Denon AVR-X4700H 2020 AVR Review

[carlob]

What they need to do is allocate less budget to the marketing dept and more budget to the engineering dept.

 

[SimpleTheater]

True, but again, as I asked before, can we realistically expect the SR6013 to perform better than the more expensive AV7705 that was measured? The 7705 couldn't do better than 75 dB SINAD if I remember right, at 4 V balanced), that is 2 V unbalanced. It only measured in the 90's when Amir made the exception on request to drop it to 2.4 V balanced, that is only 1.2 V unbalanced.

But I’m not interested in anything but the amp. Does the 7 channel line-in go straight to the amps or does it do processing. My assumption was the prior.

Edit: Found this in the Marantz SR-6012 manual “The sound mode cannot be set if the input mode is set to “7.1CH IN”.

 

[peng]

In post#97, I posted the specs of the volume control chip used in both Denon and Marantz 2016 through 2019 units, likely the same for the 2020s:
I included one for the Yamaha's just for reference (as obviously the same logic would apply to theirs too) and hope to see others comments.

My question is, those chips have a wide tolerance in THD, for the D+M, it could be typical at 0.0004%, but with a little bit of bad luck, one could end up with a unit, say the AVR-X4700H that got worse SINAD, getting a chip with less than typical THD, that is closer to the Max 0.01%, or -80 dB, am I right or wrong?

The NJU72343 is just one component, but being the volume control chip that is in every single signal path, so in theory the measured THD+N cannot be better than the THD spec of this chip, obviously.

I am an EE but I shamefully, I can't answer my own question so I am hoping experts such as @amirm , @restorer-john , @bigguyca can shed some light on this. Sorry for paging you guys, I am just too curious and anxious to hear some expert opinions as the question is obvious, why would the AVR-X3600H perform better when it has the same volume control chip, DAC, ADC, switches and all the key components in the signal chain?

So basic on such logic I see the two possible explanations (not limited to), i.e. parts tolerances, and the different HDMI board. The 2020 models must have a different HDMI board right?

https://www.njr.com/electronic_device/PDF/NJU72343_E.pdf

Yamaha A-S801:
BD3483KS2........................... Typ. 0.0004 %, Max. 0.02 %

RX-A3070:
BD34703KS2........................ Typ. 0.0004 %, Max. 0.02 %

Denon/Marantz:
NJU72343AV........................ Typ. 0.0004 %, Max. 0.01 %

 

[peng]

But I’m not interested in anything but the amp. Does the 7 channel line-in go straight to the amps or does it do processing. My assumption was the prior.

Okay so you only care about the analog inputs, specifically the 7.1 inputs. Unfortunately it does not go straight to the amp, it has to go through the preamp, that is mainly the volume control IC. I know because I have seen the schematics of quite a few D+M units including the one for the AV8805 that you can download, or could anyway when it was publicly available for a long time.

That chip is most likely the bottleneck as emphasized more than once by Dr. Rich of HTHF, I posted links to the related articles several time before if you are interested. Also, please refer to post#407 to see the THD specs of that chip.

 

[RichB]

You are right if you include the other measurements that Denon+Marantz has not included, my comments were explicitly based on comparing with their published specs.

Now that you mentioned the multi-tone test, do you know what Amir meant when he said:
"There is a common processing library these companies are using that causes multi-tone test to also be reproduced horribly...."

To me, like the Arcam's weird behavior in the significant variance in SINAD measurements that seemed to be affected by whether only the front two speakers were selected or with more speakers selected (i.e. non none..) So it's not clear cut to me if such weird measurements mean anything bad audibly speaking.

Regarding the CD's 0.0015% requirement, how many "separate" preamp/amp measured so far managed to pass that criteria? On the other hand, there are more than one AVR or streamer that seemingly came, or could come very close under certain conditions. I just find it odd that we (I include myself, just to appear unbiased, tend to complain loudly about the receivers and streamers poor performance as though they are not legit device for serious stereo music listening, implying stereo integrated amps and/or separates are the only viable options. Measurement on AH and elsewhere seem to tell a different story if we step back a little and look at all of the measurements objectively, as you obviously have done.

It is not particularly difficult to meet the 0.0015 standard in amps, especially if you are willing to move the Class-D.
According to ASR, apparently my electronics do.
Apparently, all humans are looking for validation

The multi-tone test (MTT), intended to be representative of music, it ridiculously bad.
One can argue at -76 dB audibility but not -40 dB.
It remains clear (to me anyway) that tests are performed only data points to determine transparency and performance.
Limited tests are performed that are intended to b indicative of performance.
Every so often a new test is published that proves to be disruptive.

Very few (and perhaps no other) sites perform a MTT.
Consider an that owner says: for some reason my AVR using HDMI does not sound as good as my DAC.
The chorus would begin showing measurements from multiple sites and claiming no audibility, the lack of DBT tests, etc.
The trouble is, the owner would be right.

Determining audibility is bitch. That is why there are high-performances measurement devices.
Some manufactures run MTT to evaluate performance before they ship.
Others just run the measurements that reviewers may run.
Essentially, teaching the student the test instead of striving for inaudibility.
There evidence is that Denon is doing the later.

ASR, Stereophile, AH has measured preamp outputs but most professional reviews do not.
ASR in particular has upheld the 2 volt and 4 volt output criteria.
This is not rocket science, all manufacturers are aware of this and yet, some chose to ignore them.
Oppo did not ignore them on their $1200 players.

I used to believe that AVR/AVPs did not matter because their performance would exceed my amplifiers.
It may be that cost cutting is part of it, but also consider Marantz HDAM which clearly costs more and performs worse.
Perhaps, the designers don't believe in measurements and think that high-current and slew rate are more important.
Perhaps, marketing has a Marantz differentiator to provide a nice photo and story for the website.
After measurements have emerged, HDAM and copper chassis seem to have been de-emphasized.
I am not a EE, but it seems like the money could be better spent to provide a measurable improvement and use that for the marketing pitch.

- Rich

 

[Jimbob54]

I remember when Harman Kardon had the most "luxury" power amplifier in the world. It was AU$15,800 in 1983. Showing my age.

View attachment 69775

Everybody loves the Vintage Knob. In Yorkshire, we could call him old cock.

 

[SIY]

May I toss out a philosophical question: what should a reviewer do when a product is poor? Contact the manufacturer directly and ask wtf is going on with this? Confirm that the product is functioning normally for it, then publish, good or bad? I've done both and gotten burned. Other alternatives?

BTW, Amir is the only reviewer I'm aware of (besides me, of course ) who will share project files so people can independently replicate his measurements.

 

[IslandMD]

But I’m not interested in anything but the amp. Does the 7 channel line-in go straight to the amps or does it do processing. My assumption was the prior.

Peng is correct. The multichannel 7.1 inputs go through the volume control and then to the amps. The 7.1 inputs are never digitized.

Likewise, the analog inputs (like CD, etc) are also not digitized in the Direct (or Pure Direct) mode. That signal stays analog and only goes through the volume control. However, if you switch to the Stereo (or any other) mode, the signal will be digitized in order to allow for processing. The sonic degradation caused by ADC and processing is clearly noticeable.

As Amir mentioned in his review the real problem with these AVRs is the digital subsystem. I can tell you from a lot of experimenting with 2 channel stereo gear and AVRs in my home theater that getting the best 2 channel sound means avoiding the digital subsystem of the AVR. My Denon AVR-4311's analog preamp as well as it's internal amps sound very good. To me they are essentially indistinguishable from my expensive stereo integrated amp (April Music Stello Ai500SE). Consider the following comparison that I've experimented with:

1) Topping D70 > XLR > Stereo Integrated Amp

2) Topping D70 > RCA > Denon CD input in Direct mode > Denon's Internal amps

Comparing these two setups, played into my Aerial Acoustics 7T speakers I can't hear any significant difference at all at.

 

[RichB]

Regarding the CD's 0.0015% requirement, how many "separate" preamp/amp measured so far managed to pass that criteria? On the other hand, there are more than one AVR or streamer that seemingly came, or could come very close under certain conditions. I just find it odd that we (I include myself, just to appear unbiased, tend to complain loudly about the receivers and streamers poor performance as though they are not legit device for serious stereo music listening, implying stereo integrated amps and/or separates are the only viable options. Measurement on AH and elsewhere seem to tell a different story if we step back a little and look at all of the measurements objectively, as you obviously have done.

Rarely are AVRs measured to the same standards as stand-alone amplifiers.
I have auditioned a Yamaha $300 AVR (I can't remember the model) driving the Salon2s and, subjectively thought it was lacking dynamics even at low levels. I suspect the linearity and distortion profile changes driving the difficult loads like the Salon2s.
I could have done more but, honestly, would anyone recommend driving a demanding load?
When driving Harbeth 30.2 (minimum impedance 6 Ohms) in a small room and moderate levels (say 80 dB), perhaps that would provide good performance.

Every device has its use and the educated end-user must decide which devices suit their needs.
Some are looking for absolute performance but most are consumers just want to select a component that for their system.

I was considering a Denon AVR-X3700H to replace the Onkyo in a family vacation home, but it may be that the old Onkyo has a better preamp out section than the new model. Older Onkyo's has been known to have HDMI issues but this one has been very reliable. In a vacation home, this is critical. If it dies, I may still buy the Denon for reliability.
The ability to turn off the amps, is a great feature in home without air conditioning.
Were it not for the multi-tone test results, I may have already ordered one.

- Rich

 

[RichB]

Everybody loves the Vintage Knob. In Yorkshire, we could call him old cock.

I love the packing crate.
https://www.audiosciencereview.com/forum/index.php?attachments/1592608250085-png.69775/

- Rich

 

[peng]

The multi-tone test (MTT), intended to be representative of music, it ridiculously bad.
One can argue at -76 dB audibility but not -40 dB.
It remains clear (to me anyway) that tests are performed only data points to determine transparency and performance.
Limited tests are performed that are intended to b indicative of performance.
Every so often a new test is published that proves to be disruptive.

My point is, why is it possible for the AVR-X3600H to measure so much better in that same measurement? 18 to 20 bits vs 7 bit, that's huge!!
So my logical deduction is that there is something wrong with the AVR-X4700H's HDMI board, if I am right, then we probably shouldn't consider the X4700H as a good or even fair representation at all. Again, if you compare the parts in the signal chain, the X4700H should be equal or better, unless that different HDMI board is the one causing all the troubles, or as I am suspecting, parts tolerance might have played a part, but that could only explain the SINAD in the single tone measurements.

 

[A.West]

May I toss out a philosophical question: what should a reviewer do when a product is poor? Contact the manufacturer directly and ask wtf is going on with this? Confirm that the product is functioning normally for it, then publish, good or bad? I've done both and gotten burned. Other alternatives?

BTW, Amir is the only reviewer I'm aware of (besides me, of course ) who will share project files so people can independently replicate his measurements.

Amir runs a standardized, low cost, high throughput operation, testing as many products in a month as any other operation publishes in a year. For those interested in the things he does test and the way he tests, it's great. If he tested everything and had to go back and forth with manufacturers we would have much fewer reviews. Members may also be less willing to send in equipment if it's tied up for months. You can't make everyone happy all the time, but Amir doing it his way is working for a lot of people.

 

[peng]

I was considering a Denon AVR-X3700H to replace the Onkyo in a family vacation home, but it may be that the old Onkyo has a better preamp out section than the new model. Older Onkyo's has been known to have HDMI issues but this one has been very reliable. In a vacation home, this is critical. If it dies, I may still buy the Denon for reliability.
The ability to turn off the amps, is a great feature in home without air conditioning.
Were it not for the multi-tone test results, I may have already ordered one.

- Rich

Until we hear back from Denon, I would not touch any of the 2020 models that logically speaking, would likely perform similarly bad if HDMI input is used. The AVR-X3600H would be a much safer bet, and the X4500H too (but that's just my guess based on logically thinking).

 

[Aerith Gainsborough]

As Amir mentioned in his review the real problem with these AVRs is the digital subsystem.

Except the whole point of going the AVR route are the digital subsystems because you need them for movies, convenient digital connections and room correction.
If a person is not interested in that, an analog poweramp combined with a highend multichannel DAC is the better choice.

To buy an AVR and essentially bypass everything it has to offer seems nonsensical to me.

BTW: I agree with @peng in respect to the multitone test. This may be specific to this single unit. At least I hope so because that -40dB result is truly horrible.

 

[Skylarlove1999]

May I toss out a philosophical question: what should a reviewer do when a product is poor? Contact the manufacturer directly and ask wtf is going on with this? Confirm that the product is functioning normally for it, then publish, good or bad? I've done both and gotten burned. Other alternatives?

BTW, Amir is the only reviewer I'm aware of (besides me, of course ) who will share project files so people can independently replicate his measurements.

It is definitely a tough question to answer. Especially once you yourself know about the poor measurements. Especially since manufacturers are trying to provide what they consider an acceptable noise threshold that they feel does not harm the audible sounds playing through their product. I am thankful that Amir and other posters have explained in detail the benefit of hitting certain noise levels because there is no doubt that the noise cannot be heard, where as above that level it could begin to be audible to some people. I think it is a matter of every improvement comes at a cost even if per unit it is only pennies more to achieve the desired measurement. But how many improvements are needed for both video and audio . Even pennies and up quickly per unit, per shipment , per product line, per year . I am not defending manufacturers. But the reality is people will not buy something when the price increases on even small ticket items. It is called psychological pricing, the mind has already accepted previous prices as a reference point and conceptually the mind has a hard time breaking through that price barrier. A simple example is ice cream price and volume. Ice cream typically was sold a decade ago in 64oz volume containers. The acquisition price for milk caused the price of ice cream to increase. But people's minds couldn't accept paying more so manufacturers lowered the volume to the new standard ice cream container of 48oz . The containers looked so similar people didn't really notice. And why? Because the price actually went down breaking past the price barrier in a positive way. There have been numerous studies confirming this phenomenon when introducing new prices. Same with peppering the box with all those stickers about 4K, HDR, DTS:X, Atmos , etc. The brain sees all those colorful stickers and believes it is getting all of that so it views it as a "higher "value than the box without all those fancy marketing stickers. The point of all of this is that there are reasons why more money is directed towards marketing than engineering. People's brains aren't yet familiar with the numbers that prove this unit sounds better than another so you can't slap DAC and SINAD numbers on the side of a box and get the same response from most consumers brains. So consequently the manufacturer doesn't spend time measuring for those the way an engineer like Amir does. Amir is absolutely correct in what he is fighting for to improve products . It is an uphill battle, not with the engineers but with the marketing department. So I don't know how you convince manufacturers to improve the DAC and SINAD numbers when it won't help sell more units to the general public. Not saying this is right but it is the reality. Educating people like Amir is doing is a step in the right direction because the more people who understand the numbers they will begin to demand those numbers to improve and maybe someday you will see those numbers on the sticker and beloved by the marketing department. It will be a long uphill climb.

 

[SIY]

Good engineering costs no more than poor engineering.

 

[Skylarlove1999]

Good engineering costs no more than poor engineering.

Good engineering can require more time , more research and more development of the product. Costs increase with all three. Not trying to be argumentative but good engineering does have an added cost. That cost may not translate into any higher production cost. That could absolutely be true. How you arrived at a better engineered product that costs the same as the poorly engineered product does have its own inherent cost. One could argue that you make up for the additional expenditure in the R and D phase by selling more units of the better engineered product but unfortunately that hasn't been proven to be true in the world of AV. Just look at how successful Bose has been in fooling people for decades. Once again not trying to be argumentative. I am not an engineer. I know most of you here have way more technical expertise than I will ever acquire. Speaking just from the cost perspective of engineering a better product. Not saying it isn't a worthwhile endeavor but companies are generally more concerned about their bottom line than the DAC and SINAD numbers destroying the competition.

 

[SIY]

Good engineering can require more time , more research and more development of the product.

Nope, that's the opposite of good engineering. And the proof is the cheap products with amazing performance. Smart engineers are painfully aware of the cost of time-to-market, and run their projects accordingly.

 

[ririt]

Peng is correct. The multichannel 7.1 inputs go through the volume control and then to the amps. The 7.1 inputs are never digitized.

Likewise, the analog inputs (like CD, etc) are also not digitized in the Direct (or Pure Direct) mode. That signal stays analog and only goes through the volume control. However, if you switch to the Stereo (or any other) mode, the signal will be digitized in order to allow for processing. The sonic degradation caused by ADC and processing is clearly noticeable.

As Amir mentioned in his review the real problem with these AVRs is the digital subsystem. I can tell you from a lot of experimenting with 2 channel stereo gear and AVRs in my home theater that getting the best 2 channel sound means avoiding the digital subsystem of the AVR. My Denon AVR-4311's analog preamp as well as it's internal amps sound very good. To me they are essentially indistinguishable from my expensive stereo integrated amp (April Music Stello Ai500SE). Consider the following comparison that I've experimented with:

1) Topping D70 > XLR > Stereo Integrated Amp

2) Topping D70 > RCA > Denon CD input in Direct mode > Denon's Internal amps

Comparing these two setups, played into my Aerial Acoustics 7T speakers I can't hear any significant difference at all at.

is there clear evidence from pictures & information provided by Denon that analogs input are not digitized in Direct or Pure Direct mode?

 

[IslandMD]

Except the whole point of going the AVR route are the digital subsystems because you need them for movies, convenient digital connections and room correction.
If a person is not interested in that, an analog poweramp combined with a highend multichannel DAC is the better choice.

To buy an AVR and essentially bypass everything it has to offer seems nonsensical to me.

BTW: I agree with @peng in respect to the multitone test. This may be specific to this single unit. At least I hope so because that -40dB result is truly horrible.

I fully agree. The digital bottleneck in these AVRs has to be fixed or you're missing out on the main reason AVRs exist in the first place.

Regarding these substandard digital subsystems, I've noticed a trend in Amir's AVR reviews that correlates well with my own listening tests. The results on the coax input are almost always better than the HDMI input. In my own listening tests I've noticed that audio quality seems to follow this pattern from best to worst: DLNA Network streaming > coax > HDMI. It's nice to see that something I've been hearing for years is explainable and correlates with Amir's objective measurements.