Denon AVR-X3800H Review

[GalZohar]

Read everything I could on the latest AVRs, AVCs and AVPs. Cosidered them all, Onkyo/Integra/Pioneer, NAD, Anthem, etc. Settled on Denon.

Was undecided on which way to go, 3800 or 4800. The 3800s were on sale not too long ago, so I grabbed one to try from Crutchfield for $1500. That model sold out the next day and mine arrived that day.

Swapped out the only other AVR I've ever owned, the Sherwood Newcastle R-972, then proceeded with setup. Got the Audyssey app, ran the DRC using 3 captures at the same mic location, as there is only one listening position. XO is 80Hz. Disabled the midrange bump, enabled "Dynamic EQ", selected "Reference" on the AVC. Running 2.1 in an apartment at low to medium volume.

This thing sounds really good! I'd like to have the extras offered by the 4800 but don't see a good reason to drop another $1k. Looking at an upcoming Martin Logan 5.0 purchase, the money saved will step me from the EM to the bigger X. I am enjoying this so much that I am hesitant to do anything further, fully-satisfied with stereo reproduction.

I imagine that someday I'll dig into Dirac (have that on my 1st gen XMC-1 and appreciated it) and maybe even explore MultEQ X.

I am still not 100% on keeping it or not, but so far, so good. I do wonder why these went on sale for a few days.

Limit EQ in the app to 500Hz (you can also try 300Hz), and don't measure the same point 3 times, but rather take 3-5 measurements symmetrically around the listening position to get some kind of spatial averaging, which will prevent over-corretion of issues that might only be present at a very specific location point in space.

 

[fer2503]

hi, on my denon x1200 in the display I see the type of incoming audio format (atmos, DD+, dts), in the new 3800h instead I only see the audio mode set. Is there a way to change what you see on the display? Is there a way to change what you see on the display?

 

[tesseractASR]

Limit EQ in the app to 500Hz (you can also try 300Hz), and don't measure the same point 3 times, but rather take 3-5 measurements symmetrically around the listening position to get some kind of spatial averaging, which will prevent over-corretion of issues that might only be present at a very specific location point in space.

I've seen (and listened) people have good luck with single-point Dirac. What you suggests makes good sense, though. I'll try it.

I've blind tested Dirac limited to 500Hz and full-range. I could hear a clear difference 100% of the time. Preference is another matter, that flip flopped.

 

[peng]

I've seen (and listened) people have good luck with single-point Dirac. What you suggests makes good sense, though. I'll try it.

I've blind tested Dirac limited to 500Hz and full-range. I could hear a clear difference 100% of the time. Preference is another matter, that flip flopped.

Since you can hear the difference 100% of the time, then it is better to try different ways so that you will know for sure which way is best for you. What others prefer may or may not apply to you. To me, I prefer full range with Dirac and limit to below 3,000 to 5,000 Hz with Audyssey, but that's just for my applications.

 

[tesseractASR]

Since you can hear the difference 100% of the time, then it is better to try different ways so that you will know for sure which way is best for you. What others prefer may or may not apply to you. To me, I prefer full range with Dirac and limit to below 3,000 to 5,000 Hz with Audyssey, but that's just for my applications.

I will try XT32 @ 5kHz and walk it down an octave at a time.

 

[tesseractASR]

I am seeing refurb/open box 3800s and 4800s. While I am not inclined to return my new 3800 for an open box, I might spend the extra $500 on a refurb 4800.

Seems like the 4800 performs better as a pre/pro, but the 3800 SINAD is fine as long as only the internal amps are being used?

 

[peng]

Seems like the 4800 performs better as a pre/pro, but the 3800 SINAD is fine as long as only the internal amps are being used?

It may seem like that on paper but if you don't need more than 1.6 V pre out voltage for your power amp, the 3800 can do 90 dB or higher SINAD vs the 4800's 92 dB (at 1.6 V) You are not likely going to hear a difference.

 

[dlaloum]

With my 1980's vintage Quad power amps, full rated power is at 0.5V (for the older models) and 0.7V (for the newer ones) ... so the peak SINAD point is totally in the wrong place.... (for these power amps)

I still feel that the standard line level unbalanced signal should never be more than 1V....

CD players started introducing variable volume levels on a source component, and some of them then started running "over-voltage" with outputs going up to 2V.... and people then started to use that as a direct feed into power amps.... this then led to lower gain power amps ....lower gain, means improved SINAD... so the specmanship battle naturally followed... and now we have whole generations of amps that required "high voltage" inputs to drive them to spec.

The actual formal specifications for pre-out on many of the AVR's in the market is 1V for unbalanced outputs.... everything beyond that is bonus.
The flaw that gets me annoyed, is that rather than having peak SINAD within that 0V to 1V range, most of the AVR's end up with peak SINAD out at 3V or more.... which even with low gain power amps, is almost never going to be used.

Where we need the SINAD is at the low end, when things are in the quiet parts of movies or the quiet parts of music.... once we reach a crescendo... we aren't ever going to be able to hear the base noise!

Specmanship has led to a well nigh useless spec.

 

[Alice of Old Vincennes]

With my 1980's vintage Quad power amps, full rated power is at 0.5V (for the older models) and 0.7V (for the newer ones) ... so the peak SINAD point is totally in the wrong place.... (for these power amps)

I still feel that the standard line level unbalanced signal should never be more than 1V....

CD players started introducing variable volume levels on a source component, and some of them then started running "over-voltage" with outputs going up to 2V.... and people then started to use that as a direct feed into power amps.... this then led to lower gain power amps ....lower gain, means improved SINAD... so the specmanship battle naturally followed... and now we have whole generations of amps that required "high voltage" inputs to drive them to spec.

The actual formal specifications for pre-out on many of the AVR's in the market is 1V for unbalanced outputs.... everything beyond that is bonus.
The flaw that gets me annoyed, is that rather than having peak SINAD within that 0V to 1V range, most of the AVR's end up with peak SINAD out at 3V or more.... which even with low gain power amps, is almost never going to be used.

Where we need the SINAD is at the low end, when things are in the quiet parts of movies or the quiet parts of music.... once we reach a crescendo... we aren't ever going to be able to hear the base noise!

Specmanship has led to a well nigh useless spec.

Anthem 740/1140 pre-out max out at 5 Vmrs.

 

[dlaloum]

Anthem 740/1140 pre-out max out at 5 Vmrs.

Yes, and when listening to them at normal levels (ie: circa 70 to 75db at MLP) - what would the requisite V out level be? - 0.2V? what is the SINAD at that point?

Both the Denon's and the Onkyo's can drive 3.5V to 4V output - but the really important thing is not what their max output V is, but rather what is their SINAD, at the low V output levels, at which levels any noise/distortion is most audible... and typically, those are not the points at which the SINAD is at its lowest.... sadly.

 

[HarmonicTHD]

Yes, and when listening to them at normal levels (ie: circa 70 to 75db at MLP) - what would the requisite V out level be? - 0.2V? what is the SINAD at that point?

Both the Denon's and the Onkyo's can drive 3.5V to 4V output - but the really important thing is not what their max output V is, but rather what is their SINAD, at the low V output levels, at which levels any noise/distortion is most audible... and typically, those are not the points at which the SINAD is at its lowest.... sadly.

Running the math gives ca. 0.95W into 8ohm from 0.2VRMS preout assuming poweramps with 26dB gain. For most speaker that would be louder than ca 75dBSPL. The SINAD from the Denon at 0.2V can not be read off the chart Amir published, so unless there is more data somewhere else - nobody really knows for sure.

If you want to be anal about it, get poweramps with low gain and preouts max voltage eg 4V and 20dB gain to optimize noise and distortion reaching the speakers.

 

[Mr Wolf]

If you want to be anal about it, get poweramps with low gain and preouts max voltage eg 4V and 20dB gain to optimize noise and distortion reaching the speakers.

Quite right but the point he's making is not an anal one at all. He's simply highlighting the fact that most of the headline SINAD measurements bandied about on this forum are, for all practical purposes, totally useless. And for the great many people that appear to not appreciate this then they're worse than that, they're misleading.

The reality is that most people's pre-out/speaker voltages are far lower than they think and the only SINAD levels we should actually care about are the ones at relatively tiny voltages.

Using my 2400Ft3 HT room as an example where the MLP is 10-11ft from 89-90dB sensitive LCR speakers, the most demanding speaker power wise measures as requiring 25W at 8 Ohms to hit maximum 105dB reference peaks. That's 14.1V output to the speaker on a 29dB (28.2x) gain AVR so a theoretical maximum peak of 0.5V at the pre-outs. Like most people I listen no higher than -10dB MV so in my case that maximum possible peak pre-out voltage falls to 0.158V. And this is on the loudest 0dBFS transient peaks that last a fraction of a second which will be crash, bang, wallop sound effects so not exactly what you'd call THD+N% sensitive critical listening content, that will require much lower voltages. So let's look at loud music score passages which are typically mixed closer to -20dBFS (i.e. 85dB at reference). On my system at -10dB MV the peak pre-out voltage on those would be a tiny 0.0158V and even at reference level that only increases to 0.05V. And for those systems that might require say 4x the amplifier power (larger rooms, greater distances, less sensitive speakers etc.) that's still maximum pre-out voltages of only 0.1V at reference and 0.03V at -10dB MV on these loud music passages.

The positive is that the differences between SINAD levels at these lower voltage output levels might actually be audible and worth caring about unlike the 80dB+ figures that people get unduly excited about at the moment.

 

[HarmonicTHD]

Quite right but the point he's making is not an anal one at all. He's simply highlighting the fact that most of the headline SINAD measurements bandied about on this forum are, for all practical purposes, totally useless. And for the great many people that appear to not appreciate this then they're worse than that, they're misleading.

The reality is that most people's pre-out/speaker voltages are far lower than they think and the only SINAD levels we should actually care about are the ones at relatively tiny voltages.

Using my 2400Ft3 HT room as an example where the MLP is 10-11ft from 89-90dB sensitive LCR speakers, the most demanding speaker power wise measures as requiring 25W at 8 Ohms to hit maximum 105dB reference peaks. That's 14.1V output to the speaker on a 29dB (28.2x) gain AVR so a theoretical maximum peak of 0.5V at the pre-outs. Like most people I listen no higher than -10dB MV so in my case that maximum possible peak pre-out voltage falls to 0.158V. And this is on the loudest 0dBFS transient peaks that last a fraction of a second which will be crash, bang, wallop sound effects so not exactly what you'd call THD+N% sensitive critical listening content, that will require much lower voltages. So let's look at loud music score passages which are typically mixed closer to -20dBFS (i.e. 85dB at reference). On my system at -10dB MV the peak pre-out voltage on those would be a tiny 0.0158V and even at reference level that only increases to 0.05V. And for those systems that might require say 4x the amplifier power (larger rooms, greater distances, less sensitive speakers etc.) that's still maximum pre-out voltages of only 0.1V at reference and 0.03V at -10dB MV on these loud music passages.

The positive is that the differences between SINAD levels at these lower voltage output levels might actually be audible and worth caring about unlike the 80dB+ figures that people get unduly excited about at the moment.

Correct. Yes I would also appreciate SINAD tests which go down further in voltage.

I btw run 20dB gain in my poweramps ;-)

As for audibility at lower levels, I wouldn’t be so certain also because of Fletcher-Munson, but of course one would have to try on an individual base. You know about the Distort test where you can test your own sensitivity to distortion. It is humbling for most people.

Distort audibility tester | Distort documentation

distortaudio.org

As for noise (depending on the volume control) it decreases in absolute value with output voltage and therefore is often not an issue (in SOTA amps).

 

[Mr Wolf]

As for audibility at lower levels, I wouldn’t be so certain also because of Fletcher-Munson, but of course one would have to try on an individual base. You know about the Distort test where you can test your own sensitivity to distortion.

Oh I'm far from certain which is why I put the word "might" in there but it would be much closer to being audible for sure, unlike the nonsensical measurements that's causing some people to return their new AVRs for one with "better SINAD". That's just plain silly if not sad.

 

[peng]

Where we need the SINAD is at the low end, when things are in the quiet parts of movies or the quiet parts of music.... once we reach a crescendo... we aren't ever going to be able to hear the base noise!

You are again, appearing to focus on some very specific user cases, just as your own!! Ideally, I would agree with you that we want high SINAD across the audible range. Since things are not ideal in the mid to upper mid price range gear, tradeoffs bound to happen. In that case, the SINAD vs measured output levels graph that the mid range gear such as Onkyo, Denon, Marantz, Anthem's AVRs/AVPs seem reasonably good.

Take a good look of the Onkyo RZ50 (I picked it because it is a current model, same with the Denon 3800):
Yes, SINADS is lower at the lower output level, in this case it might have been 500 mV ore lower, we don't know for sure but let's assume it is 500 mV, so it is about 88 dB Vs 96 dB at 2 V. That sounds like a lot, but consider the following:

1) 88 dB SINAD, or 0.004% THD+N is pretty good and I am 100% sure other reviewer such as Gene or JA would consider that level low enough to be considered below the threshold of audibility. Even Amir would probably agree though he might say that depends on if it is distortions dominated.

2) Even if -88 dB THD+N is audible to enough people in their HT rooms, because they have super quite rooms, unless they have speaker with very high sensitivity, how likely are going to be bothered by the difference between -88 dB and -96 dB when listening at spl well below reference leve?

Example: 88 dB/2.83V/m speakers will produce 78 dB at 10 ft, -88 dB THD+N is -10 dB below the fundamental, or -30 dB below a noise floor of 20 dB average and that's even if that -88 dB is all noise. How likely would most HT users be bothered by that level of noise related to the noise in a quiet room? On top of that, in real world use, there's also masking effects. Keep in mind many users find the likes of the NAD T758V3 or Anthem MRX that offers SINAD below 70 dB at 2 V, presumably much lower at below 500 mV.

That's the Onkyo, now back to the Denon 3800:
Here you see >90 dB at below 300 mV

So, even when you listen to classical music, or movies that has a lot of "quiet parts" that you refer to, it really won't be as serious as you seem to be suggesting.
To be clear, I agree with you and others who pointed out the fact that we do need SINAD to be high at the low output levels, but let's not help perpetuating the myth that it is a common issue, when it isn't. I am not sure if you are just trying to make a point by exaggerating the potential issue somewhat, but it may end up misleading others to think it is an issue for them when it may not be at all.

Again, I agree with you in principle, but it would be great if we all get less extreme in making our points on matters that are not always black and white.

 

[tesseractASR]

CD players started introducing variable volume levels on a source component, and some of them then started running "over-voltage" with outputs going up to 2V.... and people then started to use that as a direct feed into power amps....

Three decades ago, I put together a minimalist system for a friend using a Sony w/variable out, a Hafler amp w/low input sensitivity and a pair of Jensen-era Acoustic Research M1s. The same speakers Levinson used to show off his new Cello line, heh.

 

[tesseractASR]

I am very much enjoying the X3800H, aside from the overzealous level control. But the thought of possible issues using the internal amps for immersion and external amps for the bed layer prompted me to order a refurb X4800H for only $500 more. The selective channel pair amplifier disconnect, superior jitter spec, backlit remote and Japanese-build quality seem worth it, to me. So soon, I'll have an X3800H and X4800H here, with one of those being returned.

I read earlier that the circuit board amps tend to perform better than the monoliths, but both types seem good enough and eventually, the AVC will be feeding another mains amp.

 

[dlaloum]

You are again, appearing to focus on some very specific user cases, just as your own!! Ideally, I would agree with you that we want high SINAD across the audible range. Since things are not ideal in the mid to upper mid price range gear, tradeoffs bound to happen. In that case, the SINAD vs measured output levels graph that the mid range gear such as Onkyo, Denon, Marantz, Anthem's AVRs/AVPs seem reasonably good.

Take a good look of the Onkyo RZ50 (I picked it because it is a current model, same with the Denon 3800):
Yes, SINADS is lower at the lower output level, in this case it might have been 500 mV ore lower, we don't know for sure but let's assume it is 500 mV, so it is about 88 dB Vs 96 dB at 2 V. That sounds like a lot, but consider the following:

1) 88 dB SINAD, or 0.004% THD+N is pretty good and I am 100% sure other reviewer such as Gene or JA would consider that level low enough to be considered below the threshold of audibility. Even Amir would probably agree though he might say that depends on if it is distortions dominated.

2) Even if -88 dB THD+N is audible to enough people in their HT rooms, because they have super quite rooms, unless they have speaker with very high sensitivity, how likely are going to be bothered by the difference between -88 dB and -96 dB when listening at spl well below reference leve?

Example: 88 dB/2.83V/m speakers will produce 78 dB at 10 ft, -88 dB THD+N is -10 dB below the fundamental, or -30 dB below a noise floor of 20 dB average and that's even if that -88 dB is all noise. How likely would most HT users be bothered by that level of noise related to the noise in a quiet room? On top of that, in real world use, there's also masking effects. Keep in mind many users find the likes of the NAD T758V3 or Anthem MRX that offers SINAD below 70 dB at 2 V, presumably much lower at below 500 mV.

That's the Onkyo, now back to the Denon 3800:
Here you see >90 dB at below 300 mV

So, even when you listen to classical music, or movies that has a lot of "quiet parts" that you refer to, it really won't be as serious as you seem to be suggesting.
To be clear, I agree with you and others who pointed out the fact that we do need SINAD to be high at the low output levels, but let's not help perpetuating the myth that it is a common issue, when it isn't. I am not sure if you are just trying to make a point by exaggerating the potential issue somewhat, but it may end up misleading others to think it is an issue for them when it may not be at all.

Again, I agree with you in principle, but it would be great if we all get less extreme in making our points on matters that are not always black and white.

We mostly agree quite vigorously....

And what is a discussion, without a bit of drama... (much gesticulation, raised voices, a bit of emotion!)

 

[Acerun]

I don't use my Denon 4700 for anything but movies. For stereo, I use the RME ADI-2 FS connected to a MBP running Audirvana. MBP (Audirvana) -> ADI-2 FS -> Purifi amp -> Revel F328Be. This combination offers a clarity of sound with music that my 4700 can't match. In answer to your question, I would suggest turning the AVR off for stereo listening.

As far as ATMOS movies I have not had any problems with the Denon 4700 performance. I wouldn't recommend or buy a Denon 3800. It's not the value proposition I look for and adding an external DAC to it doesn't improve the sound as much as simply removing the AVR from the chain.

Sorry for commenting on a thread from some time ago but I've considered doing something like this but you would lose Audyssey and subwoofer integration right?

 

[OpticonPrime]

I haven't read the whole thread but despite that the unit doesn't measure great in comparison with the introduction price (you should get more for that money). The price is much lower now.
Payed €700 for a X2700h (no hdmi bug model) a year ago and doubting to upgrade to a x3800h for better Audyssey and the possibility of Dirac.
The price of the X3800h dropped from €1699 to €950.
So it sounds like a good deal for me where I use it just in my living room.
What do you guys think of the price in comparison with the performance of the x3800h.