Denon AVR-X4700H 2020 AVR Review

[lashto]

Fair question, not sure what the answer should be, but I'll try 30 to 60% may be?

how about -40dB fair ?

 

[raistlin65]

I don't want to disclose what I know but there is a distinct difference between 3700 and 4700. When I get a 3700, I can verify if that solves a few of the issues.

Without revealing any details, could that "distinct difference" be addressed by firmware update, or only through hardware change?

The reason I ask is that I have a new X4700H unopened still in the shipping carton. Out of protest, I'm willing to return it. I can wait if Denon can't address this, instead of contributing to their sales numbers for the X4700H.

 

[lashto]

Woah hang on there, chief. Where are you seeing -40dB “distortion” measured for the x4700?

glad you asked
First post. On the first page. Multitone distortion at -40dB. Actually, I might be wrong and it's -39dB. Sorry for my aging eyes.
And even using some of the "better" numbers like the 75dB SINAD, your car analogy is still off by a level of magnitude. It'll need about 2minutes in that case. Sorry, your "latest Camry" is still thoroughly broken.

If you don't mind me asking, why did you think people were calling the x4700 catastrophe and trash and so on?!

 

[Archaea]

I don't want to disclose what I know but there is a distinct difference between 3700 and 4700. When I get a 3700, I can verify if that solves a few of the issues.

Any gut feel on where this possibly distinct difference might fall in relation to the x6700h design?

 

[A.West]

Without revealing any details, could that "distinct difference" be addressed by firmware update, or only through hardware change?

The reason I ask is that I have a new X4700H unopened still in the shipping carton. Out of protest, I'm willing to return it. I can wait if Denon can't address this, instead of contributing to their sales numbers for the X4700H.

I bet the best case is that it would take 2 months for Denon to fix and for Amir to confirm something is improved. That's only if a firmware fix is possible. So you may want to return during your return window, especially if ita free return. There is always something new for audiophiles to lust after. Or you might buy it 6 months from now at a lower price.

 

[RichB]

Agreed to both of you because a) preload is right but only if Amir or Denon can explain why the 32 tone test results and the linearity (to lesser extent) were so bad, and b) if a) is explained and would not have been bad, but we still focus on the "engineering excellence part..".

Otherwise, i.e. if a) is in fact not so bad as it appeared for valid reasons, then while I fully agree we as hobbyist are doing the right thing complaining about the poor results of a higher/more expensive model, we should make some effort in the way we complain, that it may (and evidently it has) lead less informed/technically less knowledgeable potential buyers and owners (worst in that case) to think that this is in fact a product to be avoided, and may even end up buying a competing product that could have been shown worse, if measured by ASR.

So to summarize, I think it is good that we complain like we have been to Denon, but do it in a more balanced way. For example, we can say 75 dB SINAD using HDMI input is much worse than the lower model and totally unacceptable, but it is as good as some separate processor only device such as the AV7705 (when measured at 4 V XLR), to put things in perspective, etc. etc..

Virtually all owners will use HDMI with the 4700H and something is amiss.
Until this is found the be an unusual case, not indicative of normal use, then no criticism can be considered unfair, IMO of course.

- Rich

 

[preload]

glad you asked
First post. On the first page. Multitone distortion at -40dB. Actually, I might be wrong and it's -39dB. Sorry for my aging eyes.
And even using some of the "better" numbers like the 75dB SINAD, your car analogy is still off by a level of magnitude. It'll need about 2minutes in that case. Sorry, your "latest Camry" is still thoroughly broken.

If you don't mind me asking, why did you think people were calling the x4700 catastrophe and trash and so on?!

That's what I suspected you were looking at. I don't think you're interpreting the IMD chart correctly. Look at the x-axis - it's "generator level." The IMD is only -40dB when the digital input level is -47dB. To use your example of recorded explosions in a movie, where the digital input level would be around 0dB, the IMD is actually -82dB (=not audible). Also, Amir's annotations indicates noise is an issue, so it's also not even clear whether -40dB at very low input levels is a function of IMD at all or if it's a function of noise (which would be much more benign). This is how I interpret the chart, but perhaps I'm mistaken.

 

[raistlin65]

I bet the best case is that it would take 2 months for Denon to fix and for Amir to confirm something is improved. That's only if a firmware fix is possible. So you may want to return during your return window, especially if ita free return. There is always something new for audiophiles to lust after. Or you might buy it 6 months from now at a lower price.

I know. But if it's not possible to fix without a hardware update, that pretty much makes the decision for me.

 

[amirm]

Without revealing any details, could that "distinct difference" be addressed by firmware update, or only through hardware change?

If I were to make a wild guess, my answer would be yes. But please don't set your expectations one way or the other. It is too early.

 

[amirm]

Any gut feel on where this possibly distinct difference might fall in relation to the x6700h design?

I did not ask that question but I assume the 6700 is the same as 4700 in this regard.

 

[davidc]

First poster.

Will anyone be sending Amirm a 6700H to test against the 4700H?

I'd be interested to know the difference. The 6700H is assembled in Japan I believe?

I may, but it could be towards the end of the year.

 

[lashto]

That's what I suspected you were looking at. I don't think you're interpreting the IMD chart correctly. Look at the x-axis - it's "generator level." The IMD is only -40dB when the digital input level is -47dB. To use your example of recorded explosions in a movie, where the digital input level would be around 0dB, the IMD is actually -82dB. Also, Amir's annotations indicates noise is an issue, so it's also not clear whether -40dB at very low input levels is a function of IMD or if it's a function of noise (which would be much more benign). This is how I interpret the chart, but perhaps I'm mistaken.

Not exactly sure what you are trying to say.
The multitone graph I am looking at has frequency on x-axis and level on Y-axis. And look at the text on that graph, it says "horrible 7 bits of distortion-free range". 7bits ~= 40dB. (actually, I think 6 bits is a closer number but I guess someone was trying to be extra 'fair')
And yes, that measurement/result is quite strange and still an unexplained mistery.

 

[davidc]

I have just joined the site after purchasing my new Denon X4700 and I made an “unboxing video with a nice twist. My first you tube video I have ever made so don’t be too hard on me. I hope the audio quality will be firmware patched cause I’m just going to stick with the Denon as I already paid for it so her it is everyone.

Awesome motorcycle, annoying music (sorry).

What was the "nice twist"?

 

[peng]

how about -40dB fair ?

As I said, that was terrible, or did I say horrible? I am not sure if Amir is 100% sure about that particular test, he did say something about it, alluding to something.. So I think the jury may still be out on that specific test, if the results was truly -40 dB, or even - 60 dB, I would not recommend the 4700 at all, fair or not. My comments about "fair....". Put it this way, if RichB call it a trash, which he has not, I would say fair, because he specifically cited the 32 tones test results. I think you know what I have been trying to say, if not, then I think there is nothing I can add to change your mind.

 

[davidc]

Thanks for posting these. I see TI PCM5100A as the DAC chip.
View attachment 70448
I don't know if they are used for all the channels or just the non-main ones. They are pretty low spec DACs from TI:

View attachment 70445

Change the sign from negative to positive in the last column and you get SINAD. They opted for the lowest grade of this DAC chip getting it only 90 dB SINAD. Here is the actual graph akin to THD+N vs level that I show:

View attachment 70446

So it can get a bit better than 90 dB which is what I showed.

This very much verifies my measurements to be correct as I got 92 dB with Toslink input.

With CD having a dynamic range of 96 dB, clearly even in ideal case this AVR can't resolve CD let alone higher resolution.

The DAC chip is stereo so gives you two channels and it costs just $1 in thousand piece quantity. In other words, 50 cents per channel was invested to produce sound out of this AVR.

The above would be fine if we didn't have flowery words about how great its audio subsystem is. But we do have such words:

View attachment 70447

Earlier we talked about reference levels of 105 dB, etc. You can't get to those numbers with 16 bits which caps out at 96 dB.

We need an honest, 18 bits at least to sell products like this with a straight face. 15 bits and lower just doesn't do it.

Now, if we had a $300 AVR, that would be fine. But when charging over $1000 let's deliver the bits to people please.

So, this also means that this AVR cannot fully resolve hi-res audio streaming such as TOTAL or Amazon Music hi-res?

 

[peng]

Virtually all owners will use HDMI with the 4700H and something is amiss.
Until this is found the be an unusual case, not indicative of normal use, then no criticism can be considered unfair, IMO of course.

- Rich

I absolutely agree with you, the 32 tones test results is prohibitively bad, but it helps that at least you included a condition, not a blanket condemnation that would otherwise seem premature, or at least jumping to conclusion a little too quickly. My comments about fairness was not directed to comments you made, but only to those who I got the feeling (just a feeling so I could be totally wrong) that they didn't base their comments on the 32 tones test, might not even have paid attention to it.

 

[RichB]

I absolutely agree with you, the 32 tones test results is prohibitively bad, but it helps that at least you included a condition, not a blanket condemnation that would otherwise seem premature, or at least jumping to conclusion a little too quickly. My comments about fairness was not directed to comments you made, but only to those who I got the feeling (just a feeling so I could be totally wrong) that they didn't base their comments on the 32 tones test, might not even have paid attention to it.

A fair criticism of unfair criticism

It would be wonderful if the multi-tone issue could be addressed with firmware.
That happened with Emotiva.

- Rich

 

[lashto]

Nobody is idiot...

Are you sure sure? For example I was really close to push the buy button on x4700 before the review came up. Really close to being a (modern) idiot won't you say?

Also the ancient greeks will 100% disagree with you because they had tens of thousands of idiots. Most people will probably be surprised to find out that the original greek word Idiot meant ~selfish. A very democratic word that one.

 

[davidc]

I think that could be a myth, depending what compares to what. For example, I have owned the AVR-3805, 4308, 1912, 1710, X3500H and now the X4400H, also had the Marantz AV7005, AV8801. So for me "all those in the past" would include 6 of the models I mentioned and the "present" is the X4400H. In that case, I would say my "present" AVR perform not only as good, but actually better than those in the past based on available bench test measurements done by AH, S&V, HTHF, HCC, AVtech and of course ASR. That is, not based on poor ears/brains, but specs and measurements.

Again, not trying to contradict your point at all, just saying it depends..

LOL. I'm old I guess. Past to me means 1980's 1990's. Not just AVR times!

 

[peng]

So, this also means that this AVR cannot fully resolve hi-res audio streaming such as TOTAL or Amazon Music hi-res?

Not relevant actually, because the PCM5100A chips were for the zone 2/3 and network. For the other digital inputs, it is still a 32 bit AKM chip, mostly likely the same AK4458 used in the 2016 through 2018 models such as the X3600H.