Arcam AVR850 Home Theater AVR Review

[fordiebianco]

Why is top quality expected here? These are mass consumer products. They will never achieve high quality. It is produced quickly and has to be cheap.

I personally don't think that at its initial $6000 price tag this is a mass consumer product. It might be engineered that way, but for that price I would at least think I buy an upper mid-fi, audiophile product.

In my mind a mass consumer product would be something like the Yamaha HTR-2071 for 199 pounds.

Thanks to this website we all know that there is probably not a big difference between the two.

 

[Xyrium]

At least the side channels performed admirably....

 

[TimoJ]

At the same time, small setting is very common by far compared to large so this performance is closer to actual use case.

With the same logic then you should also publish SINAD results of the devices that have better performance with balanced outputs around 2 volts than 4 volts. Power amplifier balanced input sensitivity around 2 volts is closer to actual use case.

 

[SimpleTheater]

If ANY AVR has Dirac shouldn’t it get the highest honors possible? </sarcasm>

 

[amirm]

With the same logic then you should also publish SINAD results of the devices that have better performance with balanced outputs around 2 volts than 4 volts. Power amplifier balanced input sensitivity around 2 volts is closer to actual use case.

It is not the same logic. Same logic would be to set them to small and see if their performance improves.

What you are rehashing will encourage companies to continue to keep the broken architectural of driving internal amps when the user is using external amp. You need to find a reviewer who doesn't know any better to run with that argument.

 

[Tks]

Nice review bossman!

I just had a question, what exactly would digital overflow be?

Also that THD Ratio graph was funny, 2.5% lol

 

[infinitesymphony]

It’s 400$ per channel, cheaper than the Minidsp SHD. And includes amps, so not so expensive as it looks.

Setting aside the fact that the MiniDSP SHD is not a home theater receiver and contains features that this device does not have, this Arcam is a 7.2-channel receiver for $6,000 -- wouldn't that equal $667/channel? And if we're judging equipment by the $/channel metric, the Pioneer VSX-LX504 has almost the same performance for around $90/channel (9.2 channel, $999 MSRP).

 

[Dimifoot]

this Arcam is a 7.2-channel receiver

No, it’s 7.1.4.
11 channels+sub. Amps for 7, preouts for 11 (+sub)

 

[crypticc]

Hey Amirm

Thanks for the review. I have avr850 and logging in to say hi but also a question...
Needless to say that you turned Dirac off when measuring, but was the amp digital out set to 96kHz? I'd read elsewhere that unless set to CD direct (the PCM digital options in the menu and not to be confused with the Stereo direct button when audio source is analogue) and also the digital output set to 96kHz, then amp resamples internally to whatever is the digital out frequency, e.g. 48kH. Also does that, even for analogue inputs unless the Stereo Direct option is picked. Could this explain your observations with poor digital resolution?


Thank you

 

[Thomas Savage]

@amirm could you put a ' Hi Rez ' sticker on this and retest ? Until this happens I view these measurements as incomplete

 

[Ron Texas]

Due to their complexity, I find tests of AV receivers to be far more interesting than tests of DAC's and Amps. Too bad this fairly expensive unit did not shine.

 

[North_Sky]

No, it’s 7.1.4.
11 channels+sub. Amps for 7, preouts for 11 (+sub)

I think he meant it has 7 channels of amplification and 2 subwoofer outputs.
To become a full 7.2.4 Dolby Atmos system it needs an extra 4 more additional external channels of amplification. And 2 subwoofers of course.

For $6,000 I'd like to see 11 channels of internal amplification inside the unit.
Looking inside there is enough space for them. And as there are already three fans to cool them off...or say two.

It is what it is

 

[RickSanchez]

Meanwhile at some audio manufacturer's HQ:
[development team]: Hey boss, we've just finished our last hi-fi design, it's ready for production!
[boss]: Calm down your enthusiasm boys. Send a prototype to Amir, he'll tell us if we are finished yet.

I'll admit that my experience is in software, not hardware. But I have the funny feeling you have the conversation reversed:

[development team]: "Hey boss, we've just finished our last hi-fi design; let's send a prototype to ASR for measurement."
[boss]: "Calm down your enthusiasm boys. Let's start shipping the unit, Marketing will clean up the 'rough edges'."

 

[amirm]

I just had a question, what exactly would digital overflow be?

Digital samples have a limit which you cannot exceed. If you perform math on those samples in the process of resampling, filtering, etc. you may wind up with math that results over max. Care must be taken to not allow this to happen.

Let's say you only have whole numbers to work with. I give you 9 and say to divide by 2. You get 4.5 but can't have decimal places so you round up to 5. Now I ask you to double that and you get 10, not 9! A proper system has higher internal resolution and uses proper dither to convert back and forth as to not create errors like this. In this example, your internal representation could carry 4.5 so can multiply by 2 and get 9.

Since music is loudness compressed, it is very easy to get near max PCM samples. So the excuse from companies like Arcam that these test tones are too close to max doesn't hold water. System better handle them.

 

[Tks]

Digital samples have a limit which you cannot exceed. If you perform math on those samples in the process of resampling, filtering, etc. you may wind up with math that results over max. Care must be taken to not allow this to happen.

Let's say you only have whole numbers to work with. I give you 9 and say to divide by 2. You get 4.5 but can't have decimal places so you round up to 5. Now I ask you to double that and you get 10, not 9! A proper system has higher internal resolution and uses proper dither to convert back and forth as to not create errors like this. In this example, your internal representation could carry 4.5 so can multiply by 2 and get 9.

Since music is loudness compressed, it is very easy to get near max PCM samples. So the excuse from companies like Arcam that these test tones are too close to max doesn't hold water. System better handle them.

Thanks bossman. I was just wondering one last thing since you explained this to me in detail. I was wondering just hardware-wise, and computationally speaking, this seems like almost a trivial thing to account for in the modern day (meaning not expensive to implement and build into your system basically). Am I correct to say this? Or is this some crazy expensive thing where a company can be excused for not implementing it only because the companies with the biggest resources can do this?

To be it simply seems like a massive amateurish blunder of embarrassing proportions.

 

[GXAlan]

Thanks bossman. I was just wondering one last thing since you explained this to me in detail. I was wondering just hardware-wise, and computationally speaking, this seems like almost a trivial thing to account for in the modern day (meaning not expensive to implement and build into your system basically). Am I correct to say this? Or is this some crazy expensive thing where a company can be excused for not implementing it only because the companies with the biggest resources can do this?

To be it simply seems like a massive amateurish blunder of embarrassing proportions.

I think Yamaha went to “64-bit” numerical processing to address the mathematics part (but the Cx5200 has a low quality volume IC). There was an interview with Denon Japanese engineers which said that higher precision math was not feasible within the cost constraints of even the target X8500h price point.

Presumably, Trinnov with a full software stack avoids this mathematical precision problem.

 

[amirm]

Thanks bossman. I was just wondering one last thing since you explained this to me in detail. I was wondering just hardware-wise, and computationally speaking, this seems like almost a trivial thing to account for in the modern day (meaning not expensive to implement and build into your system basically). Am I correct to say this? Or is this some crazy expensive thing where a company can be excused for not implementing it only because the companies with the biggest resources can do this?

Well, when it comes to cost and laziness, there is no limit to what happens in embedded processors/DSP and programming.

Doubling the precision usually halves the computational speed. When you have 11 or more channels to process, this becomes a big deal.

FYI the AV40 processor I tested recently has two independent DSPs for above reason.

The issue can very well also be poor programming skills, and or lack of testing. The resolution may be there and not the limit.

 

[rccarguy]

Are there any issues from the speaker binding posts when speakers are set to large? Is there same issue between small and large when using internal amps, or just the pre outs?

 

[amirm]

Are there any issues from the speaker binding posts when speakers are set to large? Is there same issue between small and large when using internal amps, or just the pre outs?

It impacts both since the DAC feeds the amp. See the Amp section of the review.

 

[rccarguy]

Yeah I had a feeling it would talking to guy on av messaging as basically signal out from dac onwards is effected.

Do other arcam avr have this issue or just this model?