Merging Anubis

[Miska]

Copied from the other thread to move here:

 

  33 minutes ago, malarz said:

@Miska 

That's very true, but here one has only 4 line outs for a room correction what is very often not enough

 

Hmm, why not? For ordinary stereo pair two line outs are enough. I'm not talking about digital cross-overs, for such Hapi or NADAC would be better. Although two separate channel pair line outs are good for digitally crossing subwoofer(s) and main channels. But just room/speaker correction.

 

But separate outs for speakers and headphones are good because then I don't need to keep switching my matrix setup when switching between headphones and loudspeakers. Most pro-audio interfaces expose these as separate channels. While most consumer interfaces expose just two channels towards the computer.

[Miska]

6 minutes ago, Em2016 said:

@Kal Rubinson mentioned :

 

According to what I have heard, the Anubis is not particularly convenient for multichannel because "Two channels come out through XLR connectors, two through line level phone jacks, and the remaining two through the headphone jacks."

 

It is very convenient, but in a different way! As mentioned in other message it allows splitting stereo headphone and speaker processing to two simultaneous independent outputs. Plus one extra pair for subwoofer!

 

[Miska]

11 minutes ago, Em2016 said:

But one big question... How does headphone output volume control work?

 

Digitally controlled analogue vol control (so DSD content is not decimated)?

 

Or pure digital vol control (PCM only like RME)?

 

I don't know how it is implemented in Anubis. But ESS has non-decimating digital volume control for DSD. Similar way as the older CS4398 (when not in Direct DSD mode). This piece is always on the way with ESS since it doesn't have a direct mode...

 

[Miska]

37 minutes ago, Em2016 said:

It's very small and light and PoE just adds another useful function for ease of use.

 

Ahh, it has PoE? That is very nice indeed! I have other PoE equipment already, such as WiFi access points where the reduced number of wiring makes life so much easier and installation cleaner.

 

[Miska]

11 minutes ago, Kal Rubinson said:

Yes but that is not really what is meant by multichannel audio (as distinct from stereo).

 

Yes, that is more like multichannel in pro-audio sense. But since the product is primarily for pro-audio, it fits that bill.

 

[Miska]

18 minutes ago, matthias said:

Anubis uses ES9026PRO and not ES9028PRO as DAC chip like DA8P.

@Miska what are the differences?

 

I don't know... I haven't seen datasheet for that one. And ESS is very secretive in general. Seems to be somewhat lower specs and less pins (smaller chip). I'm not sure what are the pins that have been cut...

[Miska]

18 minutes ago, matthias said:

@Miska

Do you have plans to get an Anubis?

 

Maybe in some scope for testing HQPlayer Pro recording features... So far I've been testing DSD256 recording with RME ADI-2 Pro...

 

 

[Miska]

1 minute ago, Em2016 said:

Based on specs on paper (that’s all we have for now) do you expect better recording resolution than ADI-2 Pro?

 

Or impossible to tell from current info.

 

Someone said it uses AKM chip too? 

 

You can compare the tech specs if you like. But both are probably very good. ADI-2 doesn't have mic preamps, so that is one difference. So if you count mic pre's in as an extra component, pricing of the two is also similar. Quite different feature set though...

 

[Miska]

4 minutes ago, bmoura said:

Since HQ Player 4 Pro supports level changing and channel mixing in the DSD domain up to DSD512, having a Merging Anubis on hand would be a great way to make sure HQP 4 Pro's recording and mixing features work with the Anubis as well as the Hapi and Horus.  A win/win for HQ Player, Merging and their pro recording product owners I'd think. 

 

Yes, first I want to get Hapi verified on all three OS platforms. And then look for doing the same with Anubis. Just a bit short of hands here with all the things! Three OS, and now both input and output, in both PCM and DSD to verify is making twice as complex than just player side!

 

[Miska]

2 hours ago, matthias said:

Yes, AK5578 like Hapi.

 

ADI-2 Pro has AK5574, so a 4-channel version of the same. So in that respect performance is similar.

 

[Miska]

1 hour ago, ferenc said:

By the way, in the last few weeks I was trying all sorts of different preamps with my DACs, to see if there is a benefit to let the mentioned DACs running without volume control or in direct DSD (where it is possible)  mode to my PMC and Quested active studio monitors. There is a definitive benefit. I just could not find any preamp which did not give more weight, more life to the sound of my system.

 

On my desktop system, at the moment I have Schiit Jotunheim as a headphone/pre-amp. Sources are ADI-2 and Spring1. Very good price/performance ratio.

 

But I'm planning to switch to a Benchmark HPA4 that has all the things I need. Two balanced inputs, two unbalanced inputs, plus balanced outputs, including mono output for subwoofer. They have also LA4 pre-amp only version of the same.

 

It is not very easy to find a headphone focused preamp with balanced I/O. Most preamps have a headphone connector, but not so much thought or effort put into the actual headphone amp section...

 

[Miska]

With ADI-2 you need to run it at 705.6/768k PCM or preferably DSD Direct mode at DSD256. With DAC chips, just forget about using DAC chip digital volume controls, be it AKM or ESS.

 

[Miska]

8 minutes ago, Em2016 said:

Ethernet doesn't make it automatically better SQ... I'm sure @Miska will confirm that a poorly implemented network interface can measure worse than RME ADI-2 DAC, the latter which doesn't seem to have any of the normally discussed USB issues...

 

So my only point is, implementation matters even with networked interface...

 

Certainly, in technical jargon I would call Ethernet an "enabler". Something that has number of properties that can be used to an advantage. But it doesn't happen all by itself, or as a byproduct.

 

Some things you get almost automatically, unless messed up. You get isolation between devices as long as you don't use something like SFTP cables, and you get all the networking flexibility as long as you don't try to do point-to-point.

 

4 minutes ago, matthias said:

I can imagine that Ravenna is superior to Ethernet UPnP/DLNA.

 

This is like comparing web browser to a BitTorrent client. Two completely different things. I can have HQPlayer Embedded as UPnP Renderer with output to RAVENNA. Sure, both can go over Ethernet, but that's pretty much it... They are just conceptually entirely different.

 

[Miska]

4 minutes ago, matthias said:

Therefore it makes sense to compare SQ of devices in the same price range but with different interfaces like Anubis vs. Holo Spring2.

 

If you want to compare interfaces, it would be better to use something that has both Ethernet and USB to the same DAC. If you compare Anubis and Spring you are certainly comparing mostly other things than the way you connect a DAC. The two have very very few things in common.

 

[Miska]

1 hour ago, kravi4ka said:

Does anyone have an idea how to control the Anubis? I mean what software player could be used?

 

You can use any player, on Windows anything that supports ASIO, on macOS anything that supports CoreAudio and on Linux anything that supports ALSA.

 

On RAVENNA you can deal with device configuration using web browser and audio routing with Aneman:

https://www.merging.com/products/networked-audio/aneman

 

[Miska]

28 minutes ago, Em2016 said:

But actually there are already Ravenna input speakers but not Zman based.

 

 

Ahh, another Finnish product...

 

But I think Ravenna/AES67 will go with pro market ahead and then trickle down to consumer products over time. Just due to the technology design history / background. And it may be better that way in the end.

 

[Miska]

1 minute ago, Em2016 said:

No I meant they just generally may not want to announce product plans, separate to NDA.

 

That's normal for companies, of course. Merging themselves won't tell you what their next Zman based product is...

 

I think that's the best. I hate vaporware that gets announced years before anything materializes. Such is bad for everybody in the end.

 

What I liked about some Apple announcements over the time are the ones "this thing is available in our store right after this event"!

 

[Miska]

5 hours ago, Em2016 said:

 

Hi Jussi, which ethernet cable would you recommend for PoE?

Would you recommend sticking with Cat 6 UTP or does PoE changing things and you'd recommend STP?

 

I see Belden recommend their Cat 6A 10GX/10GXS series for PoE applications, which uniquely has the 'floating shield' design.

 

 

Some of my PoE feeds are on STP and some are on UTP. Depending on overall grounding topology.

 

If the shield is disconnected at the other end, then it doesn't pose typical problems. Just good to pay attention which end of the cable has the shield connected.

 

 

For example for my WiFi access points shielded cables would be pointless, because the shield wouldn't be connected anywhere. Cabling coming to wall sockets is unshielded and the connectors are plastic body both in the patchbay and in the wall sockets. In addition the access point's connector is plastic body.

 

[Miska]

28 minutes ago, asdf1000 said:

I know you normally recommend UTP but for using PoE (like with Anubis) would you recommend STP cable, since it is quite higher voltage compared to normal, non-PoE? 

 

Will the shielding provide some benefit with PoE?

 

Will using PoE with UTP cable mean significantly higher radio emissions from the cable? Especially with analogue gear around nearby?

 

STP will easily create grounding trouble, so I would avoid mixing it with audio. PoE will naturally break the galvanic isolation Ethernet would otherwise have. But it doesn't affect radio emissions because it is DC voltage and the Ethernet signal itself is is differential on the twisted pair unaffected by the DC offset.

 

[Miska]

7 hours ago, asdf1000 said:

If PoE will break galvanic isolation with UTP cable, does this mean ground loops are still possible with UTP cable with PoE?

 

Ground loops not, unless there is something fishy with the PoE feeder implementation. It should be floating power feed, a bit like the 48V phantom power for microphones.

 

[Miska]

1 hour ago, asdf1000 said:

So with HQP OS and Ravenna support to Merging DACs - do they support DSD256x48? Or only DSD256x44.1 ?

 

DSD only at multiples of 44.1k

 

1 hour ago, asdf1000 said:

Is HQP OS the only supported way to play from HQP Embedded to Merging DACs?

 

If you build the driver yourself, you can use something else too. But you may have limited functionality.

 

From my side, HQPlayer OS is the official way.

 

[Miska]

2 hours ago, asdf1000 said:

Hmm hopefully my 9900K can handle 48k base rates to DSD256x44.1 

 

With ext2 and ASDM7EC

 

Lot of CPUs can do such rate conversions if you use poly-sinc filters like -s2 variants or ext2. My i5-7700T can do that to 44.1k x512 too.

 

ASDM7EC is the heavy part where 9900K helps, but there the source rate doesn't matter.

 

9900K can certainly do that combination.

 

[Miska]

1 hour ago, asdf1000 said:

So I have Anubis but I am currently thousands of kilometres from my i9-9900K HQPe server...

 

I can wait til I am back to my server but I remember you saying NAA can be other side of the world from server... I'd like to confirm if that is true or a lie ! 😀

 

Is there a simple and secure way? Or is it complicated?

 

You can use some VPN that forwards also multicasts. I think something like OpenVPN with Ethernet tap should work. Everything else is straightforward as long as you get the discovery working over VPN.

 

[Miska]

On 2/6/2021 at 9:04 AM, asdf1000 said:

 

Hi @Miska when I look at the attached, where is the galvanic isolation technically broken? Or does this image not really show it?

 

Is it where the voltage loop formed by the +/- shown for both Mode A and Mode B? That's the phantom power similarity you mentioned?

 

 

 

It does, at the point where it picks up the electricity. With PoE you have that extra wire that goes to primary side of the ethernet transformer.

 

[Miska]

31 minutes ago, asdf1000 said:

So technically this wire connecting ethernet transformers for power, can be path for SMPS 'common mode noise' (leakage currents) to go downstream?

 

Yes, theoretically. Depending on how the power is used. Likely it goes through some switching regulators and transformers though, so not necessarily an issue.