Audiocontrol Hyperion Processor

[Kal Rubinson]

From their website:

View attachment 495789

View attachment 495790

I've seen that but I have been assured otherwise. We will find out soon.

 

[Highfidelity]

I've seen that but I have been assured otherwise. We will find out soon.

In order for the CI to have both the Digigram Dante PCI-E card, and also have a hardware based Ravenna card the unit would need to be redesigned from the ground up before the official launch. Because the piece of hardware they have now called the Altitude CI doesn’t have the space for it.

One thing that is feasible is the Digigram Dante card could be swapped for the Digigram Ravenna card: https://www.digigram.com/products/s...smpte-st-2110-pcie-sound-card/#specifications

And after a bunch of custom programming the unit could be Ravenna hardware based. But then lack proper Dante and native AES67 support. And still not have the ability for on the fly sample rate switching. As it doesn’t use the Merging Zman for Ravenna.

 

[Highfidelity]

This is what Trinnov uses for Ravenna/AES67:

bondagit/aes67-linux-daemon.git: from https://github.com/bondagit/aes67-linux-daemon.git

bondagit/aes67-linux-daemon.git: from https://github.com/bondagit/aes67-linux-daemon.git

gitee.com

 

[welwynnick]

At this point I’m thinking the winning setup will be the Trinnov CI over Dante with Dante network set at 24/96. This way only 44.1 and 48 will be up-sampled. And SoX based 64 bit floating point DSP SRC that Trinnov uses is very transparent upsampling to 24/96. And native 24/96 content will be sent bit-perfect through the whole chain to DAC.

I thought we'd established that there's always an ASRC in front of a Dante output because the source clock and Dante clock aren't synchronised.
Trinnov said so here. Therefore a Dante output can't be bit perfect even when the audio format is the same.

I also have a Appsys Flexisys AES with Dante daughter card.

That's interesting. With or without ASRC?

 

[Highfidelity]

I thought we'd established that there's always an ASRC in front of a Dante output because the source clock and Dante clock aren't synchronised.
Trinnov said so here. Therefore a Dante output can't be bit perfect even when the audio format is the same.

That's interesting. With or without ASRC?

Ya but there’s no need to alter the audio in the ASRC if the incoming and outgoing sample rates are the same. All it does is aligns the clocks for timing. If outputting a different sample rate then there’s a possibly for artifacts in the algorithm to degrade the sound quality. However this is minimized a lot when using 64 bit floating point DSP.

I have the one without SRC. This way I know the signal is bitperfect when passed though the box. On the downside it only works when my incoming sample rate matches the Dante rate. I use an Apple TV as my source and it’s a non issue as it outputs everything at 48.

 

[Highfidelity]

I suppose when we focus on ASRC breaking the system from bitperfectness, we are forgetting about all the other sections of an AVP that also does.

For example even with the holy grail setup being the AES/EBU outputs on an Altitude CI, rarely will we have bit-perfect output to the DAC in real-world use, here’s the complete list of what forces processing:


• Dolby Atmos decoding (object rendering) → always breaks bit-perfect


• DTS:X or Auro-3D decoding → always breaks bit-perfect


• Optimizer room correction → breaks when active (bypass = bit-perfect)


• Any upmixer (Trinnov Active Decoding, Dolby Surround, DTS Neural:X, Auro-Matic, etc.) → breaks when active


• Bass management / crossovers → breaks when active


• Speaker remapping / 3D acoustic relocation → breaks when active


• Any target curve or manual PEQ → breaks when not flat


• Loudness management / Dynamic EQ → always breaks


• Digital volume control not at exactly 0 dB (reference level) → breaks


Only when every single one of the above is completely bypassed/off/flat do you stay 100 % bit-perfect on the AES/EBU outputs. In practice that means:


– playing native PCM multichannel (no object decoding)


– Optimizer bypassed


– no bass management, no remapping, no EQ, volume at reference


The moment you turn on the features people actually buy an Altitude for (Atmos, Optimizer, bass management, etc.), bit-perfect is gone – even on AES/EBU.

 

[welwynnick]

I suppose when we focus on ASRC breaking the system from bitperfectness, we are forgetting about all the other sections of an AVP that also does.

Quite right. Though you might have missed one - DACs themselves often have ASRC these days.
Yet ironically it was you that flagged up the audible degradation of the CS8422 ASRC that led to the use of the Appsys Flexisys without an ASRC.
Before that, I never really had a feel for the sonic sins of an ASRC.
I sincerely hope that Trinnov do avoid ASRC with AES/EBU outputs, but I'm certain they use ASRC with Dante output, which implies resampling. They said they did.
Remember that it was gdickens himself (Elytone Dolby/Dante architect) who said:

So generally, all cases of HDMI -> Dante/AES67 will be going through some form of ASRC.

Having been a hair-shirt audiophile, I now find myself pondering using both Audiolense/HLC, PLUS an upsampling tool like HQP or PGGB, to get the digital filtering right.

 

[PGAMiami]

HX16-AIP & HX16-AES

Now that H16-AIP is shipping and no longer in back order, we can start to talk about HX16-AIP (Customers who purchased an H16 series unit directly from Nuprime-X.com are eligible to upgrade to the HX16 series by paying the price difference. Please note that this trade-in program may not be availa...

audiophilestyle.com

Should be similar to DPR16

 

[PGAMiami]

It will be interesting to see if a two box solution of the Nuprime HX16 or Audio Control DPR16 with an Okto or Merging DAC outperforms the APR16

 

[Highfidelity]

Not even close. That Nuprime box is a joke.

 

[PGAMiami]

Not even close. That Nuprime box is a joke.

Have you seen the HX? It’s almost 4x the price of the H.

These are all very complicated. One thing I can say about my lowly Denon 4800h is it just works. The older I get, the more reliability matters to me.

Which brings me back to the question of an APR-16 versus a DPR-16 plus a Merging Hapi versus an AV-10 or 20.

 

[Highfidelity]

The HX is a vaporware concept. Based on all the promises and failures of their other processors I have zero confidence this box will be good even if they ever deliver it which I doubt.

 

[PGAMiami]

I don’t doubt you. These processors are asked to decode many formats and interface with other components for CEC, copy protection, etc.

Do you have experience with Audio Control? De bugging one of these must be a nightmare.

The APR-16 seems very carefully laid out. Wonder how quiet the DAC and analogue outs are. Alternatively, how reliable a DPR-16 with outboard DACs could be. And whether it’s best to use AoiP or AES/EBU. Would these outputs have a well implemented volume control.

 

[welwynnick]

The APR-16 seems very carefully laid out. Wonder how quiet the DAC and analogue outs are. Alternatively, how reliable a DPR-16 with outboard DACs could be. And whether it’s best to use AoiP or AES/EBU. Would these outputs have a well implemented volume control.

Hyperion are still claiming 130dB DNR and 120dB THDN. It's not entirely clear whether they're quoting DAC chip specs, or actual measurements for a real, live, complete, production standard, piece of equipment. They're state-of-the-art results for a minimalist stereo DAC at any price, and quite without precedent for a jack-of-all-trades AVP. The absolute best Anthem, Marantz and McIntosh AVPs don't quite reach 110dB THDN yet. I'll have to remain cautiously sceptical for now, but if Hyperion can be shown to be honest, the functionality and performance of the APR takes away much the need for digital audio outputs. Those numbers are comparable with Merging and Okto.

Has anyone seen a user manual for the APR-16 yet?

 

[welwynnick]

Not even close. That Nuprime box is a joke.

PGAMiami was talking about the HX16, which won't have the same limitations as the H16. To your credit, you pointed out that the H16-AIP uses SRCs that cause audible degradation (hence your use of the Appsys). However they're used on the AOIP board to enable it to synchronise with Dante, and are not needed or used on the H16-AES. That has it's own limitations, such as a number of outstanding bugs. Without exception, all AVPs have bugs. To their credit, Nuprime have issued two firmware updates, they have publicly stated what the outstanding bugs are, how they're prioritised, and are engaged with the user community. The H16 also has only 12 output channels, not 16, and it performs DSP at 48kHz (like Storm). Maybe all that wouldn't be acceptable if it cost $10,000. Sure, the Trinnov ALCI will be better. But the H16 costs less than a 2 channel upgrade for the Trinnov.

Yes, we all know by now you don't like the Nuprime. So don't buy it. Please change the record.

 

[PGAMiami]

As I understand, the when using Dirac most every processor and AVR will resample to 48khz. That’s not necessarily a bad thing if the SRC is done with appropriate math, like 32 bit floating point. One question is will DPR-16 connected by AES to two Okto Pro8 outperform an APR-16? We will have to wait until these are available. The HX-16 is expected to sell for $3000, and I don’t know that includes Dirac ART. Either way, these three options are all similarly priced in the $7000 to $8000 range. A bit more than a Marantz AV20 or AV10 on sale. Dirac and Trinnov both attempt to manage room modes. Dirac seems more flexible on placement of the subwoofers. Should be a fun year. For now I’ll optimize my speaker layout using the Denon 4800h in prepro mode with Dirac ART?

 

[welwynnick]

To the best of my knowledge, the Datasat RS20i (with all options added) will process Dirac at 96kHz , but that's perceived as obsolete now.
From what Kal has said a few times, Dirac will also work at 96kHz on a suitable PC.
But with all other applications, it works at 48kHz.
It remains to be seen what the APR/DPR-16 will do.
My guess is that a DPR + Okto will out-perform an APR-16. The Okto is a proven performer.
I wouldn't be tempted to pull the trigger just yet, though.

 

[Kal Rubinson]

From what Kal has said a few times, Dirac will also work at 96kHz on a suitable PC.

Dirac will work up to 192kHz on a PC with suitable SW.

 

[Highfidelity]

PGAMiami was talking about the HX16, which won't have the same limitations as the H16. To your credit, you pointed out that the H16-AIP uses SRCs that cause audible degradation (hence your use of the Appsys). However they're used on the AOIP board to enable it to synchronise with Dante, and are not needed or used on the H16-AES. That has it's own limitations, such as a number of outstanding bugs. Without exception, all AVPs have bugs. To their credit, Nuprime have issued two firmware updates, they have publicly stated what the outstanding bugs are, how they're prioritised, and are engaged with the user community. The H16 also has only 12 output channels, not 16, and it performs DSP at 48kHz (like Storm). Maybe all that wouldn't be acceptable if it cost $10,000. Sure, the Trinnov ALCI will be better. But the H16 costs less than a 2 channel upgrade for the Trinnov.

Yes, we all know by now you don't like the Nuprime. So don't buy it. Please change the record.

They made it clear that they didn’t even know how many channels their processor could process until it was actually in the hands of customers! They bought some cheap third party Chinese board and popped it in a box and pumped it up before they even knew its capabilities and limitations. When the same company who has proven they have no idea what they’re doing make some pie in the sky promises about some vaporware box that is only a conception in their minds, my expectations are low.

Then there’s the clown they hired to represent the company online. Who has absolutely no idea how the product he represents even works. I could pull some random fentanyl junkie off the street, sit them down for 3 hours teaching them about the product, and they would do a better job representing the company than that guy ever did.

 

[Highfidelity]

Regarding the APR-16, there’s absolutely no reason the THD+N they claim wouldn’t be honest. I seen how well the engineering department designs the internal boards. And with my experience designing several dacs using the ES9039Q2M, I can say there’s never been a dac chip made to date that is so easy to match the specs of the datasheet performance in a real world implementation. They made it almost idiot proof to get flawless results very easily. Which wasn’t the case with any chip that came before it. So knowing this, unless you have specific needs such as having Dante input active speakers, the APR-16 is likely the way to go. If you’re just running Dante or AES from the DPR-16 to a DAC in the same rack you’re not taking advantage of any of the reasons audio over IP was deployed in the domestic home theatre market.

The purpose of audio over IP in domestic HT applications is distributed audio. Put the AVP in a single room and network to DAC/amps, or active speakers located around the home. Map the hdmi from several TV’s through the AVP so a single AVP can be the processor for several rooms. And the dacs/amps can be located as close to the speaker drivers as possible without analog cable losses. If you’re not taking advantage of a system like this, and 16 channels is enough, just get the AVP-16. That’s the reason they make it. For you.