Hyperion APR-16 AV Processor Review

[amirm]

This is a review and detailed measurements of the Audio Control Hyperion APR-16 home theater Audio/Video Processor and pre-amplifier. It is on kind loan from a member and costs US $8,000.

A gorgeous, large and responsive touchscreen anchors the front panel. Both the power button and rotary control feel good. I did not have to use either as I managed it using its web interface. Speaking of that, it is elegantly designed with large fonts/graphics that are very easy to intuit and use, especially for this class of device. By class I mean a processor that is aimed at Custom Integration (CI) channel. Much attention is paid to aid those integrator deal with myriad of HDMI issues such as EDID display and override. There is even a Linux console window where you can see technical messages and maybe interact with the OS directly! I don't remember any processor being so well designed in this regard.

Back side shows extensive connectivity of the kind you want:

Of course you have XLRs for all 16 channels but also RCA (I focused my testing exclusively on the former). You have Dante support for digital transmission over Ethernet. I was also very pleased to see balanced XLR analog input.

If there is one down side is lack of a physic power button. The front panel control turns off the display but I can see a number of blue LEDs still lit. Left the unit overnight and the back side is somewhat warm, indicating some amount of power usage. Leaving the device on all the time likely helps with HDMI compatibility (not breaking the connection) and control even when the device is "on." I did not look in the menus to see if there is a power save mode.

I upgraded the unit to latest firmware before testing:

Hyperion APR-16 DAC Measurements
I focused on the front left and right XLR channels and set the volume to get the nominal 4.0 volt output. Configuration was 2.0 so there was no processing involved:

From memory, this matches company spec and lands the APR-16 at the top of every AV processor and AVR tested:

More output can be had at the cost of some distortion:

Dynamic range @ 1 KHz) is the best I have seen in this class:

Multitone distortion is again class leading:

I didn't see any filter choices. What is there is pretty good:

Linearity is also the best I have seen in AV products:

There is a bit of jitter but again, for this category, this is very good and not an audible concern:

EDIT: New firmware addresses the following issues:

Oh wait.... what is going on with IMD test?

Distortion starts to rise around -40 dB. Strange. Let's put a pin in this and run our noise+distortion vs frequency test:

Oh boy. I have never seen such a rise in low frequencies. Remembering that the IMD test is 60 Hz + 7 kHz, it must be the same problem causing the 60 Hz component raising IMD.

Investigating more, we see a jump in noise floor as frequencies go below 300 Hz:

To confirm, here is our dashboard using 50 Hz instead of 1 kHz:

We see the higher noise floor and drop of 22 dB in SINAD!

What a shame... we were doing so well.

EDIT: here are the new results with the new firmware:

IMD is much improved:

As is THD+N:

Hyperion APR-16 Pre-amp (Analog) Measurements
I switched to XLR input leaving everything else the same:

Note that I had to set the sensitivity to 4 volts. Not sure why it is 2 volt by default as that caused severe clipping. What we have here is worlds better than what we get out of typical AVR. Someone paid attention to this less used pipeline/ADC.

Frequency response is nice and flat with a cliff above 20 kHz, indicating 44.1 kHz digitization/ADC sampling:

SNR is quite good for this class:

Finally, I ran the wideband THD+N test again, this time with analog input:

We see the same rising low frequency noise but not as bad as digital.

Conclusions
Gosh... we came so close to have a home run here. A professional AV processor at reasonable price with chart busting performance, only to get trapped with low frequency response. Given the accuracy of company specs, I have to think the design was fully optimized using measurements. But somehow low frequency noise escaped the test suite. I am wondering if the increased noise is due to some kind of signal processing, even though everything is turned off. I have a vague memory of another processor company telling me pure stereo may have this issue. If so, hopefully a firmware fix can be applied.

EDIT: new firmware addresses the issues I originally found.

As a result, I am going to recommend the Hyperion APR-16 AV Processor.

------------
As always, questions, comments, recommendations, etc. are welcome.

Any donations are much appreciated using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/

 

[amirm]

 

[MediumRare]

I don’t like seeing the anomalous rise in noise (and distortion) as shown in the 20Hz SINAD, BUT it’s still extremely low (-120 dB) and totally inaudible. Right? Given all the device’s many other strengths I gave it a Golfing Panther.

 

[Billy Budapest]

Hopefully this is something that can be addressed in a firmware update. Otherwise, I am very impressed.

 

[Power Pop 23]

'Gosh, we came so close to a home run'......

With his objective measurements, Audio All-Star Outfielder Amir, 'Robs' another piece of gear of a home run.

I am hopeful the manufacturer can address the anomaly.

 

[Fidji]

16 channels is not enough for system with 8k USD processor, in my opinion. 9.x.4 is a minimum for immersive setup, so unless you invest into separate sub controllers you end up with 3 subs, which is also suboptimal.

 

[xanalog]

Thank you for the review, @amirm,

So, what we are looking at here is the 20dB increased ('absolute') noise floor, at all times that 'music' is playing.

At this $8K price-level, would there be any need for BOTH single-ended AND balanced outputs, or any need for RCA outputs to begin with?
I'd be guessing that the required power-amplifiers would equally be equipped with XLRs, especially since some of these output signals may need long cable-runs, to the destination of many of the 16 channels.

I almost gave it a Happy Panther solely because this AVpre/pro has built-in Dante.

 

[darth]

The review got posted on AVS forum, lets see if someone from the brand looks at the low freq stuff

 

[phoenixdogfan]

Thanks for the very fine review, Amir.


Aside from the rising distortion in the bass when using HDMI (but not XLR analog) input, I also see that it supports Dirac ART which is an essential feature, IMHO, at this price point. Still, $8 K is a stiff price for just a processor, getting up into Storm and Trinnov territory. Also have to wonder what it has that makes it worth 2.5 x the price of something like the Mini DSP Tide 16 which retails for $3500 with 16 channels, Dirac ART, a full set of miniDSP tools including PEQ, Bass Management, routing for inputs and output, limiting, etc, and supports of both Dolby ATMOS and DTS-X with at least equally good measured performance.

Maybe Dante, but still.

 

[Mort]

Thanks for the very fine review, Amir.


Aside from the rising distortion in the bass when using HDMI (but not XLR analog) input, I also see that it supports Dirac ART which is an essential feature, IMHO, at this price point. Still, $8 K is a stiff price for just a processor, getting up into Storm and Trinnov territory. Also have to wonder what it has that makes it worth 2.5 x the price of something like the Mini DSP Tide 16 which retails for $3500 with 16 channels, Dirac ART, a full set of miniDSP tools including PEQ, Bass Management, routing for inputs and output, limiting, etc, and supports of both Dolby ATMOS and DTS-X with at least equally good measured performance.

Maybe Dante, but still.

Except it hasn't been tested yet. What if it fails something like this did? It's also way more than 3500 for us residents because of tariffs taxes and shipping.

 

[Sokel]

I would bet that this noise coming from down low is the good, old quantization noise coming from the DSP processing as we have seen before.

The amount is right on the money, we usually see 20-30dB noise rise.

Nice piece of gear though.

Thanks Amir!

 

[welwynnick]

I'm very grateful indeed to Amir for testing this unit when he did.

I'm very grateful indeed to PGAMiami for sending his unit in to test.

I can't say how much I was looking forward to seeing this.

I'm sure AudioControl will be all over the LF measurements. Did they get to see the results first?

 

[milosz]

$8,000 is way too much money for this.

 

[TimoJ]

$8,000 is way too much money for this.

And $799 extra for full Dirac, since it's not included.

 

[amirm]

Did they get to see the results first?

No but the owner has messaged them. The company is close to me so if there are fixes, it can be coordinated easily.

 

[Mnyb]

Gave it "not terrible" due the low frequency distorsion performance , a bug like this can hide other sins in the DSP system I think ?

If they fix it it will be great , suggests an more expanded DSP test suite if they improve on this , we have seen wonky performance in for example Mini DSp when for example EQ are in use

 

[welwynnick]

I would bet that this noise coming from down low is the good, old quantization noise coming from the DSP processing as we have seen before.
The amount is right on the money, we usually see 20-30dB noise rise.
Nice piece of gear though.
Thanks Amir!

I'm a bit hesitant to say this, but I don't think that's true. Amir and MDSimon2 have indeed measured increased LF noise associated with DSP processing, but what they found was a bit different. They did measure increased LF noise, but that was independent of the test signal. There are also multiple instances where noise increased simply because an HDMI input was connected (even if it wasn't the selected input) for example with the Marantz AV-10.

The APR-16 is rather odd in that the noise appears to be a function of the test signal, and is flat, broadband noise. It's most odd.

It means that with a mixed signal input (normally with an LF component) the whole noise floor will always be elevated.

@amirm - I wonder if the increased noise is a function of the LEVEL of the LF test signal?

I'm sure AudioControl will fix this quickly, though, so I voted "Fine".

 

[Sokel]

I'm a bit hesitant to say this, but I don't think that's true. Amir and MDSimon2 have indeed measured increased LF noise associated with DSP processing, but what they found was a bit different. They did measure increased LF noise, but that was independent of the test signal. There are also multiple instances where noise increased simply because an HDMI input was connected (even if it wasn't the selected input) for example with the AVR-10.

The APR-16 is rather odd in that the noise appears to be a function of the test signal, and is flat, broadband noise. It's most odd.

It means that with a mixed signal input, where there's almost always LF signals, the whole noise floor will always be elevated.

I'm sure AudioControl will fix this quickly, though, so I voted "Fine".

I'm not sure either of course, is just a guess.
But the pattern looks familiar and the rise is about where it should be.

To be honest here, it also looks like the problems that come from a DI box for example or a small transformer, so yep, just guesses.

For a piece of gear with lots of subsystems though thorough investigation by the company is expected.
Let's see.

 

[amirm]

@amirm - I wonder if the increased noise is a function of the LEVEL of the LF test signal?

You can sort of see that effect in the IMD vs Level. I could test it directly but I am not sure what that would tell us. The company can troubleshoot this much faster than we can.

 

[welwynnick]

You can sort of see that effect in the IMD vs Level. I could test it directly but I am not sure what that would tell us. The company can troubleshoot this much faster than we can.

I was curious that the problem didn't seem to manifest itself in the multi-tone test, which looked really good for an AV product.