

[Oddball]

ART is definitely very interesting and a game changer for bass response and I am sure there's space for improvement where it hands over the calibration to DL but it will require a license to work and will only be useful to a few and even fewer will appreciate. I am working on using MIMO filters through Audyssey interface these days. I don't know why, I am not as enthusiastic as before lately. It's been weeks since I watched a movie from start to finish in my system.

We are really looking forward to any of your projects. But take some time to rest and enjoy a bit.

I just finished Midway and wow, it was breathless as it ever was. Not because of ART or anything else. Moving on to Masters of Air to have my mind blown again by the story and similar non-technical things.

 

[kawauso]

We are really looking forward to any of your projects. But take some time to rest and enjoy a bit.

I just finished Midway and wow, it was breathless as it ever was. Not because of ART or anything else. Moving on to Masters of Air to have my mind blown again by the story and similar non-technical things.

I did my ART calibration today and measured 13 positions each for Auro and Dolby.
It was pretty exhausting, so I hope you’ll give me a little credit for it.
I won’t be changing my setup for a while, so Kawauso-chan really put in some effort this time.

 

[Sal1950]

I did my ART calibration today and measured 13 positions each for Auro and Dolby.
It was pretty exhausting, so I hope you’ll give me a little credit for it.
I won’t be changing my setup for a while, so Kawauso-chan really put in some effort this time.

Un hun

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?

We, so are you on Dirac's payroll now.

 

[Dj7675]

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.
View attachment 489984

I will be doing this as well with 2 centers, one above, one below a solid screen. Set you height 4 speakers to center height, change the pre amp on that speaker output to output center instead. And you now have 2 centers. You can use pink noise do it by ear watching content to set levels. You can also set delays by adjusting the distance. You are limited to .5dB level adjustments. Once that is done run Audyssey/Dirac and it will measure them as 1 speaker.
Another use would be to create surround arrays. Set your speaker layout with wides and they change output to surround. It could be useful if you have multiple rows in some home theaters. It is a super useful feature and they way Denon handles presets etc it makes them quite flexible.

 

[kawauso]

Un hun


We, so are you on Dirac's payroll now.

Not sure what that emoji means, but I just added the 4th subwoofer and was recalibrating the system, that’s all. Guess I’m just a little too excited about bass!

I installed the HB1500. It’s really good.
View attachment 490133

 

[kawauso]

I will be doing this as well with 2 centers, one above, one below a solid screen. Set you height 4 speakers to center height, change the pre amp on that speaker output to output center instead. And you now have 2 centers. You can use pink noise do it by ear watching content to set levels. You can also set delays by adjusting the distance. You are limited to .5dB level adjustments. Once that is done run Audyssey/Dirac and it will measure them as 1 speaker.
Another use would be to create surround arrays. Set your speaker layout with wides and they change output to surround. It could be useful if you have multiple rows in some home theaters. It is a super useful feature and they way Denon handles presets etc it makes them quite flexible.

"It's great that D&M has this feature! Can the levels of the two center speakers be adjusted separately? The surround array idea could be really useful too, especially if there are big differences in viewing positions between the screen and the TV. It’s much more convenient than switching speakers with a selector, since it only takes a single button press. I’m using 12 height speakers myself, and I’ve set up 6 for Dolby and Auro, switching between them depending on the format. It works really well.

 

[Dj7675]

Can the levels of the two center speakers be adjusted separately?

Yes. Once you set the speaker layout with a Center Height it will be in the speaker list to adjust levels and delays. What makes it a viable solution is to be able change levels for sure.

 

[kawauso]

Yes. Once you set the speaker layout with a Center Height it will be in the speaker list to adjust levels and delays. What makes it a viable solution is to be able change levels for sure.

Center height speakers are part of the Auro-3D format, but are they functioning properly? I think they wouldn't work if you switch to Dolby Atmos. How are you making use of them?

 

[Dj7675]

Center height speakers are part of the Auro-3D format, but are they functioning properly? I think they wouldn't work if you switch to Dolby Atmos. How are you making use of them?

I don’t know a lot about Auro-3D, but I do know that CH is a DTS:X format speaker location. You are just setting the Speaker to that speaker type in that layout to have another speaker so you can change the signal to center instead of center height. Then you have 2 speakers, driven by 2 different amps on the A1H. In Dolby Atmos playback both the center and the CH are getting the same center signal. There are of course pros/cons to this with comb filtering etc. From my experience in my previous room I liked it (used a stormaudio at that time).

Here are the actual steps….
1-Set you layout you want but it needs to include center height. In my case to test I set it to a 7.1.5 layout for example.
2-Set the TS/CH speaker to CH
3-Click on the “Settings” button on the bottom and it will open the custom amp assign settings
4-Change the height 3 amp settings to center instead of center height
5-Set the levels of the center speakers so that the combined levels come out of the center of the screen.
6-Proceed to calibration of Audyssey/Dirac as normal and it will calibrate them as one speaker (I think)

 

[kawauso]

I don’t know a lot about Auro-3D, but I do know that CH is a DTS:X format speaker location. You are just setting the Speaker to that speaker type in that layout to have another speaker so you can change the signal to center instead of center height. Then you have 2 speakers, driven by 2 different amps on the A1H. In Dolby Atmos playback both the center and the CH are getting the same center signal. There are of course pros/cons to this with comb filtering etc. From my experience in my previous room I liked it (used a stormaudio at that time).

Here are the actual steps….
1-Set you layout you want but it needs to include center height. In my case to test I set it to a 7.1.5 layout for example.
2-Set the TS/CH speaker to CH
3-Click on the “Settings” button on the bottom and it will open the custom amp assign settings
4-Change the height 3 amp settings to center instead of center height
5-Set the levels of the center speakers so that the combined levels come out of the center of the screen.
6-Proceed to calibration of Audyssey/Dirac as normal and it will calibrate them as one speaker (I think)

Got it, so in one setup you're using it as a center height and in another setup, you're treating it as two centers. If the level balance is maintained after calibration, it really feels like a bit of a workaround.

 

[kawauso]

Had a bit of a rough day today, so I blasted some music for about a minute (Explosive Kawauso style) and it felt so much better. With JBL speakers, even though ART correction reduces headroom, they still manage to handle it.

JTR subwoofers moving.

https://twitter.com/x/status/1989233237164716176

 

[RenPa]

I never found their calibration algorithm worth investigating.

I always read your comments on various blogs with great interest. Over time, you have acquired an impressive level of expertise in these areas. I have myself learned the basics of acoustics, speaker measurement, phase relationships, delays, etc. on the job in order to achieve specific results. I know enough to recognize when Dirac is more of a hindrance than a help. But I am nowhere near as skilled as you are. That's why I hope that one day you'll explain why your lack of interest for his algorithm or you'll have the time and the means to get on board with Dirac, because for us mere mortals, it's the application that allows us to achieve the most. For example, what intrigues me is the fact that you recommend (if I'm not mistaken) several measurements in the same place, whereas Dirac has made a (surely well-researched) choice of 9 measurements distributed around the MLP.

Translated with DeepL.com (free version)

 

[Fidji]

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.

This is interesting area in general, for those with neighbors/families - is there any experience on whether ART is improving bass bleed out to adjacent spaces, thanks to its modes cancellation? [it works like this with DBA/SBA] I would expect yes, as you are cancelling lot of energy already inside the room, but it might also be, that it is just moved outside. Checking it should be simple

I don’t know a lot about Auro-3D, but I do know that CH is a DTS:X format speaker location. You are just setting the Speaker to that speaker type in that layout to have another speaker so you can change the signal to center instead of center height. Then you have 2 speakers, driven by 2 different amps on the A1H. In Dolby Atmos playback both the center and the CH are getting the same center signal. There are of course pros/cons to this with comb filtering etc. From my experience in my previous room I liked it (used a stormaudio at that time).

Here are the actual steps….
1-Set you layout you want but it needs to include center height. In my case to test I set it to a 7.1.5 layout for example.
2-Set the TS/CH speaker to CH
3-Click on the “Settings” button on the bottom and it will open the custom amp assign settings
4-Change the height 3 amp settings to center instead of center height
5-Set the levels of the center speakers so that the combined levels come out of the center of the screen.
6-Proceed to calibration of Audyssey/Dirac as normal and it will calibrate them as one speaker (I think)

While it is technically possible, I would expect results to be mixed from such setup. There are lot of things to consider - one needs to consider dispersion patterns, SPL capabilities of each speaker and ability to have individual steering and EQ for each speaker, outside of automatic calibration. I have calibrated Double Center manually and in order to get acceptable results there was hell lot of manual tweaking and measurements - even when i had.2 centers [as opposed to center + repurposed height speaker]. If I remember correctly I had upper center high-passed at 250Hz with 4dB attenuation and slighlty different target curve, so that combined response was coherent with L-R.

Anyway - I think with 115+ in TV’s becoming more common we will see new HT design approaches and best practices - as with the sheer size and it will be more difficult to retain audio-video coherence. I have 124 in screen and lot of flexibility - low end of the screen is like 50 floor in order to have comfortable viewing height. I can see lot of images of HT’s with screens/TV’s so high that it feels like sitting in first row of IMAX. Not good long term for sure.

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.

Seat to Seat consistency is for sure good to have, especially for bass. Obviously one needs lot of subs and/or advanced correction.
And yes, there are another 20-30 parameters to take care of in order to get immersive bubble.

My checklist:
L-C-R cohesion, Bass Quality [FR, decay, impulse, GD, remove resonances] and align with L-C-R so that it creates uniform sonic landscape in front of me. I spend probably 80% of effort to gt this rigth
Surrounds - have enough of them, place them properly, check for FR and SPL/Distrotion
Heights/Ceiling is mostly placement and quantity with appropriate SPL/Distortion/FR - here I trust Trinnov to get it right
I have my tweaks - e.g. I run ceiling speakers 2dB hotter and I high pass them at 100Hz fot upmixers to get controlled bass response.

 

[Dj7675]

While it is technically possible, I would expect results to be mixed from such setup. There are lot of things to consider - one needs to consider dispersion patterns, SPL capabilities of each speaker and ability to have individual steering and EQ for each speaker, outside of automatic calibration. I have calibrated Double Center manually and in order to get acceptable results there was hell lot of manual tweaking and measurements - even when i had.2 centers [as opposed to center + repurposed height speaker]. If I remember correctly I had upper center high-passed at 250Hz with 4dB attenuation and slighlty different target curve, so that combined response was coherent with L-R.

I expect it to take some effort when I set it up, but expect it to work just fine…
-Select appropriate speakers. I am using KEF CI3160RLM horizontally above/below a microled. Lower is 10 deg below ear height and upper is 20 degrees above ear height. Measurements show sound is nearly the same as on axis.
-Use pink noise to set levels of lower and upper to get audio centered on the screen
-Run automated EQ. Both centers will be used with automated EQ and will be treated as a single speaker.
-Listen to the results with real content, do post measurements to try to mitigate any issues and adjust the center channel curve if needed or use the A1H Dialogue enhancer with does broad EQ boosting
Not everyone is bothered by center channel below or above a solid screen. But as TVs get larger, the issue is more prevalent. Even though dual centers will create some issues with comb filtering, in my experience in my other room, the positives of not being able to locate the center outweigh any issues created.

 

[peng]

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.

May be you got defensive and not bother to understand what I wrote, your 1103 and 687 seems like some numbers in absolute sense? In my post, I clearly mentioned "we don't know exactly the basis of the calculations especially when comparing two different platforms". So, your 687 active out of 704, 1003 active out of 1024 just show that you can't just base on a number without knowing the conditions, is that 687, 1003 active at all times, maximum, or under certain use conditions right, so don't be sorry about the breaking the news lol..

I did search the internet expensively and again, I found people's claims are all over the map, any who do a search with the help of the so called AI such as Gemini, Co-pilot will know what I mean. By the way, I said right from the beginning as I am just skeptical when I see some single absolute numbers, I have never said you were wrong about those number, really never. If I did, I would apologize, so again, only expressing my skepticism for reasons I explained already.

Also, I really wanted to use your freeware but loss the opportunity when I decided to replace my Denon with an Anthem, and now I look forward to you coming up with something that can work with Anthem's. ARC G is really not that great vs Audyssey XT32 and Dirac Live, based on my own experience, yet their fans thought ARC is the best.

Please don't think I was doubting your knowledge in this field, I do admire your knowledge in this field, but I repeat, I naturally am skeptical about certain things when it is stated as though it is a fact, without supporting data, conditions, caveats or references etc. Now you included the highlighted information, it helps, and thank you very much for that.


Lastly, if someone do a quick search just to see how confusing this number game is, I just tried and go the following AI search, if you are not interested, don't waste your time, and just to be clear, I am not saying, in fact, don't believe such AI search will get one correct information. In fact, it just shows how one needs to be skeptical about such simple questions and answers, in this case, a matter of how many FIR tap Audyssey is capable, answer: 1024, all I am saying to such Q/A is, really, any caveats, conditions?

Here's my Gemini search as a demo, the contents may be from mostly wrong, partially true, to mostly true, hard to say..


Search question:

so all you know is 512x but not an absolute tap count number, such as that given by devices like minidsp what could have spec of say 1024 taps


Gemini response:

That is absolutely correct. What I can provide you is the filter resolution number (512×), but not the absolute, fixed tap count (like 1024 taps) for a very specific technical reason: Audyssey uses a proprietary, non-linear method called a Warped FIR filter.

Why Audyssey Doesn't Use a Simple Tap Count​

The fundamental difference lies in how Audyssey allocates the filter's processing power (taps or coefficients) compared to a standard, linear FIR filter used by devices like a MiniDSP:

[td]Feature[/td][td]Audyssey MultEQ XT32[/td][td]Standard FIR (e.g., MiniDSP)[/td] [td]Filter Type[/td][td]Warped FIR Filter[/td][td]Linear FIR Filter[/td] [td]Tap Allocation[/td][td]Non-linear: More resolution (taps) is dedicated to the low frequencies (below ∼300Hz), where room modes are most problematic.[/td][td]Linear: Taps are evenly distributed across the entire frequency range.[/td] [td]Key Metric[/td][td]Filter Resolution (512×): Indicates the effective corrective power.[/td][td]Absolute Tap Count (e.g., 1024): Indicates the total number of filter coefficients.[/td] [td]Effective Resolution[/td][td]Provides note-by-note (about 1/12-octave) correction from 50Hz up, and coarse control down to 25Hz.[/td][td]Resolution is fixed. For a 1024-tap filter at 48kHz, the low-frequency resolution is very poor (e.g., only two frequency bins below 100Hz).[/td]

Audyssey's Claimed "Effective" Tap Count​

In an interview, an Audyssey representative (Jeff Clark) explained this approach and made an illustrative comparison:

The technology is based on a patented method of reallocating the FIR filtering power from high frequencies toward the low frequencies...1 The XT32 FIR system is effectively 2$32$ times more efficient—that is, the taps we run are 3$32$ times more effective than they would be without XT32.4

While this doesn't give a simple number, they have stated that the resolution of XT32 is equivalent to 16,000 taps per channel if implemented using a standard linear FIR filter.

Conclusion​

• Absolute Tap Count: Audyssey does not publish a single, fixed tap count for XT32 because the "taps" are non-linearly applied using proprietary algorithms.
• Credible Source Metric: The official number is the $512\times$ filter resolution.
• Comparison Estimate: The effective power of XT32, particularly in the bass region, is claimed to be equivalent to a system with thousands of standard FIR taps, far exceeding the resolution of a simple $1024\text{-tap}$ linear filter in the critical low frequencies.

 

[Fidji]

I expect it to take some effort when I set it up, but expect it to work just fine…
-Select appropriate speakers. I am using KEF CI3160RLM horizontally above/below a microled. Lower is 10 deg below ear height and upper is 20 degrees above ear height. Measurements show sound is nearly the same as on axis.
-Use pink noise to set levels of lower and upper to get audio centered on the screen
-Run automated EQ. Both centers will be used with automated EQ and will be treated as a single speaker.
-Listen to the results with real content, do post measurements to try to mitigate any issues and adjust the center channel curve if needed or use the A1H Dialogue enhancer with does broad EQ boosting
Not everyone is bothered by center channel below or above a solid screen. But as TVs get larger, the issue is more prevalent. Even though dual centers will create some issues with comb filtering, in my experience in my other room, the positives of not being able to locate the center outweigh any issues created.

Yes, KEFs Coax are very sensible choice for dual centers [and overall]. But with 10 degrees I would not bother with dual center.
How big is your screen and how far you sit from it?

 

[Dj7675]

Yes, KEFs Coax are very sensible choice for dual centers [and overall]. But with 10 degrees I would not bother with dual center.
How big is your screen and how far you sit from it?

Screen is 119”wide and 54” high (2.22 AR Microled) had a similar screen size in my old theater 122”x52”, solid non AT screen. I had center below the screen and thought it was fine. I then set up dual center with the processor at the time (Storm) and found the difference to be a nice upgrade. If watching content, and you perceive the center channel below the screen and audio not coming from the center of the screen, it is a problem worth trying to address. Some may not notice, or care, but it is quite bothersome if you can localize the sound coming from below the screen/tv.

 

[OCA]

May be you got defensive and not bother to understand what I wrote, your 1103 and 687 seems like some numbers in absolute sense? In my post, I clearly mentioned "we don't know exactly the basis of the calculations especially when comparing two different platforms". So, your 687 active out of 704, 1003 active out of 1024 just show that you can't just base on a number without knowing the conditions, is that 687, 1003 active at all times, maximum, or under certain use conditions right, so don't be sorry about the breaking the news lol..

I did search the internet expensively and again, I found people's claims are all over the map, any who do a search with the help of the so called AI such as Gemini, Co-pilot will know what I mean. By the way, I said right from the beginning as I am just skeptical when I see some single absolute numbers, I have never said you were wrong about those number, really never. If I did, I would apologize, so again, only expressing my skepticism for reasons I explained already.

Also, I really wanted to use your freeware but loss the opportunity when I decided to replace my Denon with an Anthem, and now I look forward to you coming up with something that can work with Anthem's. ARC G is really not that great vs Audyssey XT32 and Dirac Live, based on my own experience, yet their fans thought ARC is the best.

Please don't think I was doubting your knowledge in this field, I do admire your knowledge in this field, but I repeat, I naturally am skeptical about certain things when it is stated as though it is a fact, without supporting data, conditions, caveats or references etc. Now you included the highlighted information, it helps, and thank you very much for that.


Lastly, if someone do a quick search just to see how confusing this number game is, I just tried and go the following AI search, if you are not interested, don't waste your time, and just to be clear, I am not saying, in fact, don't believe such AI search will get one correct information. In fact, it just shows how one needs to be skeptical about such simple questions and answers, in this case, a matter of how many FIR tap Audyssey is capable, answer: 1024, all I am saying to such Q/A is, really, any caveats, conditions?

Here's my Gemini search as a demo, the contents may be from mostly wrong, partially true, to mostly true, hard to say..


Search question:

so all you know is 512x but not an absolute tap count number, such as that given by devices like minidsp what could have spec of say 1024 taps


Gemini response:

LLM AI is merely next word prediction based on probability derived from training data. In many audiophile subjects, that's 90% forum gibberish. It's not even truly context aware although that's improving rapidly. "List me resources that claim 512 for taps for Audyssey" kind of prompts would be a better use of AI for research imo.

1024, 512 and 128 are all valid constants taken from the actual interface but it's quite a bit more complicated than that.