Denon AVR-X3800H Review

[Brian6751]

THD+N and hence SINAD is a fundamental performance metric for any audio device. Consumers won't care if no one provides it. But when someone like me does, it impacts a segment of their market. Why else were they unhappy with my measurements in the past if it didn't matter?

because you showed them that something was broken.

If better measurements had a positive impact for the company why would they not continue striving for as good or even improve them? That dog dont hunt

 

[amirm]

Look folks: I already created a special lower tier of performance for AVRs. I no longer slot them in the desktop products. I have also substantially curved my requirement for AV products to get my recommendation. But I am drawing a line in taking a design and substantially lowering its performance. The argument that "these are complex AV products so can't have XYZ performance" no longer holds. Their own product from last year and the year before had far better performance. They chose to go to market with far lower performance with much higher retail cost. There is no logical stance for me to give a pass to this.

 

[hmt]

The main problem here is that you pay more for something that is the same from previous generation. Yeah, you pay for a promise of Dirac, software tha you´ll have to pay extra to get.


Why should I get a Denon for Dirac and pay extra when I can get something that includes Dirac already? Because as a customer, that would be my main reasoning discounting all tech jargon.

You pay more for every AVR than 3 years ago. And yes you coul get something with dirac already included. But the price will be higher.

 

[amirm]

If better measurements had a positive impact for the company why would they not continue striving for as good or even improve them?

I can't speak to their skills in product management and management in general. Maybe folks just found out that the performance has gone backward this much. Maybe there are talks to deal with it. I expect some kind of PR response first but hopefully there will be a technical solution that brings the performance back to top.

Let's remember: it is not our job to solve the manufacturer's problems. We need to make our desires clear and see who delivers, and who doesn't.

 

[CyrusTheGreat_600BC]

It is acceptable as long as the competition don't offer better products, as unfortunate as it may be.

Onkyo probably already finalized RZ70 design, but now they have a good chance to give something better than Denon and escape bankruptcy.

 

[Brian6751]

... The argument that "these are complex AV products so can't have XYZ performance" no longer holds. ...

I dont think thats the argument. I think the argument is to label a product that does so many things so well unacceptable or not recommended over something that no-one can hear begs the question, Why?

Just to put pressure on the manufacturer to improve said inaudible measurement does not make sense to me either.

 

[amper42]

What did we learn? Buy a slightly used or refurb 4700 for 40% less than the 3800 for better performance.

Masimo needs to hire someone at Denon to complete measurements before blessing prototypes for production.

 

[dlaloum]

Sure, a second hand computer (say) with HDMI out and the appropriate software loaded would also be required. Maybe another $1k-$1.5k, but still insane value with massive functionality and performance when compared to some of the HiFi jewellery being espoused by so called "up market" manufacturers.

Depends on what you value, bells and whistles or good technology producing good sound. Both have their place. My preference would be the latter.

Still missing the cost of software licencing...

 

[dlaloum]

If I remember right, one user confirmed that it could be done on individual channel pair (even the center, presumably) basis. The manual isn't clear on this but that's just my opinion.

At board level, the 4700 has seperate amp boards per channel. - This is an architecture that would lend itself to turning off channel pairs
The 3800 has monolithic power boards with multiple channels (2 boards I believe) - I doubt they can be turned off per pair.... perhaps idled?

 

[Brian6751]

I already see people talking about returning their 3800, or cancelling their order, or choosing models with less options based on this "not recommended" result over something not even audible. Most of the time they spend the same amount of money or more. These are the people who are paying more and getting less. How does this help them?

 

[SKBubba]

How many of these things does Denon sell?

This was a serious question. Is it 1000 units? 10,000? 100,000? 1,000,000?

Just curious about economies of scale for parts (like dacs) and tooling and licenses and whatnot.

 

[oivavoi]

I dont think thats the argument. I think the argument is to label a product that does so many things so well unacceptable or not recommended over something that no-one can hear begs the question, Why?

Just to put pressure on the manufacturer to improve said inaudible measurement does not make sense to me either.

Agreed. And just to point out to @amirm that I see great value in the SINAD approach, as do many others! - even when it goes beyond audibility. Getting numbers on that in such a clear way has been a great service to the audio community.

The issue I may have here rather has to do with how much importance SINAD numbers get in the grand scheme of things when it comes to recommendations. It's perfectly fine to say that SINAD numbers have gone down, and that's not optimal. That in itself would send a signal. But then add that it's still probably audibly transparent, and it delivers some functional improvements, so, whatever, it will get a pass.

But that's obviously your call Amir
Ok now I'll stop discussing things on the internets and practice some piano instead.

 

[Vader540is]

"Time to return my X3800"

Lol

 

[dlaloum]

It does:

That doesn't tell you whether those channels are actually powered down, or merely "idled" ... measurements are needed to determine that!

 

[amirm]

I dont think thats the argument. I think the argument is to label a product that does so many things so well unacceptable or not recommended over something that no-one can hear begs the question, Why?

What does it do so well and by what standard? Just the wet thumb in the air that its objective performance is good enough?

Should I stop testing them then and you all just tally the feature list and number of logos on the front as measure of how good it is?

 

[Brian6751]

"Time to return my X3800"

Lol

Seriously! people are doing it! Im like what are you doing!!! Stop!!! LOL

 

[dlaloum]

I somewhat disagree. AVR is a very complex product that engineers have to make compromises. whether "acceptable" or not is up to the market to decide (hint: 90% if not more people do not care about SINAD, features, price and brand would triumph anything else)

Denon have already demonstrated that they can do better... they have done better in the 3700.

A marketing decision was made to bring the new series to market based on a specific timeline... the AKM chipsets would not have been available on that timeline... so they compromised.

Alternatively they could have re-designed the DAC boards for a higher performance non AKM chipset - but they went the path of evolution.... so stuck with their existing solutions.

Hence performance is the same as the late versions of X3700....

Most of the complaints can be put down to the fact that the early production of the X3700 used higher performance AKM chipsets... and caused quite a furor through the community due to its (relatively) high performance.

 

[amirm]

The issue I may have here rather has to do with how much importance SINAD numbers get in the grand scheme of things when it comes to recommendations. It's perfectly fine to say that SINAD numbers have gone down, and that's not optimal. That in itself would send a signal. But then add that it's still probably audibly transparent, and it delivers some functional improvements, so, whatever, it will get a pass.

You can say it is "probably" transparent. I can't do that. For me to do that there needs to be study of trained listeners, with specific content aimed at showing these types of distortion. Such a study doesn't exist. Nor has the manufacturer provided one. The only alternative is then to aim for objectively transparent response which calls for absolute noise+distortion floor that is at threshold of audibility. 87 dB SINAD does not remotely get you there. Neither does SNR of 107 dB by the way.

As I noted earlier, I had already caved into having far lower bar for AVRs. What you are asking is that said bar be knocked down by another 10 to 12 dB. How does this help anyone as a consumer? Did you get a discount in exchange for that? It makes no sense to me.

 

[Brian6751]

What does it do so well and by what standard? Just the wet thumb in the air that its objective performance is good enough?

Should I stop testing them then and you all just tally the feature list and number of logos on the front as measure of how good it is?

What does it do so well and by what standard? Just the wet thumb in the air that its objective performance is good enough? Everything that it did just as well as the 3700 that was so highly praised, plus four independent subwoofer outs AND you can even use the #4 sub out as a dedicated transducer preout

Should I stop testing them then and you all just tally the feature list and number of logos on the front as measure of how good it is? No. No one is saying that. I am extremely thankful for the work you do. What we are saying is that if we are going to gig disregard a audio product for measuring badly in a area, it should at least be audible.

 

[PeteL]

This was a serious question. Is it 1000 units? 10,000? 100,000? 1,000,000?

Just curious about economies of scale for parts (like dacs) and tooling and licenses and whatnot.

It will be in the 5 Digits and start by 1 or 2