Active Room Treatment (ART) by Dirac

[Fidji]

The expensive Bryston and Storm units have excellent external UI and channel management software. Consumer units don't come close.
I contacted Storm and asked if they would ever release low channel count units. They said they have been considering that, but I doubt they will go through with it. They have a clear niche, which I don't see them relinquishing.
The other issue is that consumer software is terrible. And it's the hardware manufacturer's responsibility to make sure that Dirac works without problems. Consumer AVPs have a poor track record there.
I have low expectations for Marantz and Denon units performing well. I would very much like to be wrong.

After loosing ART exclusivity, only thing that really distinguishes Storm from D&M is ability to accommodate 24 or 32 Channels in their higher level AVP’s and of course unique AoIP full digital units - which is actually super cool feature, for complex setups to avoid all the cable management hassle. So top of the line AVP’s still do have use cases, this is why I specifically mentioned base model.
I fully get the flexibility of bass management and channel allocation + internal equalizer, but this are very specific use cases. Yes, it is nice being able to set crossover to LR2 at 74.6Hz instead of 80Hz. If it is worth 3x the price, is for the prospective buyer to answer.

D&M are some of the the most reliable AVR/AVPs out there, IMO.

I've had a perfectly working Marantz SR7012 for 7 years now, purchased brand new as the 7013 came out (inferior unit as it had that can't use Neural X with Dolby sources nonsense on it) for a mere $1100.

It's now expanded to 11.1.8 (plus optional SH and optional 4 floor speakers plus front stage "lift" to gone from the screen using Scatmos, speaker switches and a mixer) for a mere $1800 with none of the Atmos renderer issues (like no Front Wide output for fixed and/or bed objects despite correlated info).

25 speaker home theater for $2900 (3 sub output with Mini-DSP correction using multi-sub optimizer plus seat shaker channel) for a total of 11.4.10.4.

Screw Trinnov and Storm's overpriced crap! Yeah, consumer equipment doesn't come close. LOL. Keep buying into that audiophile snake oil marketing!

I've even got Logic 7 courtesy of a Lexicon MC-1 plugged into the 7.1 inputs of the 7012. With Scatmos expansion it's really Logic 11 (plus front stage lift and the 4 floor ambient speakers taken from the Front Wide Scatmos Surround outputs) has 11.4.2.4 in operation with Logic 7 for truly holographic 2-channel music.

There's also a Carver Sonic Holography option for 2-channel output as well that's particularly interesting in Multi-Channel Stereo mode with some albums.

Books Shade's Dear Future Self and Yello's Point are mind blowing in Dolby Atmos on it (24 foot long room).

I admire your ingenuity in creating immersive audio out of 11.2 receiver. But the hard truth, [similar to bass nad LFE discussion] even with 25 speaker you are still processing 7.1.4 signal. Adding 4 subs does not make it 7.4.4 it is still 7.1.4.

Can not speak of Storm, but all that you describe can be done super easily in Trinnov, including “Scatmos” [you have full Matrix routing capability so 28x28 matrix in my case] and on top you have 29 separate presets and 32? [not sure] logical sources.

E.g. Apple TV same HDMI input - you can have source called Netflix Atmos, Netflix Scatmos, Netflix DTS upmix, Netflix DTS HD MA, Chris Nolan [for super screwed soundtracks], Apple Music Stereo, Apple Music Spatial Audio, Apple Music Stereo Upmix. Each can have different preset, target curve and speaker count/routing setup and bass management settings. And will be finetuned for that specific use case.
Same with Blu-Ray player, where you have different setting for Classical Music DTS, Live Concerts, Movies, Action Movies, DTS to Scatmos and Atmos Native, Atmos Scatmos.

And all of those you can multiply to be optimised for MLP only, 2 seats, 3 seats, 2 rows independently. Each with different measurements and calibration.

Do you want to cross-over between your subs with LR4? Bessel3? Butterworth 2? LR2? LR8? at 61Hz? do you want to invert polarity of subs for your surround backs? Different gain for LFE? Do you wan to compensate for stacked subs? Compensate for AT screen? And another probably 200 variants and options to get what you desire.

This is the real difference between 35k processor and consumer grade. Do you really need all of this stuff? 99% don’t.

This is why I am thinking for secondary setup to go for Marantz, as ART is the solution for what will be pretty standard setup, with 115in TV and corner loaded subs. There is one thing I need to understand still and that is possibility to go dual center, which is critical for this super big screens,

 

[Magnus]

I meant specifically that the benefits come from the UI and channel mapping software, as well as the reliability of Dirac implementation.

There's no snake oil to these products.

The myth of ultra high-end DACs being worth the price of the paper their specs are printed on for one thing. I can't tell the DAC in my 6-disc Pioneer CD player from 1989 (still works) from the one in my Marantz AVR (I guess I'm deaf), but Trinnov is definitely pushing the envelope with even pricier DACs in their newer units!

I guess you should get something for over $30k for an Altitude32, but I can't hear it. The price hasn't dropped one scintilla since 2014 despite the hardware stagnating beyond DAC and HDMI changes and other minor changes (CPU might be better, but it has no effect as 2014 units have the same features).

Electronics used to trend downward so it's definitely an oddball thing that home theater equipment keeps moving ever higher, even with inflation. A single chip could handle what four or eight did in 2014, but 11.2 costs almost twice as much as it used to. With inflation being 35 cents more per dollar now than in 2014 and electronics are roughly 2x-10x faster depending on the application yet 11.x has doubled in price in recent years from $2k retail to $4k retail new. No wonder home theater is losing popularity again (brief bump during Covid).

 

[Magnus]

I admire your ingenuity in creating immersive audio out of 11.2 receiver. But the hard truth, [similar to bass nad LFE discussion] even with 25 speaker you are still processing 7.1.4 signal. Adding 4 subs does not make it 7.4.4 it is still 7.1.4.

If you understand how the Atmos and DTS:X renderers worked, you'd realize there's no real functional difference between rendering (Dolby) and matrix extracting the channels discretely (Neural X or Scatmos). You get the same benefits (off-axis seating resolution is greatly improved with more discrete speaker locations as there are no "new locations" outside the 7.1.4 skeleton, only speakers between those locations, allowing them to be extracted relatively discretely via matrix steering logic like PLII or DTS' own Neural X. There might be audible differences, but they are small.

Can not speak of Storm, but all that you describe can be done super easily in Trinnov, including “Scatmos” [you have full Matrix routing capability so 28x28 matrix in my case]

That's the first time I've ever heard anyone claim they can extract more channels (outside of Neural X) on a Trinnov. Do tell more of how this is done. Of course, at 10x the cost (not counting amplification which using PLIIx AVRs includes), you've got to expect some conveniences from Trinnov.

The one feature I am super jealous of in a Trinnov is the ability to digitize its 7.1 inputs. Theoretically, it should be able to run Neural X on the output of my Lexicon MC-1, giving a top end (ceiling) to its Logic 7 output for music. That I would be curious to hear.

 

[EWL5]

I meant specifically that the benefits come from the UI and channel mapping software, as well as the reliability of Dirac implementation.

There's no snake oil to these products.

Dirac on Denon/Marantz has been exemplary! Not so much on the PAC and Arcam units (due to DLBC woes)...

 

[Keith_W]

The myth of ultra high-end DACs being worth the price of the paper their specs are printed on for one thing. I can't tell the DAC in my 6-disc Pioneer CD player from 1989 (still works) from the one in my Marantz AVR (I guess I'm deaf), but Trinnov is definitely pushing the envelope with even pricier DACs in their newer units!

Umm, not to point out the obvious but if you buy a Trinnov you're not buying it for the DAC. You are buying it for the DSP.

 

[Magnus]

Umm, not to point out the obvious but if you buy a Trinnov you're not buying it for the DAC. You are buying it for the DSP.

I would have been buying the 16 to get 20-channels without a cheaper way. But people buy things for different reasons.

 

[GXAlan]

I guess you should get something for over $30k for an Altitude32, but I can't hear it. The price hasn't dropped one scintilla since 2014 despite the hardware stagnating beyond DAC and HDMI changes and other minor changes (CPU might be better, but it has no effect as 2014 units have the same features).

The Altitude is basically a licensing dongle.

The 2014 hardware allows

• Dolby Atmos® Native Decoding & Up-mix: Up to 34 discrete channels
• Auro-3D® Native Decoding & Up-mix (Auromatic): Up to 13.1
• DTS:X® Pro Native Decoding & Up-Mix: Up to 30.2
• Data precision: 64 bits, floating point
• Max Sampling Rate (up to 24 channels): Native, up to 192kHz
• Max Sampling Rate (up to 32 channel with +4 option): Native, up to 96kHz
• Max Sampling Rate (above 24 channels): Native, up to 96kHz
• Max Sampling Rate (above 32 channels): Native, up to 48kHz

And Waveforming.

You need to upgrade the hardware when these specs are no longer adequate. They have updated the HDMI board over time…

Think about Dirac. If your product from 2014 had plenty of power to handle ART, would you really worry about the hardware being stagnant? Even today, it’s hard to find a Dirac home theater product handling >48 kHz sampling rate. (Theta Casablanca is one that comes to mind).

No wonder home theater is losing popularity again (brief bump during Covid).

It may very well be the opposite. Because home theater is losing popularity, companies have to shift to higher prices to maintain the same level of revenue.

People seem happy streaming short form content for their entertainment.

there's no real functional difference between rendering (Dolby) and matrix extracting the channels discretely (Neural X or Scatmos). …There might be audible differences, but they are small.

Agree about small differences. Even taking a Christopher Nolan, flat 7.1 mix and upsampling to include heights is good.

That's the first time I've ever heard anyone claim they can extract more channels (outside of Neural X) on a Trinnov. Do tell more of how this is done.

As you point out earlier, what’s the difference between a rendered object and more channels via matrix’d setup? In the case of Atmos, it becomes semantics. One thing that distinguishes Trinnov from other systems like Dirac is that it knows exactly where the speakers are in 3D space. If you think about the acceptable “ranges” for speaker position for Atmos, and having a dynamic object, Trinnov can precisely pinpoint which speaker(s) to deliver that audio. The precision in delivery is one element of why discrete objects versus matrixed audio hits the “audible differences” but “small” based upon the part of your comment I quoted.

I still argue that a good soundbar system gets you 70% of the audio performance of a Trinnov and you begin your path toward the point of diminishing returns.

Get great speakers and now you jump to 85% performance with a basic Sony receiver. You get to 95% with a flagship Dirac processor and then 100% with Trinnov.

For us audiophiles, we like to magnify the differences that actually exist. Take your spouse or kid or random neighbor? Many built in TV speakers results in inaudible dialogue at times, which impacts the movie watching experience. Jumping up to a good soundbar resolved the dialogue issue and then after that, the rest is just enhancements.

Take tactile bass. We love it, but the story and visuals with audible but not tactile bass is still very good. The cost of adding technologies like ART or Waveforming is going to be lost on your *truly* average viewer. By average, think about what that really means…

 

[Oddball]

Scatmos aside at really a separate topic, we do still have issues to address with ART on D&M.

By specs, it it not matching Trinnov solution and I have not heard any practical answer how ART package deals with 15hz issues. Obviously Trinnov requires much more hardware and in very inconvenient positons for subs, but at least they are being open and honest about the requirements.

So what will the ART actually do for various systems? Any case studies that one could refer to? For sure my system is not common and would like some sense of what it is could do before splashing $1K on DIrac - which I am not really inclined to do as nobody is telling me what would be the benefit?

 

[Goodfellas27]

Instead of just correcting individual speakers, ART makes all the speakers in a system (including subwoofers) work together as a unified, intelligent system. It uses patented MIMO (Multiple-Input Multiple-Output) technology. The speakers collaborate acoustically, using the strengths of one speaker to compensate for the weaknesses of another, or to actively counteract problematic room interactions like resonances and reflections.

ART primary goal is to reduce bass decay times and eliminate lingering bass (often described as "muddy" or "boomy" bass), especially in the critical frequency range below 150 Hz. This leads to significantly clearer, more impactful sound, better bass clarity, and improved soundstage and imaging across the listening area. It aims to make the "room disappear" from the sound, allowing the content to take center stage.

ART actively uses bass enabled speakers and subs in the system to send out "anti-reflections" or "anti-noise" signals that effectively cancel out problematic room resonances and lingering bass. It's akin to active noise cancellation, but applied to room acoustics. A key benefit of ART is its ability to significantly reduce bass decay times (reverberation), which leads to much tighter, clearer, and more impactful bass. This is something traditional room correction struggles with and typically requires extensive physical bass trapping.

While it has a strong focus on bass (below 150 Hz), ART is part of the broader Dirac Live ecosystem, which also handles correction for the full frequency range (above 150 Hz) using traditional impulse response correction.
Regarding below 20hz treatment, you 1st need speakers or subwoofers that are able to go below 20hz and a microphone with enough resolution below 20Hz (Not UMIK1) for Dirac to address that ranges by moving it's curtains below that range.

Regarding Trinnov Waveforming; typically requires multiple subwoofers (often 4 or more) strategically placed on both the front and rear walls of the room, sometimes in specific heights (planar arrays) or distributed along the width (cylindrical arrays) The front subwoofer array is used to create a single, coherent wavefront of bass that is steered towards the listening area, minimizing interactions with side walls, floor, and ceiling. The rear subwoofer array (and potentially all subs) then acts as an "absorber," actively canceling out the reflections from the back wall and other room modes. This is achieved by precise timing and phase control.

Rather than correcting existing problems, Waveforming aims to prevent room modes from forming in the first place by controlling the bass wave's propagation directly.

Both solutions are a expensive proposition for many. Trinnov more is so than Dirac's take. Not even actual Movie Theaters are using something like Waveforming, only "select few partners" which I cannot visit and see the latest movie release AKA - exhibition space.

Check out YouTube @SoundUnitedTraining for future Dirac videos for M&D products.

Dirac MIMO Whitepaper: https://www.dirac.com/wp-content/up...pRdKB-OkPEczdejW2zzryVWxKyzPtIFNjKnT7nelldP6u

Trinnov Waveforming: https://www.trinnov.com/en/technologies/active-acoustics/waveforming/?hl=en-US

Scatmos aside at really a separate topic, we do still have issues to address with ART on D&M.

By specs, it it not matching Trinnov solution and I have not heard any practical answer how ART package deals with 15hz issues. Obviously Trinnov requires much more hardware and in very inconvenient positons for subs, but at least they are being open and honest about the requirements.

So what will the ART actually do for various systems? Any case studies that one could refer to? For sure my system is not common and would like some sense of what it is could do before splashing $1K on DIrac - which I am not really inclined to do as nobody is telling me what would be the benefit?

 

[Oddball]

Instead of just correcting individual speakers, ART makes all the speakers in a system (including subwoofers) work together as a unified, intelligent system. It uses patented MIMO (Multiple-Input Multiple-Output) technology. The speakers collaborate acoustically, using the strengths of one speaker to compensate for the weaknesses of another, or to actively counteract problematic room interactions like resonances and reflections.

ART primary goal is to reduce bass decay times and eliminate lingering bass (often described as "muddy" or "boomy" bass), especially in the critical frequency range below 150 Hz. This leads to significantly clearer, more impactful sound, better bass clarity, and improved soundstage and imaging across the listening area. It aims to make the "room disappear" from the sound, allowing the content to take center stage.

ART actively uses bass enabled speakers and subs in the system to send out "anti-reflections" or "anti-noise" signals that effectively cancel out problematic room resonances and lingering bass. It's akin to active noise cancellation, but applied to room acoustics. A key benefit of ART is its ability to significantly reduce bass decay times (reverberation), which leads to much tighter, clearer, and more impactful bass. This is something traditional room correction struggles with and typically requires extensive physical bass trapping.

While it has a strong focus on bass (below 150 Hz), ART is part of the broader Dirac Live ecosystem, which also handles correction for the full frequency range (above 150 Hz) using traditional impulse response correction.
Regarding below 20hz treatment, you 1st need speakers or subwoofers that are able to go below 20hz and a microphone with enough resolution below 20Hz (Not UMIK1) for Dirac to address that ranges by moving it's curtains below that range.

Regarding Trinnov Waveforming; typically requires multiple subwoofers (often 4 or more) strategically placed on both the front and rear walls of the room, sometimes in specific heights (planar arrays) or distributed along the width (cylindrical arrays) The front subwoofer array is used to create a single, coherent wavefront of bass that is steered towards the listening area, minimizing interactions with side walls, floor, and ceiling. The rear subwoofer array (and potentially all subs) then acts as an "absorber," actively canceling out the reflections from the back wall and other room modes. This is achieved by precise timing and phase control.

Rather than correcting existing problems, Waveforming aims to prevent room modes from forming in the first place by controlling the bass wave's propagation directly.

Both solutions are a expensive proposition for many. Trinnov more is so than Dirac's take. Not even actual Movie Theaters are using something like Waveforming, only "select few partners" which I cannot visit and see the latest movie release AKA - exhibition space.

Check out YouTube @SoundUnitedTraining for future Dirac videos for M&D products.

Dirac MIMO Whitepaper: https://www.dirac.com/wp-content/up...pRdKB-OkPEczdejW2zzryVWxKyzPtIFNjKnT7nelldP6u

Trinnov Waveforming: https://www.trinnov.com/en/technologies/active-acoustics/waveforming/?hl=en-US

Mic is a valid concern, but then there is no real solution if Dirac does not tell us. UMIK-1 is recommended for Dirac overall and is what I understand they use to check the performance of various versions internally. One would certainly not want to solve 15hz mode by compromising the results across the band. Trinnov mic solution is expensive but obviously the only serious mic in the business.

Otherwise I do fully understand how ART and Waweforming work and have heard both. Waweforming is improved version of the double bass array that was around as long as subwoofers were - and that in a primitive version I use to this day with amazing results.

I do understand that everyone is excited about ART, but don't really understand why are most happy with the current Dirac approach that provides only high level information. I would want to know all there is for $1K, and not having to acquire that knowledge by watching hours of videos and scrolling through thousand of pages of various threads as they expect me to pay them good money and not trinkets. This is not really D&M issue as Dirac's signature is behind ART. Hope D&M will be more helpful than Dirac as implementation approaches.

Call me spoiled, but I would expect more for $1K.

EDIT: Lol, forgot to mention my real concern. Don't want to spend 200 hours implementing and tweaking ART, which based on the current state of things is my conservative estimate.

 

[alvigg]

ART is not $1000. The whole Dirac suite is. ART will be (I think) $300 if you have DLBC. That already works well with multiple subs.

 

[Oddball]

ART is not $1000. The whole Dirac suite is. ART will be (I think) $300 if you have DLBC. That already works well with multiple subs.

The only reason I would get Dirac is ART, so will be close to a grand.

 

[GXAlan]

The price hasn't dropped one scintilla since 2014 despite the hardware stagnating beyond DAC and HDMI changes and other minor changes (CPU might be better, but it has no effect as 2014 units have the same features).

To follow up on my other comment:

The StormAudio current gen uses in TI K2G and the 4x ADSP-21489

The Marantz AV10 uses the ADSP-21593.

Again, we can talk about efficient versus inefficient software code, but the Intel CPUs haven’t been upgraded in 10 years because the hardware is fine.

You also have to consider system RAM and system memory bandwidth too…

 

[Oddball]

To follow up on my other comment:
View attachment 453469

The StormAudio current gen uses in TI K2G and the 4x ADSP-21489

The Marantz AV10 uses the ADSP-21593.

Again, we can talk about efficient versus inefficient software code, but the Intel CPUs haven’t been upgraded in 10 years because the hardware is fine.

You also have to consider system RAM and system memory bandwidth too…

Can't recall where I read it but @OCA noted that at current levels Storm and Marantz are not really maxed out? Obviously Trinnov is a wet dream, but I would take it only with a playmate that would implement it over the course of 2 or more weeks as that would be a minimum of my requirements.

Leaving me alone in the room with Trinnov would not really fare well. I would not want or desire to spend countless hours exploring literally endless options thay have. Based on the contact with the dealers in my region, the are pretty much the sell and go option. They think that 8 hours of setting up Trinnov with my system would be more than enough .

 

[GXAlan]

Can't recall where I read it but @OCA noted that at current levels Storm and Marantz are not really maxed out?

That’s right. They are NOT maxed out except for the fact that they restrict to 48 kHz. (Which is fine).

But that also explains why Trinnov hasn’t changed their setup much. It’s still way overkill.

Leaving me alone in the room with Trinnov would not really fare well. I would not want or desire to spend countless hours exploring literally endless options thay have.

The default approach is fine. It’s trickier than the Marantz but probably not a Storm.

Based on the contact with the dealers in my region, the are pretty much the sell and go option. They think that 8 hours of setting up Trinnov with my system would be more than enough .

I think Trinnov’s default mode is like setting the time on your analog quartz wrist watch by going to time.gov and pushing in the crown as soon as the seconds hand hits zero.

Bringing in a professional calibrator is like fine tuning the thermocompenstation of the quartz circuitry and then pulling atomic clock data from several different NIST sources, and averaging them to account for network latency and stochastic variability of quantum mechanics to ensure that your watch is timed to the 1/32768th of a second that represents the true current time and maximum theoretical performance of a quartz based clock.

Sometimes being a split second off is just fine. Other times, you feel like you have already put in $30K so you want to use all the features you paid for. Nothing wrong with having a pro do your home theater calibration, and a pro will do a better job than a single point calibration wizard — but instead of doing a head to head comparison, if you measured your overall enjoyment of a movie — it probably doesn’t change between the two as much as one would lead you to believe.

Put a different way. If you are watching a movie alone, or watching with someone else or a group of others, even if you are sitting in your ideal seat, the presence of others changes the reflections. Ambient temperature changes the sound. Amir’s measurements have revealed this. But all this nitpicking is fine tuning.

The real negatives of Trinnov are extreme cost and poor ergonomics (no HDMI CEC, remote controlled power on/CEC power on). But if you got it for free, it’s not too hard to set it up and get performance that is as good or better than a normal Marantz/Denon setup.

 

[Magnus]

To follow up on my other comment:
View attachment 453469

The StormAudio current gen uses in TI K2G and the 4x ADSP-21489

The Marantz AV10 uses the ADSP-21593.

Again, we can talk about efficient versus inefficient software code, but the Intel CPUs haven’t been upgraded in 10 years because the hardware is fine.

You also have to consider system RAM and system memory bandwidth too…

You miss the point entirely. The cost of the computer parts inside the Trinnov cost maybe 1/100 of what they did in 2014 because newer stuff is so much faster, but the price hasn't dropped a dime! They're essentially REAMING you, but you don't care! It's a Trinnov! Hooray!

 

[Oddball]

That’s right. They are NOT maxed out except for the fact that they restrict to 48 kHz. (Which is fine).

But that also explains why Trinnov hasn’t changed their setup much. It’s still way overkill.



The default approach is fine. It’s trickier than the Marantz but probably not a Storm.

I think Trinnov’s default mode is like setting the time on your analog quartz wrist watch by going to time.gov and pushing in the crown as soon as the seconds hand hits zero.

Bringing in a professional calibrator is like fine tuning the thermocompenstation of the quartz circuitry and then pulling atomic clock data from several different NIST sources, and averaging them to account for network latency and stochastic variability of quantum mechanics to ensure that your watch is timed to the 1/32768th of a second that represents the true current time and maximum theoretical performance of a quartz based clock.

Sometimes being a split second off is just fine. Other times, you feel like you have already put in $30K so you want to use all the features you paid for. Nothing wrong with having a pro do your home theater calibration, and a pro will do a better job than a single point calibration wizard — but instead of doing a head to head comparison, if you measured your overall enjoyment of a movie — it probably doesn’t change between the two as much as one would lead you to believe.

Put a different way. If you are watching a movie alone, or watching with someone else or a group of others, even if you are sitting in your ideal seat, the presence of others changes the reflections. Ambient temperature changes the sound. Amir’s measurements have revealed this. But all this nitpicking is fine tuning.

The real negatives of Trinnov are extreme cost and poor ergonomics (no HDMI CEC, remote controlled power on/CEC power on). But if you got it for free, it’s not too hard to set it up and get performance that is as good or better than a normal Marantz/Denon setup.

Not really sue what to tell out of all this. Trinnov is definitively not user friendly but can do anything you ask of it and based on the measurements that are not derived by toy-like mics like UMICs or Audy.

I did not put in $30K in Trinnov but only EUR 5K in Marantz AV-10, but for sure have more than $30K in the rest of the system. The really sour part was finding the right subwoofer positioning. That alone was probably 200 hours - but I guess was just what came with a courtesy of a new new large room.

The bad part was being able to tune my system to my requirements. Took probably 200 hours before I gave up and called a professional calibrator. After all that work I did he was pretty lucky but still did 100's of various REW measurements to tune it in OCD fashion as he really seemed to car about the outcome, not just the fee which was actually pretty low per hour. Based on the records from my reception, he was either chilling in my flat to 8 hours a day for 5 days, or he was doing something remotely meaningful - which is likely because he did a great job in integrating the whole system.

So I have pretty much a custom setup with all advanced bass management that Audy can take, phase adjustment to the extent possible, group delay considerations, and we even calculated what is the optimal setup for winter months when humidity in the flat goes low due to central heating.

Do I need this drama with Trinnov or ART? Probably could take it with Trinnov as similar path as Audy, but not feeling compelled with ART as so little information is available and at the end can't see what the ultimate outcome would be.

 

[masterhw]

Since its quasi-related: after OCA's insightful comment I was curious why more processing horsepower wouldn't necessarily yield improvements, so I spent some time learning about decimation techniques and the implications of different filter types (phase type is particularly important).
I was surprised at how effectively something like decimation, which processes lower frequencies at a lower sample rate, can reduce demand. You don't need to process the full 20-20,000 Hz at 48kHz sampling, the bulk of that range can be subsampled. The bass region in particular can be run at 1/64th without compromising quality, and since this is where the bulk of correction is happening then this gain is massive. Working harder is not working smarter.
I also got schooled on the fact that the latency associated with certain DSP engines is often not a hardware problem. Linear phase FIRs, which Audyssey uses exclusively, increase in latency with more filters (mathematical property, not a computational one), and would produce the same exact result even with a hardware upgrade. Using the headroom gains to add more filters or up the sampling to 96kHz (which would require doubling the filters to maintain their resolution) would double the latency even on a supercomputer.
For Audyssey to step past this would require an architectural overhaul to use other kinds of filters, like Dirac and Trinnov. But again the results of using other filter types are not necessarily superior, just have different sets of tradeoffs.

Dirac and Trinnov (among others) use minimum phase and mixed phase filters, but the result can be working much much harder to produce a similar result a couple milliseconds faster. Trinnov's hardware has impressive GFLOPS but the fact that they need it could indicate their architecture is a poorly optimized mess. Also may not perform much better on a modern i7. We just don't know. And there are enough places for clever engineering optimizations or brute force that I don't feel we can tell very much about these architectures from their footprint alone. I think this is why @Oddball seeks deeper understanding.

 

[Oddball]

Since its quasi-related: after OCA's insightful comment I was curious why more processing horsepower wouldn't necessarily yield improvements, so I spent some time learning about decimation techniques and the implications of different filter types (phase type is particularly important).
I was surprised at how effectively something like decimation, which processes lower frequencies at a lower sample rate, can reduce demand. You don't need to process the full 20-20,000 Hz at 48kHz sampling, the bulk of that range can be subsampled. The bass region in particular can be run at 1/64th without compromising quality, and since this is where the bulk of correction is happening then this gain is massive. Working harder is not working smarter.
I also got schooled on the fact that the latency associated with certain DSP engines is often not a hardware problem. Linear phase FIRs, which Audyssey uses exclusively, increase in latency with more filters (mathematical property, not a computational one), and would produce the same exact result even with a hardware upgrade. Using the headroom gains to add more filters or up the sampling to 96kHz (which would require doubling the filters to maintain their resolution) would double the latency even on a supercomputer.
For Audyssey to step past this would require an architectural overhaul to use other kinds of filters, like Dirac and Trinnov. But again the results of using other filter types are not necessarily superior, just have different sets of tradeoffs.

Dirac and Trinnov (among others) use minimum phase and mixed phase filters, but the result can be working much much harder to produce a similar result a couple milliseconds faster. Trinnov's hardware has impressive GFLOPS but the fact that they need it could indicate their architecture is a poorly optimized mess. Also may not perform much better on a modern i7. We just don't know. And there are enough places for clever engineering optimizations or brute force that I don't feel we can tell very much about these architectures from their footprint alone. I think this is why @Oddball seeks deeper understanding.

Thanks for deeper understanding. As always you do seem to get to the bottom of things. I am just trying to point out couple of things that could benefit all. What is really the benefit of ART in 80hz bass managed systems, and what is the benefit in larger systems. It is a a question as 80hz bass managed system could have 4 subs and the full-range system (with bass management) could have only 2 of them. Or the other way around. And this is just a really simple question compared to the bass management that has been out there for years.

Don't want to detail the ARC worship thread, but not really sure what to worship at this point expect that it sounds grand just because its new and untested.

 

[GXAlan]

You miss the point entirely. The cost of the computer parts inside the Trinnov cost maybe 1/100 of what they did in 2014 because newer stuff is so much faster, but the price hasn't dropped a dime! They're essentially REAMING you, but you don't care! It's a Trinnov! Hooray!

The computer parts have dropped in value but newer versions of the software have been given without added cost.

I have a Trinnov. Every new software update has come to me for free, so I don’t feel like I am getting “reamed.” Maybe there is incentive to keep the price high since it is like US Social Security where the people paying currently by buying new Trinnov’s are helping those who are existing owners…

Put another way. Think about Roon and a Lifetime License. Now think about Trinnov as selling lifetime licenses with the dongle being the hardware…

The original 32 ch Altitude 32 was €40000. Today, it’s €36000, with newer HDMI and DACs.

To keep up with inflation, that Altitude 32 should be €54000.