This is a review and detailed measurements of the Arcam AVR10 7.2 channel Audio/Video Receiver (AVR). It was purchased by a member and kindly drop shipped to me. AVR10 costs US $2,500.
Notes: our company Madrona Digital is a dealer for Harman which owns Arcam so feel free to read any level of bias in my review. As a courtesy, I sent a copy of my measurements to Harman/Arcam engineering a week ago. Alas, I have received no feedback. Meanwhile I had to return the unit to its owner.
The AVR10 has a solid and attractive look like rest of the line:
A large, informative display lets you easily navigate the user interface. Sadly the same is not replicated on the display output of the unit which is strange. Usually the AV receiver is less accessible in a home theater application so it is highly desirable to be able to navigate its menu on TV/Projector.
The back panel is as you expect with only modern inputs and outputs:
As you see there is a fan in the back. Inside there are actually two fans but I don't recall either one of them turning on even under load.
DAC Audio Measurements
As usual we start by feeding the AV receiver a 1 kHz tone (digitally over HDMI) to see how well it converts that to analog (Front Right and Front Left pre-out):
I had set the volume control to output 2 volts per CD standard and what every 2-channel DAC provides. Unfortunately at this level we have quite high distortion. This is typical of most of AVRs which allow their amp to clip and drag the performance of DAC with it. It makes no sense to do this as in this scenario the internal amplifier is not used (user would be using an external amp). Power is wasted and performance is reduced for no good reason or benefit.
In addition to clipping, we see elevated low frequency noise which seems to have a strange correlated signature (multiples of 10 Hz). An internal counter of sorts is causing noise to bleed into the output of the DAC. Needless to say, the AVR10 doesn't win any awards for distortion and noise:
To see where performance drops, I ran a sweep, varying the digital input and measuring output SINAD:
This test is volume control sensitive so I tested it at a couple of settings (made no difference). Basically best performance is achieved at just 0.9 volts. Unless your amplifier has unusually high gain, power will be limited in external amplification at 0.9 volt. You would need to turn up the volume and with it, lose the fidelity of pre-amp outputs as you see in the sharp drop above 0.9 volt. As a way of random example, here is the spec for Emotiva XPA-3 amplifier:
Intermodulation+noise versus input level nicely shows this clipping behavior:
32-tone signal emulating "music" shows the same clipping behavior we have seen in other Arcam products:
I building digital signal processing chain it is very important to check to make sure that there is no overflow condition for full amplitude signal. As some of you know, due to "loudness wars," a lot of music is mastered with peaks near or at 0 dBFS. It is important that then audio chain doesn't add its own massive distortion on top of an already bad situation in creation of the music. I have run this test on literally hundred of products and only Arcam AV products create this level of distortion. A firmware fix should remedy this if the will is there to fix it.
Linearity test shows that noise takes over and randomizes the output despite the heavy filtering that linearity test already includes:
If this were a $99 DAC, I would complain bitterly but AV products as a rule do so poorly on this test that I take this for "good enough" measurement result.
Test of jitter, noise and spurious signals shows fair bit of problems as we could have guessed from the noisy output in our dashboard earlier:
Nothing other than a single peak at 12 kHz should be there. We know this is interference as running the test twice created slightly different output (blue vs red). The wide "skirt" around the 12 kHz tone indicates random but low frequency noise in the DAC clock. Mass of spikes at 18 kHz can easily be traced to some activity inside the AVR at that frequency (networking, video, DSP, etc.). People are not running in streets fortunately because levels are too low to be very audible. Bad engineering "fixed" by the poor discrimination of our hearing.
A sweep of THD+N versus frequency using a much wider bandwidth than our dashboard shows quite high levels:
To see what is going on, I usually follow this test with a spectrum analysis of one sample frequency point at 10 kHz:
A perfect system would produce just one spike at 10 kHz and nothing else. Instead, we have all kinds of unwanted signals here. Yes, most are ultrasonic so "not audible" but show poor engineering and design hygiene. Part of the problem is the reconstruction filter in the DAC:
The filter setting Arcam has selected has poor out of band rejection. This means that mirror images of our signal relative to sample rate wind up showing in ultrasonic range. Arcam should allow the user to select the right filter already available in the DAC chip they use by the user. Or else, pick a more sensible one than this. Should take a couple of hours of development time to provide this capability to the user (setting a register in the DAC chip).
Now that we have sufficiently depressed ourselves, let's see how the amplifier section performs.
AVR Amplifier Measurements
As usual, we start with our 1 kHz Dashboard, producing 5 watts into 4 ohm load:
For this test I am using analog input in special "CD direct" mode. Large amount of noise is taking the performance way down. Harmonic distortion is actually low but noise is the problem for low SINAD of just 68 dB. Switching to HDMI input improves performance good bit:
The rankings for each input relative to other AVRs looks like this:
This is quite a bit worse than Arcam's own AVR390. In the larger context of all amplifiers tested to date (91 amps), performance is average with HDMI and far worse with analog input:
Poor noise performance bits us on the behind in signal to noise ratio at both 5 watts and full power:
These are embarrassingly poor numbers for analog input.
Crosstalk is decent:
CD direct input provides wide bandwidth indicating it is not digitized:
Once processed though, sample rate appears to be just 44.1 kHz with some odd ball antialiasing filter causing that strange response prior to 20 kHz. Really, in this day and age we better get full transparency for CD content. Why not sample higher so in-band response is dead flat?
Anyway, let's measure power vs distortion using the better digital input:
Spec is conservative which is good. This is not a lot of power though for a home theater application. Which may prompt you to use external amplification only to run into the distortion problem with the pre-out!
Streaming Audio Measurements
Streaming from Roon player is nicely supported allowing me to measure both our dashboard performance and jitter:
Performance is the same as local playback as it should be. Jitter however, gets a bit worse:
Proof positive that we are dealing with internal interference which varies with what the AVR10 is doing.
Conclusions
After seeing rather poor performance form many AV products, many of us were holding hope that Arcam would do better. Company has a great heritage of caring about measurements and superb performance. Alas, whatever pedigree was there, is no longer found. Maybe they and us are victim of one million logos on these devices, forcing the project to be mostly software/firmware development and hardware execution has become an afterthought. Regardless, for a multi-thousand dollar device, our expectation should be very well engineered AVR but we don't remotely get that.
In some sense, the AVR10 is worse than the last generation AVR390 which had great amplification stage. Seems like Arcam took the only thing that was good there and messed that up too.
Needless to say, I cannot recommend Arcam AVR10.
-----------
As always, questions, comments, recommendations, etc. are welcome.
Given the shortage of masks, I am thinking of planting an N95 mask tree. Someone is selling the transplants online. I understand all you have to do is water the thing and fertilize it and you get a bunch of certified masks!!! Help me get there by funding the seedlings by using : https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
Notes: our company Madrona Digital is a dealer for Harman which owns Arcam so feel free to read any level of bias in my review. As a courtesy, I sent a copy of my measurements to Harman/Arcam engineering a week ago. Alas, I have received no feedback. Meanwhile I had to return the unit to its owner.
The AVR10 has a solid and attractive look like rest of the line:
A large, informative display lets you easily navigate the user interface. Sadly the same is not replicated on the display output of the unit which is strange. Usually the AV receiver is less accessible in a home theater application so it is highly desirable to be able to navigate its menu on TV/Projector.
The back panel is as you expect with only modern inputs and outputs:
As you see there is a fan in the back. Inside there are actually two fans but I don't recall either one of them turning on even under load.
DAC Audio Measurements
As usual we start by feeding the AV receiver a 1 kHz tone (digitally over HDMI) to see how well it converts that to analog (Front Right and Front Left pre-out):
I had set the volume control to output 2 volts per CD standard and what every 2-channel DAC provides. Unfortunately at this level we have quite high distortion. This is typical of most of AVRs which allow their amp to clip and drag the performance of DAC with it. It makes no sense to do this as in this scenario the internal amplifier is not used (user would be using an external amp). Power is wasted and performance is reduced for no good reason or benefit.
In addition to clipping, we see elevated low frequency noise which seems to have a strange correlated signature (multiples of 10 Hz). An internal counter of sorts is causing noise to bleed into the output of the DAC. Needless to say, the AVR10 doesn't win any awards for distortion and noise:
To see where performance drops, I ran a sweep, varying the digital input and measuring output SINAD:
This test is volume control sensitive so I tested it at a couple of settings (made no difference). Basically best performance is achieved at just 0.9 volts. Unless your amplifier has unusually high gain, power will be limited in external amplification at 0.9 volt. You would need to turn up the volume and with it, lose the fidelity of pre-amp outputs as you see in the sharp drop above 0.9 volt. As a way of random example, here is the spec for Emotiva XPA-3 amplifier:
Intermodulation+noise versus input level nicely shows this clipping behavior:
32-tone signal emulating "music" shows the same clipping behavior we have seen in other Arcam products:
I building digital signal processing chain it is very important to check to make sure that there is no overflow condition for full amplitude signal. As some of you know, due to "loudness wars," a lot of music is mastered with peaks near or at 0 dBFS. It is important that then audio chain doesn't add its own massive distortion on top of an already bad situation in creation of the music. I have run this test on literally hundred of products and only Arcam AV products create this level of distortion. A firmware fix should remedy this if the will is there to fix it.
Linearity test shows that noise takes over and randomizes the output despite the heavy filtering that linearity test already includes:
If this were a $99 DAC, I would complain bitterly but AV products as a rule do so poorly on this test that I take this for "good enough" measurement result.
Test of jitter, noise and spurious signals shows fair bit of problems as we could have guessed from the noisy output in our dashboard earlier:
Nothing other than a single peak at 12 kHz should be there. We know this is interference as running the test twice created slightly different output (blue vs red). The wide "skirt" around the 12 kHz tone indicates random but low frequency noise in the DAC clock. Mass of spikes at 18 kHz can easily be traced to some activity inside the AVR at that frequency (networking, video, DSP, etc.). People are not running in streets fortunately because levels are too low to be very audible. Bad engineering "fixed" by the poor discrimination of our hearing.
A sweep of THD+N versus frequency using a much wider bandwidth than our dashboard shows quite high levels:
To see what is going on, I usually follow this test with a spectrum analysis of one sample frequency point at 10 kHz:
A perfect system would produce just one spike at 10 kHz and nothing else. Instead, we have all kinds of unwanted signals here. Yes, most are ultrasonic so "not audible" but show poor engineering and design hygiene. Part of the problem is the reconstruction filter in the DAC:
The filter setting Arcam has selected has poor out of band rejection. This means that mirror images of our signal relative to sample rate wind up showing in ultrasonic range. Arcam should allow the user to select the right filter already available in the DAC chip they use by the user. Or else, pick a more sensible one than this. Should take a couple of hours of development time to provide this capability to the user (setting a register in the DAC chip).
Now that we have sufficiently depressed ourselves, let's see how the amplifier section performs.
AVR Amplifier Measurements
As usual, we start with our 1 kHz Dashboard, producing 5 watts into 4 ohm load:
For this test I am using analog input in special "CD direct" mode. Large amount of noise is taking the performance way down. Harmonic distortion is actually low but noise is the problem for low SINAD of just 68 dB. Switching to HDMI input improves performance good bit:
The rankings for each input relative to other AVRs looks like this:
This is quite a bit worse than Arcam's own AVR390. In the larger context of all amplifiers tested to date (91 amps), performance is average with HDMI and far worse with analog input:
Poor noise performance bits us on the behind in signal to noise ratio at both 5 watts and full power:
These are embarrassingly poor numbers for analog input.
Crosstalk is decent:
CD direct input provides wide bandwidth indicating it is not digitized:
Once processed though, sample rate appears to be just 44.1 kHz with some odd ball antialiasing filter causing that strange response prior to 20 kHz. Really, in this day and age we better get full transparency for CD content. Why not sample higher so in-band response is dead flat?
Anyway, let's measure power vs distortion using the better digital input:
Spec is conservative which is good. This is not a lot of power though for a home theater application. Which may prompt you to use external amplification only to run into the distortion problem with the pre-out!
Streaming Audio Measurements
Streaming from Roon player is nicely supported allowing me to measure both our dashboard performance and jitter:
Performance is the same as local playback as it should be. Jitter however, gets a bit worse:
Proof positive that we are dealing with internal interference which varies with what the AVR10 is doing.
Conclusions
After seeing rather poor performance form many AV products, many of us were holding hope that Arcam would do better. Company has a great heritage of caring about measurements and superb performance. Alas, whatever pedigree was there, is no longer found. Maybe they and us are victim of one million logos on these devices, forcing the project to be mostly software/firmware development and hardware execution has become an afterthought. Regardless, for a multi-thousand dollar device, our expectation should be very well engineered AVR but we don't remotely get that.
In some sense, the AVR10 is worse than the last generation AVR390 which had great amplification stage. Seems like Arcam took the only thing that was good there and messed that up too.
Needless to say, I cannot recommend Arcam AVR10.
-----------
As always, questions, comments, recommendations, etc. are welcome.
Given the shortage of masks, I am thinking of planting an N95 mask tree. Someone is selling the transplants online. I understand all you have to do is water the thing and fertilize it and you get a bunch of certified masks!!! Help me get there by funding the seedlings by using : https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
