Is Dirac the frog who aspired to become as big as the ox?

[FargateOne]

I write this post reluctantly, but there comes a time when things have to be said. This is an open letter from a consumer who is disillusioned by Dirac's inability to get its software to work properly, and has been for many years. The situation of the owners of Arcam and JBL is even worse than reported here, and we are still waiting for solutions to many serious problems with the implementation of their software in high-end receivers. In Harman products case bugs are known since more than 2 years now and was begining to be investigated only in 2025...
Dirac is pushing their new ART software in what appears to be a headlong rush into overpriced products to abandon DLBC, which they are unable to make work in the too many brands of (more affordable) products to which they have undoubtedly charged large sums in copyright.
Translated with DeepL.com (free version)

To Whom It May (But Probably Doesn’t) Concern,​

​

For the past several months, users have been reporting a persistent and disruptive crossover bug in Dirac Live when used with PAC products (Onkyo, Pioneer, and Integra receivers). Despite its widespread impact—and despite users purchasing this software at a premium—there’s been no visible action, no resolution, and not even the decency of a timeline. At this point, the silence is louder than the bug itself.​

​

Now, with new leadership in place and Dirac pivoting toward larger, more lucrative ventures—such as its recently announced automotive integration partnership—it’s hard not to feel like the enthusiast AV community is being quietly sidelined. New CEOs aren’t typically brought in to fix legacy bugs; they’re brought in to scale operations, chase bigger markets, and deliver investor returns. We get it. But let’s not pretend this is compatible with continuing to sell a consumer-facing product that doesn’t consistently work.​

​

If Dirac still intends to serve this market—if the AV community isn’t just legacy baggage on the way to bigger deals—then now is the time to demonstrate it. Fixing a crossover bug isn’t a moonshot. It’s basic accountability. These are solvable software issues, and users deserve better than silence.​

​

This crossover bug? It’s fixable. It’s code. All we are asking for is the product to work as described. But that only happens if someone at Dirac is still listening.​

​

Sincerely,​

 

[Asere]

Apparently Dirac and PAC are working on a fix but when is the question. I reached out to Dirac over my LX805 but have not heard from them. We shall see.

 

[Travis Ballstadt]

I may be wrong, but the way I've always viewed Dirac is that they have a technology and a server-side software and they license to hardware manufacturers the API to communicate with their software, but the user hardware-side integration is the responsibility of the hardware manufacturer to develop. If Dirac works with some products but not with others, that seems like a hardware manufacturer responsibility to address.

Similarly, I think we're finding hardware manufacturers discovering that when they licensed the tech, promised customers the feature, and sold them hardware and licenses, they did so before implementing Dirac and later found their hardware to be incapable of performing the task at hand. I've been watching this happen since Emotiva started implementing Dirac in the XMC-1. Big promises, underpowered hardware, shampoo rinse repeat.

Admittedly I don't know anything about the crossover bug mentioned in the first post.

 

[JStewart]

I write this post reluctantly, but there comes a time when things have to be said. This is an open letter from a consumer who is disillusioned by Dirac's inability to get its software to work properly, and has been for many years. The situation of the owners of Arcam and JBL is even worse than reported here, and we are still waiting for solutions to many serious problems with the implementation of their software in high-end receivers. In Harman products case bugs are known since more than 2 years now and was begining to be investigated only in 2025...
Dirac is pushing their new ART software in what appears to be a headlong rush into overpriced products to abandon DLBC, which they are unable to make work in the too many brands of (more affordable) products to which they have undoubtedly charged large sums in copyright.
Translated with DeepL.com (free version)

To Whom It May (But Probably Doesn’t) Concern,​

​

For the past several months, users have been reporting a persistent and disruptive crossover bug in Dirac Live when used with PAC products (Onkyo, Pioneer, and Integra receivers). Despite its widespread impact—and despite users purchasing this software at a premium—there’s been no visible action, no resolution, and not even the decency of a timeline. At this point, the silence is louder than the bug itself.​

​

Now, with new leadership in place and Dirac pivoting toward larger, more lucrative ventures—such as its recently announced automotive integration partnership—it’s hard not to feel like the enthusiast AV community is being quietly sidelined. New CEOs aren’t typically brought in to fix legacy bugs; they’re brought in to scale operations, chase bigger markets, and deliver investor returns. We get it. But let’s not pretend this is compatible with continuing to sell a consumer-facing product that doesn’t consistently work.​

​

If Dirac still intends to serve this market—if the AV community isn’t just legacy baggage on the way to bigger deals—then now is the time to demonstrate it. Fixing a crossover bug isn’t a moonshot. It’s basic accountability. These are solvable software issues, and users deserve better than silence.​

​

This crossover bug? It’s fixable. It’s code. All we are asking for is the product to work as described. But that only happens if someone at Dirac is still listening.​

​

Sincerely,​

I’ve followed your story on the AVS thread. I personally can’t imagine having to deal with software issues for a couple of years as you have. You’re lucky to have used REW to verify the issues. There are probably many Arcam users with no idea.
In thinking about this, I wonder who the responsible party actually is. Dirac measures and creates filters. Why would the filters, or speaker delays and trim levels for that matter, generated within the Dirac software be any different based on the device the files are exported to?
I can’t help but think the issue is with Arcam and other manufacturers, whose job it was to implement Dirac on their devices, which implies ensuring it actually worked. In my opinion, Arcam should be responsible for doing the same testing you have, seeing the same issues, and either fixing it or stop selling their devices as compatible.
Have you reached out to Arcam as well? What do they say?

 

[FargateOne]

I personally can’t imagine having to deal with software issues for a couple of years as you have. You’re lucky to have used REW to verify the issues.

It all beegan with my ears and listening session. It is so important to hear first then measure.
I loaded the first faulty update (3.5.7) in June 2023. Since the very beginning my ears told me that something wasn't right. But you know, thee are so many variables. I began to measure with REW after the third update and asked my first quesstion in dec. 2023. Thanks to Markus767 help I openned 3 tickets to Dirac in March 2024.

Have you reached out to Arcam as well? What do they say?

In January, disgusted to see no real results from Dirac yet, I open a ticket to Harman Luxury January 8th 2025.
Here some quote from Harman since then:

There are probably many Arcam users with no idea.

March 20th 2025:
«To date, you are the only person to have mentioned it to us, so I really don't know where all these worried users are who have been suffering for many years are... they weren't concerned enough to email us about it. »
And:
«It looks like we might have found something, but we're not 100% sure if our testing methodology is correct.»
April 7th 2025
«The development team are investigating it, but I have not heard anything conclusive one way or the other.»

Dirac began to investigate seriously my complaints AFTER my January 8th email.

From Dirac 3 days ago:
«Latest Dirac DLBC update:
Currently we tested:
Behavior of Dirac Live on multiple Dirac Live versions using:

• Audio Precision
• REW
• Internal measurement software to exclude any hardware

There we found that there are some inconsistent behaviors on the Arcam devices with some of our versions.
Back in the day the implementation from our software on OEMs devices wasn’t done in the same way as it is today. It seems that this now affects older devices while using newer Dirac Live versions.
We’re looking into the actual part of our code and the implementation which causes the discrepancies. It only happens on Bass Control so we have to deep dive there.
The issue doesn’t occur on other devices so we need to investigate the reason why Arcam’s implementation and our code is conflicting. Once we have that, we discuss this more with Arcam.
I’m sorry about the long wait for this issue to be fixed but it’s a complex one. We’re working as fast and efficient as we can.»

They can't go fast for sure...

I'm really off about it.

 

[JStewart]

To date, you are the only person to have mentioned it to us, so I really don't know where all these worried users are who have been suffering for many years are... they weren't concerned enough to email us about it.

Is that really a reply from Harman? That’s rude to say the least.

Dirac’s answer implies an older version would be stable. Iirc you’re using 3.4.4 and it works properly?
Of course that knowledge won’t help all users since Dirac offers the most recent version for download when you purchase.
It’s good that they make older versions available as well though, if you know enough to go looking for them.

 

[FargateOne]

Is that really a reply from Harman?

Yes it is and it continued like this March 20th email:«If there is an issue, it is certainly going to be harder for us to commission a software update than it would have been if we had been contacted two years ago when this apparently became a problem.»

you’re using 3.4.4 and it works properly?

None of the update above this v. works with DLBC option. Because the algoritm has changed with 3.10.X we can say «Properly» if you accept the following steps:

1- Install v3.13.2 choose your curve and calculate DLBC filter;​

2- printscreen the predicted response for each groups of speakers;​

3- uninstall current version, re-install the old 3.4.4;​

4- select the same settings and the same measurements project of course and hit calculate button on DLBC;​

5- repeat step 4 as many times ( up to 12 calculations sometimes) needed to get the response already noted ;​

6- IF you finally get lucky load the filter in Arcam...​

Reducing the number of measuremnt to 5-6 and keeping all measurements in an 3D ovale figure of not more 30-45cm long and 15cm large and heath MAY reduce the calculated attemps to 5-6 only...

 

[Asere]

Dirac is likely overwhelmed with this crossover bug issue. I just emailed them and they replied with the ticket has been resolved as done and that was all. Has anyone else received such reply?

 

[FSAF]

Could anyone explain why you would bother with these receivers if you could get a mini-PC for $200, add a 7.1 USB soundcard, and Bob is your uncle?

 

[FargateOne]

Could anyone explain why you would bother with these receivers if you could get a mini-PC for $200, add a 7.1 USB soundcard, and Bob is your uncle?

Sorry I don't understand. Call me ignorant if you wantl. Who's Bob?

 

[FSAF]

Sorry, I thought that it was a well-known idiom. My wrong. https://en.wikipedia.org/wiki/Bob's_your_uncle

 

[FargateOne]

Sorry, I thought that it was a well-known idiom. My wrong. https://en.wikipedia.org/wiki/Bob's_your_uncle

Ok but I'd like to know how your proposal setup replace an Arcam AVR21 with DLBC for 5.1 system or music listening in stereo with 2 subs aligned with satellites at xo?
You mean a laptop would have Dirac for pc? Do we know that DLBC works in a pc? See my long journey and remember that all Arcam/JBL owners were sure that DLBC above 3.4.4 worked and it wasn't. I can trust DLBC predicted response anymore and anywhere...

 

[FSAF]

Arcams, etc have a DSP inside to execute long FIR filters. x86 with AVX/2/512 etc is also a capable FIR executioner. You can, of course, have 8/16 channels audio interfaces like Focusrite 18i20 (or 3x cheaper alternatives) to take in many channels with ASIO, filter/combine them, and output the audio the way you like it. If you have a stereo input and want to get a 7.1 output, you may get away with $100 Creative sound card. Everything will be under your full control.

There are a few programs to run FIR /etc for audio if you don't want to bother coding, like Equalizer APO.
You do not need a dedicated laptop; a mini PC with N100 ($200) will suffice.

 

[JStewart]

Could anyone explain why you would bother with these receivers if you could get a mini-PC for $200, add a 7.1 USB soundcard, and Bob is your uncle?

For music only, sure. As soon as you start adding more devices to switch between, Dolby and DTS decoding, have a TV hosting apps with audio via ARC or eARC or Atmos it starts to become a complex project or in some cases perhaps not even possible. Add to that significant others who will need to know how to use it and our strong desire to sleep within the shelter, then at least in my case, it’s a hard no.

 

[FSAF]

That is a good point.

On the other hand, you get incompatibility between scientists and engineers, their total inability to work together, and the resulting rotating door employment practices precluding any support, "...and the weak will suffer" where the weak is a customer. That's besides the industry being a complete rip-off. Just look at the prices of TI / Analog DSPs and compare them with receivers' prices.

 

[FargateOne]

On the other hand, you get incompatibility between scientists and engineers, their total inability to work together,

This is exactly why Dirac is in danger of losing its market. Their software is sophisticated and requires standards to be respected, which don't exist for the following reason:

That's besides the industry being a complete rip-off

 

[Asere]

Also Dirac like with any software is always going to have glitches, updates for the better and worse, etc. There won't ever be a stop to the updates unless the company goes out of business.