This is a review and detailed measurements of the Arcam AV40 AV Processor. It was purchased by a member as "z-stock" and kindly loaned to me for testing. He said that unit arrived brand new with even the plastics undisturbed. I saw nothing in my testing that would indicate it being broken. He had it upgraded to the latest version of the firmware (1.28) before giving it to me.
Note: our company Madrona Digital is a dealer for Harman company which acquired Arcam a few years ago. So feel free to read as much bias as you like into this review.
The AV40 managed to differentiate itself in a see of similar looking products with a unique configuration and attractive design:
The display is large with good resolution and excellent contrast. More so, it lets you fully navigate the unit just like you are used to seeing on the display which is great for manipulating the unit when standing in front of it. Oddly the on-screen display overlaid on your TV/projector is only two lines! This is reverse of what most companies do, skimping on front panel and giving you a full output on the display (although some give you the same two line display on both). I would have wanted to see both mirroring each other.
The controls feel very nice but something is "seriously" wrong with them if you are among world's 90% of the population. Can you see it? Why is the volume control on the left of the display and not right? If you are right handed you have to reach to the left to changing the volume and thereby blocking the display. I would put the power button on the left and volume on the right. OK, this is a minor thing but it stood out for me.
The back panel has the more modern take on AV products, leaving legacy video connectors behind:
Even though there has been serious growing pains as far as software with this product, in my testing where I only touch very little of the product, all worked solidly.
My overall impression of AV40 look and feel is very positive and well above the score I give to many other AV products.
HDMI DAC Audio Measurements
We start with our foundation measurement which is a simple 1 kHz tone, generated digitally as we measure the Front Left and Right balanced outputs. Unless noted, that is the output I am testing. And mode of operation is "Stereo Direct" although I found that it made no difference in measured performance in a couple of places I checked:
The DAC chip that is uses is ESS9026Pro which ESS rates at a SINAD of 110 dB. We are sadly way below that at just 90 dB or so. While SINAD is deteremined by both noise and distortion, here it is the second harmonic that dominates the measurements at -92 dB. This SINAD rating just rescues the AV40 from "broken" category (in red) but by a hair:
Certainly not competitive with anything we are used to even in budget 2-channel audio products:
You can't see it in a static display but as I note on the graph, the left side had a pattern that would come and go. It is interference from the rest of the subsystem bleeding into the DAC's analog output. It becomes a spray of noise as we move up in frequencies to the right of our 1 kHz tone.
Often AV manufacturers put higher quality DACs for left and right channels so I thought I measure a couple of other ones to see how they do:
What the heck??? Performance improves? Notice how the second harmonic distortion has gone way down to around -105 dB from 92. If noise was not holding it back, SINAD would improve a lot more.
You may be asking if the performance improves at other output levels so let's see if that is the case:
As is typical of these AV products, where you set the volume determines different gain stages that get turned on and off, impacting performance. Using the same volume position as previous tests, we see that best performance is at 3.1 volts than 4 volts. Improvement is negligible though to the tune of just 1 dB.
Using higher volume control position gives you more output but with worse performance. So if your amplifier has multiple gain settings, don't go for much over 4 volt "sensitivity" (i.e. level that produces max power).
Back to noise, here is our dynamic range with 4 volts output on balanced out:
Oh, come on now. We can't even clear 16 bit dynamic range in this and age?
The high noise floor seriously degrades intermodulation distortion+noise measurement:
I am not going to raise my blood pressure more by repeating the words on the graph.
I watched a marketing video for an Arcam product and the presenter talked at length how jitter makes music sound bad and that Arcam gets it right. Does it? Let's measure with HDMI first:
Clearly not. The J-test signal has a embedded 250 Hz cycle that toggles all the bits in the sample. It does so by just changing the low order bit in a 24-bit audio sample so should be invisible in our measurement above. Sadly it is anything but invisible. What this means is that what digital data arrives at the input of the DAC, determines what bleeds into the output as unwanted/spurious tones. Classic lack of isolation between digital and analog sections. There are other jitter sources that cause the "skirt" around our main tone and some other spikes close to it.
Toslink is often used for return audio from a TV so let's see how we do with that optical input:
Jitter galore! Clearly no attention was paid to reduction of jitter in this AV40.
Fortunately these levels while visually distracting, audibly are too low to be heard easily so don't panic over sound getting "gritty and such."
Any more sins? Yeh, sure. Here is our linearity graph:
Terrible results. The zigzag graph that amplifies as levels go down tends to indicate truncation of 24-bit samples without dither. There should be no truncation but if there is, for heaven's sake apply some dither to it so it linearizes the levels.
Here is our DAC reconstruction filter:
I don't mind the curve but why is the attenuation just 60 dB? Theory mandates infinite attenuation but in practice, I take 90 dB or more. But not 60 dB. This will impact tests like THD+N versus frequency which use wide band measurements that include components created due to above weak filter:
The blue line is a $99 (or less) DAC board. Let's pick a 10 kHz and see what contributes to that sum on the graph:
Wow, we have a nice buffet for dinner. Jitter, noise, out of band image components, random junk, harmonic distortion, etc. Be thankful your ears aren't nearly as good as my analyzer!
Can it get worse? Turns out it can. Let's look at intermodulation distortion this time with 32 tones resembling "music:"
This is so, so bad I thought there was something wrong with the measurement system. How do you get distortion products that are just 35 dB below our tones? Then I thought it may be some kind of clipping and indeed, it is. Reduce the level by 1 dB and it sure calms down:
This is still poor by stereo products but at least it is not severely broken.
Here is a dual tone of the same test by the way:
Analog Input Audio Measurements
Just in case you want to use an analog input to get around the poor performance of the DAC/DSP subystem, I also tested the CD input in its Direct mode. First up is our dashboard:
Well, this isn't good either. We have a spray of harmonics. This hurts THD+N versus frequency as well:
Fortunately bad news stops there. Here is our dynamic range:
This is what the digital subsystem should have produced but didn't.
To make sure nothing is being digitized, I measured the frequency response:
Wow, this is excellent! Ruler flat and so wide.
Conclusions
As with the last Arcam AV product I reviewed, we basically have a broken DSP/DAC pipeline but the analog subsystem seems to perform a lot better. And the mechanical and industrial designer did his/her job as well. I say broken in engineering design parlance. Clearly there were no specifications there to meet in digital to analog conversion. They certainly forgot about their marketing talking points of low jitter.
I know these companies can do better. But without management and marketing departments being aware, it is not going to get there. To some extent then, I blame the press that reviews these products without any measurements or cursory, hard to read and understand graphs and numbers. When put in context, we see that the Arcam AV40 lacks the design hygiene of even budget audio products let alone state of the art.
I hope as an industry in general, and Harman/Arcam in the specific, improvements are made in next generation products where we can be proud of what has gone into the implementation of audio in AV products.
Until then, I cannot recommend the Arcam AV40. I hope they can fix at least some of the software issues like multi-tone clipping.
Note: usage of Dirac EQ in this product should make it subjective sound excellent, erasing the pain of the above graphs. I hope in the future though we get to have both: great room EQ and fundamental measured performance.
-----------
As always, questions, comments, recommendations, etc. are welcome.
Amazon Prime shipping was great in how they would send you things in just 2-days. Now that has stretched to nearly a week so I have not been spending as much money with them. Should they fix that though, I need to b ready with a good stash of money to buy stuff to make for the lower purchasing now. So please donate as much as you can using : https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
Note: our company Madrona Digital is a dealer for Harman company which acquired Arcam a few years ago. So feel free to read as much bias as you like into this review.
The AV40 managed to differentiate itself in a see of similar looking products with a unique configuration and attractive design:
The display is large with good resolution and excellent contrast. More so, it lets you fully navigate the unit just like you are used to seeing on the display which is great for manipulating the unit when standing in front of it. Oddly the on-screen display overlaid on your TV/projector is only two lines! This is reverse of what most companies do, skimping on front panel and giving you a full output on the display (although some give you the same two line display on both). I would have wanted to see both mirroring each other.
The controls feel very nice but something is "seriously" wrong with them if you are among world's 90% of the population. Can you see it? Why is the volume control on the left of the display and not right? If you are right handed you have to reach to the left to changing the volume and thereby blocking the display. I would put the power button on the left and volume on the right. OK, this is a minor thing but it stood out for me.
The back panel has the more modern take on AV products, leaving legacy video connectors behind:
Even though there has been serious growing pains as far as software with this product, in my testing where I only touch very little of the product, all worked solidly.
My overall impression of AV40 look and feel is very positive and well above the score I give to many other AV products.
HDMI DAC Audio Measurements
We start with our foundation measurement which is a simple 1 kHz tone, generated digitally as we measure the Front Left and Right balanced outputs. Unless noted, that is the output I am testing. And mode of operation is "Stereo Direct" although I found that it made no difference in measured performance in a couple of places I checked:
The DAC chip that is uses is ESS9026Pro which ESS rates at a SINAD of 110 dB. We are sadly way below that at just 90 dB or so. While SINAD is deteremined by both noise and distortion, here it is the second harmonic that dominates the measurements at -92 dB. This SINAD rating just rescues the AV40 from "broken" category (in red) but by a hair:
Certainly not competitive with anything we are used to even in budget 2-channel audio products:
You can't see it in a static display but as I note on the graph, the left side had a pattern that would come and go. It is interference from the rest of the subsystem bleeding into the DAC's analog output. It becomes a spray of noise as we move up in frequencies to the right of our 1 kHz tone.
Often AV manufacturers put higher quality DACs for left and right channels so I thought I measure a couple of other ones to see how they do:
What the heck??? Performance improves? Notice how the second harmonic distortion has gone way down to around -105 dB from 92. If noise was not holding it back, SINAD would improve a lot more.
You may be asking if the performance improves at other output levels so let's see if that is the case:
As is typical of these AV products, where you set the volume determines different gain stages that get turned on and off, impacting performance. Using the same volume position as previous tests, we see that best performance is at 3.1 volts than 4 volts. Improvement is negligible though to the tune of just 1 dB.
Using higher volume control position gives you more output but with worse performance. So if your amplifier has multiple gain settings, don't go for much over 4 volt "sensitivity" (i.e. level that produces max power).
Back to noise, here is our dynamic range with 4 volts output on balanced out:
Oh, come on now. We can't even clear 16 bit dynamic range in this and age?
The high noise floor seriously degrades intermodulation distortion+noise measurement:
I am not going to raise my blood pressure more by repeating the words on the graph.
I watched a marketing video for an Arcam product and the presenter talked at length how jitter makes music sound bad and that Arcam gets it right. Does it? Let's measure with HDMI first:
Clearly not. The J-test signal has a embedded 250 Hz cycle that toggles all the bits in the sample. It does so by just changing the low order bit in a 24-bit audio sample so should be invisible in our measurement above. Sadly it is anything but invisible. What this means is that what digital data arrives at the input of the DAC, determines what bleeds into the output as unwanted/spurious tones. Classic lack of isolation between digital and analog sections. There are other jitter sources that cause the "skirt" around our main tone and some other spikes close to it.
Toslink is often used for return audio from a TV so let's see how we do with that optical input:
Jitter galore! Clearly no attention was paid to reduction of jitter in this AV40.
Fortunately these levels while visually distracting, audibly are too low to be heard easily so don't panic over sound getting "gritty and such."
Any more sins? Yeh, sure. Here is our linearity graph:
Terrible results. The zigzag graph that amplifies as levels go down tends to indicate truncation of 24-bit samples without dither. There should be no truncation but if there is, for heaven's sake apply some dither to it so it linearizes the levels.
Here is our DAC reconstruction filter:
I don't mind the curve but why is the attenuation just 60 dB? Theory mandates infinite attenuation but in practice, I take 90 dB or more. But not 60 dB. This will impact tests like THD+N versus frequency which use wide band measurements that include components created due to above weak filter:
The blue line is a $99 (or less) DAC board. Let's pick a 10 kHz and see what contributes to that sum on the graph:
Wow, we have a nice buffet for dinner. Jitter, noise, out of band image components, random junk, harmonic distortion, etc. Be thankful your ears aren't nearly as good as my analyzer!
Can it get worse? Turns out it can. Let's look at intermodulation distortion this time with 32 tones resembling "music:"
This is so, so bad I thought there was something wrong with the measurement system. How do you get distortion products that are just 35 dB below our tones? Then I thought it may be some kind of clipping and indeed, it is. Reduce the level by 1 dB and it sure calms down:
This is still poor by stereo products but at least it is not severely broken.
Here is a dual tone of the same test by the way:
Analog Input Audio Measurements
Just in case you want to use an analog input to get around the poor performance of the DAC/DSP subystem, I also tested the CD input in its Direct mode. First up is our dashboard:
Well, this isn't good either. We have a spray of harmonics. This hurts THD+N versus frequency as well:
Fortunately bad news stops there. Here is our dynamic range:
This is what the digital subsystem should have produced but didn't.
To make sure nothing is being digitized, I measured the frequency response:
Wow, this is excellent! Ruler flat and so wide.
Conclusions
As with the last Arcam AV product I reviewed, we basically have a broken DSP/DAC pipeline but the analog subsystem seems to perform a lot better. And the mechanical and industrial designer did his/her job as well. I say broken in engineering design parlance. Clearly there were no specifications there to meet in digital to analog conversion. They certainly forgot about their marketing talking points of low jitter.
I know these companies can do better. But without management and marketing departments being aware, it is not going to get there. To some extent then, I blame the press that reviews these products without any measurements or cursory, hard to read and understand graphs and numbers. When put in context, we see that the Arcam AV40 lacks the design hygiene of even budget audio products let alone state of the art.
I hope as an industry in general, and Harman/Arcam in the specific, improvements are made in next generation products where we can be proud of what has gone into the implementation of audio in AV products.
Until then, I cannot recommend the Arcam AV40. I hope they can fix at least some of the software issues like multi-tone clipping.
Note: usage of Dirac EQ in this product should make it subjective sound excellent, erasing the pain of the above graphs. I hope in the future though we get to have both: great room EQ and fundamental measured performance.
-----------
As always, questions, comments, recommendations, etc. are welcome.
Amazon Prime shipping was great in how they would send you things in just 2-days. Now that has stretched to nearly a week so I have not been spending as much money with them. Should they fix that though, I need to b ready with a good stash of money to buy stuff to make for the lower purchasing now. So please donate as much as you can using : https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
