-
Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Active Room Treatment (ART) by Dirac
- Thread starter Flak
- Start date
The closest thing I've heard that makes sense. Is that Dirac was trying to make the implementation into Denon and Marantz avrs to be as simple as possible to use. So the average Joe could set it up properly. And that they're having difficulty getting to that goal. This may be compounded by the fact that I think I heard that Dirac had changed ownership. However, considering the size of Denon and Marantz, that still doesn't make any sense why it would be taking so long assuming they have anything to do with it
I know Dirac requires measuring with specific layouts, at minimum 9 points (cube). If the position of the measuring points is known, I suppose it is possible to create a mathematical model of room reflections with machine learning algorithms and define speaker response whose sum combines to the desired IR at the MLP, regardless of the geometric and acoustic unknowns of the environment.
I understand that Dirac intelligently uses stochastic to calculate filters instead of deterministic, in order to make the calculation possible on normal PCs.
I imagine this introduces considerable complexity given the amount of variables involved, so if the algorithm does not appropriately define the limits within which to attempt the stochastic simulation, derived from the calculated acoustic model of the room, the result will be anything but good.
Personally, I am convinced that AI is crucial to these tasks and the fact that DIRAC is pointing in this direction, summed to the concept of active wave forming, is precisely what makes me believe in the validity and innovation of their approach.
I hope they manage to fine-tune ART in order to exploit total potential.
I understand that Dirac intelligently uses stochastic to calculate filters instead of deterministic, in order to make the calculation possible on normal PCs.
I imagine this introduces considerable complexity given the amount of variables involved, so if the algorithm does not appropriately define the limits within which to attempt the stochastic simulation, derived from the calculated acoustic model of the room, the result will be anything but good.
Personally, I am convinced that AI is crucial to these tasks and the fact that DIRAC is pointing in this direction, summed to the concept of active wave forming, is precisely what makes me believe in the validity and innovation of their approach.
I hope they manage to fine-tune ART in order to exploit total potential.
Good point. Considering the absolute phenomenal growth of AI/deep learning/deep, neural networking or whatever you want to call it, in the last 1 to 2 years, I would think that this would help with ART implementation immensely . But now that makes me worried that the promised implementation into the current model lineup of Denon and Moran's equipment won't end up happening because the processors in them won't be powerful enough to do the job
It already IS implemented in all StormAudio processors and every measurement I found showed incredible good results! This doesn't look like struggle at all.
The complexity of these calculations is huge - I was really surprised to see the results. But there are plenty of installations out there (storm audio processors) and they work (like Trinnovs system also works).
I bet all these amps with 4 sub outputs are made with ART in mind. No idea why they struggle to implement - an official statement would be gold.
You must know something Gene of Audioholics don't since he couldn't get his Storm working with ART and publish test results since longer than a year now:
When will Audioholics share their thoughts on Dirac ART?
Ya'll mentioned months ago that you were trying out StormAudio units with Dirac ART. When do we get to hear your impressions?
forums.audioholics.com
Denon/Marantz are fine. The same SHARC ADSP-21593 chip runs across most of the x800 lineup and is even in the flagship A1H!
- Joined
- Jun 24, 2024
- Messages
- 1,503
- Likes
- 2,245
Gene's issue there seems to have something to do with his infrasonic bass response (which I don't think ART even attempts to address with a HPF at 20Hz) and his active speaker setup (which probably introduces additional latency issues that a more common passive system wouldn't have). However, even if some people have ART working with Trinnov/Storm, those systems have way more computation resources on hand than any traditional AVR I'm aware of and have probably had tons of tinkering done by the owner, the manufacturer, and/or Dirac themselves. Getting it working on an AVR with limited compute and without requiring a lot of support time is certainly not trivial, if it's even feasible in the first place.You must know something Gene of Audioholics don't since he couldn't get his Storm working with ART and publish test results since longer than a year now:
When will Audioholics share their thoughts on Dirac ART?
Ya'll mentioned months ago that you were trying out StormAudio units with Dirac ART. When do we get to hear your impressions?
Last edited:
It seems that Gene problems are related to the infrasonic band of his particular system, not correctly managed by ART.You must know something Gene of Audioholics don't since he couldn't get his Storm working with ART and publish test results since longer than a year now:
When will Audioholics share their thoughts on Dirac ART?
Ya'll mentioned months ago that you were trying out StormAudio units with Dirac ART. When do we get to hear your impressions?
Bug also confirmed by another Audioholics article, where on the other hand they show a correction of Dirac ART as per textbook.
Before:
After:
Actually I have not yet seen other technologies achieve these results via software. It is necessary to mathematically model the acoustic response of the room and apply MIMO processing in order to do so. A simple FDW with an inversion filter is not enough.
Last edited:
Then again, ART also works with the Monoprice HTP-1 and it has over 6 years old technology/DSP.
- Joined
- Jun 24, 2024
- Messages
- 1,503
- Likes
- 2,245
Does it? I see no mention of ART on Monoprice's listing nor do I see mention of the Monoprice on Dirac asides from the usual DLRC/DLBC licenses. Also, do we have any info on what compute resources the HTP-1 has? Regardless of age, it can easily outclass an AVR with a hardware DSP chip depending on what its running.
It's been BETA on the HTP-1 for YEARS. You cannot get it as a regular person. They said they'd work on DTS:X Pro AFTER that goes full. It's 2025. The system came out in 2019. SIX years and no DTS:X Pro. Two years since they announced ART.
NOTHING
Forget it. ART isn't on anything but Storm. It clearly has major issues or it'd be on D&M by now. I wrote it all off years ago.
NOTHING
Forget it. ART isn't on anything but Storm. It clearly has major issues or it'd be on D&M by now. I wrote it all off years ago.
Yes it does, I actually have it running (alpha test version). It's not publicly available and I'm not sure if it ever will be.
HTP-1 has one DSP (MDS APM-117 board with Ti K2G,1.0 GHz) that I believe is the same as what StormAudio uses just for audio decoding (replaced now by the new ADEC). But Storms have 1-2 additional DSPs (I think Core has 1 extra DSP, Elite has 2 extras). HTP-1's OS is Armbian Linux with Olimex som204-a20 CPU.
My best guess is that the core issue isn't simply about sending a MIMO polarity-inverted signal to cancel a reflection before it reaches the wall. In theory, any speaker in the system could contribute to that with precise timing. The challenge is that the counteracting signal doesn't just eliminate the reflection and vanish - it continues to bounce back and forth in the room. This creates extra energy that wasn’t part of the original audio track. I had similar experience while working on virtual bass array filters in the past. I suspect that’s why Trinnov relies on multiple sets of subwoofers, strategically placed on non-adjacent walls for MIMO bass arrays.
@LionIT per Dirac's level2 support, PC/Mac rollout of ART is the lowest priority. To quote the email I got from them: "at the current moment all the focus is on the avr market". There is a current PC beta running (Kal Rubinson of stereophile has access), but broad rollout is from my understanding, only a distant hope (paraphrased from their email to me).
Given their lack of success rolling this out beyond storm, and even the difficulties with storm, I'd get a trinnov if its in budget for active bass correction. Or multiple PSI AVAAs. or do a DBA. or hope that @OCA can reverse engineer something like it
Given their lack of success rolling this out beyond storm, and even the difficulties with storm, I'd get a trinnov if its in budget for active bass correction. Or multiple PSI AVAAs. or do a DBA. or hope that @OCA can reverse engineer something like it
Reports by StormAudio ART users and reviewers are consistently positive....
Clearly it works, although for HT use it appears they are planning on limiting it to 150Hz (although during the beta they allowed limits of up to 500Hz).
In car audio environments, Dirac have reported ART working at up to 4kHz.... so in a controlled environment they can tune up to much higher frequencies, than what is viable in a less controlled / unknown environment.
Gene reported that his main issues related to using full range speakers and using them as corrective speakers...
Things worked fine in the most common configuration - using frequency limited speakers with subs
He wasn't talking about infrasonics. Although it has been noted that Dirac has been avoiding the infrasonic range - the focus of ART is within the sonic range -20Hz to 150Hz.
Gene said he plans to complete his ART review once Dirac have resolved the issues with the use of full range speakers.
Keith_W
Major Contributor
These are the differences between cars and homes:
- the car interior is a known shape. Once you optimise DSP for one model, you can sell thousands of cars with the exact same DSP settings. Every hobbyist has a different room, different number of speakers, etc.
- the people who tune DSP for cars are trained professionals, as opposed to untrained hobbyists. Hobbyists might take wrong measurements, use wrong settings, etc.
- car manufacturers have a huge budget for getting the DSP right in cars, as opposed to hobbyists whose budget is peanuts by comparison.
So if you were Dirac, which market would you prioritise? The one that can make you a lot of money for less work, or the one that pays you less for more work?
No, you can do a basic setup with a single point, or a 3 point measurement....
Advanced DSP models allow you to effectively control and make interact sound waves to create whatever you want, also more than one sweet spot. I had read some Princeton documents some time ago that showed the research done in these areas. Very impressive just to imagine.
Dirac I believe it sells a framework to car manufacturers and does not actively participate in the tuning of each model, Therefore, although the volume of licenses may be interesting, the work behind it should not differ much from that for private customers.
Apart from the world of AVR that surely it will be a big slice, there are also recording studios, which, as far as I know, would have enormous advantage in obtaining a software that provides such linearity of amplitude and phase, as well as perfectly controlled decay. It would be a lot of money and time saved, so even if the request for ART is $ 1000 it would the go to.
I suppose given the sophistication of the software they seek a certain degree of reliability towards the heterogeneity of the environments, otherwise in the face of an inevitable initial high demand (given the premises) they may find themselves with considerable damage to their image if the ineffectiveness and the bugs were high.
I seem to vaguely guess with Bass Control it went like this.
But who knows, we hope to see a spread soon.
The fact is that from the first results published online by third parties, also here on the forum, the linearity that can be obtained without resorting to passive treatments is unrivaled.
Dirac I believe it sells a framework to car manufacturers and does not actively participate in the tuning of each model, Therefore, although the volume of licenses may be interesting, the work behind it should not differ much from that for private customers.
Apart from the world of AVR that surely it will be a big slice, there are also recording studios, which, as far as I know, would have enormous advantage in obtaining a software that provides such linearity of amplitude and phase, as well as perfectly controlled decay. It would be a lot of money and time saved, so even if the request for ART is $ 1000 it would the go to.
I suppose given the sophistication of the software they seek a certain degree of reliability towards the heterogeneity of the environments, otherwise in the face of an inevitable initial high demand (given the premises) they may find themselves with considerable damage to their image if the ineffectiveness and the bugs were high.
I seem to vaguely guess with Bass Control it went like this.
But who knows, we hope to see a spread soon.
The fact is that from the first results published online by third parties, also here on the forum, the linearity that can be obtained without resorting to passive treatments is unrivaled.
Last edited:
Similar threads
