Arcam AV40 AV Processor Review

[Krobar]

What are MDS systems? You mean Dirac wrote the client code for the other manufacturers? What I'm hearing is that all they need to do is write the client code for one API endpoint. That seems trivial?? Like basically enable Dirac to communicate with the unit over the network. There must be a lot more to it. I don't think I quite understand.

As far as I can tell Dirac import part in the Arcam is implemented under the covers as a sort of Google module (Just like a lot of the streaming support).

 

[ririt]

Hi, I just found on a french forum a message highlighting a Q&A exchange between Harman and a customer About the poor performances of the AV40:
Q:here are reviews out in the internet that the AV40 (and all the newArcam AVR's) measured poorly in SINAD- and Jitter tests. Are you awareof this?
A: We have seen the review and notified the reviewer as to exactly what errors they made. One such mistake was that they introduced a ground loop while measuring which considerably raised the noise floor, which is a rookie mistake. They also set the bit measurement to 16 bits rather than 24 bits which gave an incorrect result. Unfortunately, inexperienced reviewers like this cause a lot of damage, but do not own up to their mistakes. We urge customers not to believe everything they read online.
Amirn, can you please comment on the above?

 

[audioBliss]

If Arcam published their own measurements then everything would be a lot easier..I mean it they're such pros then why don't they have a whole PDF of measurements of the unit you can just download from their website?

 

[Sou o teu pai]

Hi, I just found on a french forum a message highlighting a Q&A exchange between Harman and a customer About the poor performances of the AV40:
Q:here are reviews out in the internet that the AV40 (and all the newArcam AVR's) measured poorly in SINAD- and Jitter tests. Are you awareof this?
A: We have seen the review and notified the reviewer as to exactly what errors they made. One such mistake was that they introduced a ground loop while measuring which considerably raised the noise floor, which is a rookie mistake. They also set the bit measurement to 16 bits rather than 24 bits which gave an incorrect result. Unfortunately, inexperienced reviewers like this cause a lot of damage, but do not own up to their mistakes. We urge customers not to believe everything they read online.
Amirn, can you please comment on the above?

So if you measure the way they measure, with their tools and setup exactly the way they have it setup (which is unknown), you'll get Arcam results. Meaning you can just come up with whatever numbers if they fall within your results for "certain" circumstances and context. But then you wouldn't be able to publish those results as part of industry standards now would you?! Results are what they are. If to achieve Arcam's published results you need to follow certain criteria AND those criteria are part of how end users are going to use the product, then Arcam has a valid point. If not...then it's just sad.

 

[Dimifoot]

Hi, I just found on a french forum a message highlighting a Q&A exchange between Harman and a customer About the poor performances of the AV40:
Q:here are reviews out in the internet that the AV40 (and all the newArcam AVR's) measured poorly in SINAD- and Jitter tests. Are you awareof this?
A: We have seen the review and notified the reviewer as to exactly what errors they made. One such mistake was that they introduced a ground loop while measuring which considerably raised the noise floor, which is a rookie mistake. They also set the bit measurement to 16 bits rather than 24 bits which gave an incorrect result. Unfortunately, inexperienced reviewers like this cause a lot of damage, but do not own up to their mistakes. We urge customers not to believe everything they read online.
Amirn, can you please comment on the above?

In that case Arcam has to publish “proper” AP measurements that prove their point, not just vaguely respond to a forum post.

 

[amirm]

Hi, I just found on a french forum a message highlighting a Q&A exchange between Harman and a customer About the poor performances of the AV40:
Q:here are reviews out in the internet that the AV40 (and all the newArcam AVR's) measured poorly in SINAD- and Jitter tests. Are you awareof this?
A: We have seen the review and notified the reviewer as to exactly what errors they made. One such mistake was that they introduced a ground loop while measuring which considerably raised the noise floor, which is a rookie mistake. They also set the bit measurement to 16 bits rather than 24 bits which gave an incorrect result. Unfortunately, inexperienced reviewers like this cause a lot of damage, but do not own up to their mistakes. We urge customers not to believe everything they read online.
Amirn, can you please comment on the above?

Sigh.

Yes, I have their response. I have been holding my response to them confidential as professional courtesy until they get back to me. I am saddened to see them run with their opinion without closing the loop with me.

For now, their assertion as quoted is wrong. There is no ground loop or we would see mains and harmonics in the FFT response:

The noise is broadband and covers all of the spectrum, not just at 60 Hz and multiples as we would see in ground loops. Here is a quick example of mains leakage:

Their testing involve using a portable class 2 HDMI signal generator. Since they don't see the noise issue there, they automatically jumped to the conclusion that there must be ground loops in my testing. I have explained to them that they need to use a standard consumer video source, not a an HDMI signal generator that is likely to produce much cleaner signal. And that there is no evidence of ground loops per above explanation. I have heard no answer from them.

This is on top of the fact that I am using balanced XLR for my testing. So there better not be ground loop anyway.

Communication with them on these issues is difficult because they are using an older Audio Precision analyzer so have no ability to replicate my tests directly. And I am going through an indirect channel to do that instead of speaking on the phone which I have offered to do multiple times.

I have re-tested the unit across the board in response to their feedback and found a couple of things that improves its measured performance. I will be posting those corrections. The overall picture doesn't change much though.

 

[TimoJ]

Their testing involve using a portable class 2 HDMI signal generator. Since they don't see the noise issue there, they automatically jumped to the conclusion that there must be ground loops in my testing. I have explained to them that they need to use a standard consumer video source, not a an HDMI signal generator that is likely to produce much cleaner signal. And that there is no evidence of ground loops per above explanation. I have heard no answer from them.

Did you measure SINAD etc. also via optical input? What is your HDMI signal source?

 

[GXAlan]

Sigh.
Yes, I have their response. I have been holding my response to them confidential as professional courtesy until they get back to me. I am saddened to see them run with their opinion without closing the loop with me.

Though in a bit of fairness, they may feel slighted by you running your review without discussing with them. That said, science allows for peer review and science should allow reviewers to publish independently obtained equipment without running it by the manufacturer. Kudos to you, @amirm. Though, I won't fault Arcam for running damage control. The role of a scientist (@amirm) is to report the facts -- not necessarily to campaign to help others accept the facts.

I've sent a lot of gear to Amir for testing, but for disclosure to the msg boards, I was thinking about picking up an SDP-55 until I saw the software bugs reported on AVS Forum and the AV40 review here. My bias is that I like old school JBL products

Their testing involve using a portable class 2 HDMI signal generator. Since they don't see the noise issue there, they automatically jumped to the conclusion that there must be ground loops in my testing. I have explained to them that they need to use a standard consumer video source, not a an HDMI signal generator that is likely to produce much cleaner signal.

Agreed. The alternative to "standard consumer video source" is for Arcam to prove that it works "perfectly" with Samsung LED walls/QLEDs or identify specifically which premium UHD players provide the optimal performance (which Samsung no longer offers).

Communication with them on these issues is difficult because they are using an older Audio Precision analyzer so have no ability to replicate my tests directly.

Wow. You would think that for a company of Arcam/Harman/Samsung's size, they would always be upgrading to the latest AP analyzer.

I have re-tested the unit across the board in response to their feedback and found a couple of things that improves its measured performance. I will be posting those corrections. The overall picture doesn't change much though.

Looking forward to it. Have your sources commented on whether the AudioControl or JBL Synthesis designs are expected to improve significantly?

 

[amirm]

Looking forward to it. Have your sources commented on whether the AudioControl or JBL Synthesis designs are expected to improve significantly?

The only thing I know is that JBL uses a different DAC which obviously requires a different layout and design in that part. So until we test it, it remains an unknown.

 

[amirm]

Did you measure SINAD etc. also via optical input? What is your HDMI signal source?

I did measure using Toslink. For HDMI source I use my PC HDMI video out (Intel). That way, the AP software sees it as a proper audio device and is able to run full suite of tests -- something Arcam cannot do with their setup.

 

[Krobar]

I would have hoped they could at least fix the filter in firmware since the ESS DACs are often software configurable and maybe the truncation but if they are leading with this response then even correcting what they can seems unlikely.

 

[amirm]

I would have hoped they could at least fix the filter in firmware since the ESS DACs are often software configurable and maybe the truncation but if they are leading with this response then even correcting what they can seems unlikely.

That's exactly the response I gave them. It is not like they don't have a million options already in the unit. Adding a simple one to select a proper filter should be standard among these products. Or else pick a sensible one for default that has science behind it.

On truncation, that is resolved. For some reason the chain had switched to 16 bit mode. I re-ran it now with identical setup and now is running better and in 24 bit mode. It however uncovered another issue with Toslink input. I will post these corrections a little later. I wish these products would display the proper incoming format such as 16, 24-bit, etc. so we know what it is receiving.

 

[Krobar]

The only thing I know is that JBL uses a different DAC which obviously requires a different layout and design in that part. So until we test it, it remains an unknown.

Still really interested to see the Audiocontrol X9 too as that uses the 9038 DACs and different opamps. Just noticed the 9028 in the JBL has a different pinout so there must be at least one different PCB in the JBL. I would assume the Audiocontrol X7 has identical perofrmance.

 

[TimoJ]

I did measure using Toslink. For HDMI source I use my PC HDMI video out (Intel). That way, the AP software sees it as a proper audio device and is able to run full suite of tests -- something Arcam cannot do with their setup.

I can only find jitter results via Toslink, were the other measurements not published or are they pretty much the same? I assume you have somehow verified that your PC HDMI output is "clean" enough not to limit test results? (I know it was not the limiting factor in this case)

 

[amirm]

Still really interested to see the Audiocontrol X9 too as that uses the 9038 DACs and different opamps.

Our company is a dealer for Audio Control. I can reach out to them but don't know the people personally. But I really rather someone buys one to test rather than me asking for a loan with expectations that come with it.

 

[amirm]

I can only find jitter results via Toslink, were the other measurements not published or are they pretty much the same? I assume you have somehow verified that your PC HDMI output is "clean" enough not to limit test results? (I know it was not the limiting factor in this case)

I also ran the HDMI input (but has the 16-bit problem).

And no, I don't know how "clean" my HDMI source is, any more than I know how clean the USB port is for testing USB dacs. I expect the receiving product be able to produce high fidelity analog output no matter how bad HDMI signal is. Do you know how clean HDMI is from your cable box? Blu-ray player?

 

[TimoJ]

And no, I don't know how "clean" my HDMI source is, any more than I know how clean the USB port is for testing USB dacs. I expect the receiving product be able to produce high fidelity analog output no matter how bad HDMI signal is. Do you know how clean HDMI is from your cable box? Blu-ray player?

No, I don't know, mostly use Intel PC HDMI out... But I think if you measure device performance with high end gear, also your signal source should be as clean and high end as possible. Crap in, crap out? I mean maybe have a "high end" source and also a regular one to verify everyday performance. But maybe HDMI is so robust that this actually doesn't matter, I don't know.

 

[amirm]

No, I don't know, mostly use Intel PC HDMI out... But I think if you measure device performance with high end gear, also your signal source should be as clean and high end as possible. Crap in, crap out?

Not at all. I test USB DACs that produce unbelievably low distortion and noise levels with plain "dirty" USB output. Doing otherwise would be creating scenarios that don't match real customer environment.

A hallmark of a high-end audio product is insulation against what comes before it. What are we paying for if it is not superlative filtering and clean implementation?

Now, if the content is encoded poorly, yes, that is crap in and crap out. But the source signal here is digital in nature and hence as perfect as you can get. HDMI delivers that digital sample losslessly so that is accomplished as well. The job of the product then is to extract the digital data and leave behind the noise that is inherent in any external interface.

 

[Lsc]

That's exactly the response I gave them. It is not like they don't have a million options already in the unit. Adding a simple one to select a proper filter should be standard among these products. Or else pick a sensible one for default that has science behind it.

On truncation, that is resolved. For some reason the chain had switched to 16 bit mode. I re-ran it now with identical setup and now is running better and in 24 bit mode. It however uncovered another issue with Toslink input. I will post these corrections a little later. I wish these products would display the proper incoming format such as 16, 24-bit, etc. so we know what it is receiving.

Do you know if your test with the RMC1 was in 16 bit mode or 24 bit mode?

 

[GXAlan]

Our company is a dealer for Audio Control. I can reach out to them but don't know the people personally. But I really rather someone buys one to test rather than me asking for a loan with expectations that come with it.

Interestingly, the AudioControl X9 has a firmware of 1.29 whereas the JBL and Arcam are at 1.28. The JBL/Arcam are 4/28/2020 and the Audiocontrol is 4/29/2020 so it suggests that AudioControl is doing *something* extra.