Core Infinity board

[Confused]

I have shamelessly stolen this recent quote from 'barrows' on Computer Audiophile from the 'Sonore Signature Rendu SE' thread.  For those that may not know, 'barrows' works for Sonore.

"There is nothing wrong with USB when does right, in both sender and receiver.
But, I do believe that the future will be direct ethernet input DACs.  Right now it appears that only Linn and Merging are doing that very well, other DACs with Ethernet inputs appear to be compromised on their Ethernet inputs (probably using poor Ethernet input boards from relatively affordable sources, 192 only, etc...)

What occurs to me is that the new Devialet OS Core Infinity Streamer board may well off just this, although the question remains if they will do it 'very well' or not.  Some reason for optimism I would say?

[Hifi_swlon]

I have shamelessly stolen this recent quote from 'barrows' on Computer Audiophile from the 'Sonore Signature Rendu SE' thread.  For those that may not know, 'barrows' works for Sonore.

"There is nothing wrong with USB when does right, in both sender and receiver.
But, I do believe that the future will be direct ethernet input DACs.  Right now it appears that only Linn and Merging are doing that very well, other DACs with Ethernet inputs appear to be compromised on their Ethernet inputs (probably using poor Ethernet input boards from relatively affordable sources, 192 only, etc...)

What occurs to me is that the new Devialet OS Core Infinity Streamer board may well off just this, although the question remains if they will do it 'very well' or not.  Some reason for optimism I would say?

I guess it's obvious they have the technical ability to design and build complex electronics well, (I'm primarily basing this on the expert Pros and the technology they invented). They were well ahead of the pack in terms of Ethernet streaming - even the concept of AIR - but were hampered by AIRs implementation followed by what seems like a 'car crash' of QA, support, bug fixing and plan for resolution.

So I guess a lot of it depends on their commitment. If they see it as a window into Qobuz etc, rather than a state of the art, flagship, competition slaying interface, probably not. If the latter is their goal, and they've upped their software division in terms of skills and resources, and there's a full desire from senior management and investors to make it a 'best in class product', I'm sure it's possible.

Because other brands at least say something about what they're up to (the Merging QA on Computer Audiophile as an example) you get the sense that those brands are really on top of the technology. Since we dont hear anything but fluff about it from Devialet - with practically zero mention of technology - my hunch is it will be as good as the current solution, with much more useable/accessible features, but that perhaps Devialet aren't getting into all the nitty gritty of extracting 'the last 10%' from it.

I'd love it to be a class leader, and them reveal some staggering new tech that everyone raves about, but despite my enthusiasm I just don't think will happen. That said all I really cared about personally was RoonReady and since it seems like that's coming, I'm done. The differences between AIR, the microrendu, and the Allo digiOne are so small to be practically non existent in my setup so I'm not sure I'd really even detect much improvement. But if it had a good Room Correction setup for example, that would have way more obvious and audible benefit to me personally.

I guess whatever happens it will be clear where Devialet stand in terms of pushing boundaries with Expert, and whether the AIR fiasco was simply a one-off and caused by their time being diverted towards expansion/investment. If it's hailed as the best there is we can all relax and know they've taken the problems of the past on board, addressed them, and come out fighting. If it's a bit half-arsed, or even just unimproved by the standards of those here who extract the extra from Mutecs etc, then we'll know what that means too...

I'll actually just be happy if it launches on time, they have slots to fit it straight away, and the unit comes back sounding (and looking) as good or better than when it left. And that all the software related things are stable.

As a side note, I would have liked to see some sort of forum that was promised, because no matter their intentions there will be some bugs/issues, and right now they have no proper mechanism other than every individual contacting support and then sitting in silence wondering if things will get fixed (hence the 'dark' AIR years for me). For a more technically advanced product, the support/community backend has to match.

And yet jumping on the barrows quote, 'only' 192 is pretty much all us Devialet users can have anyway, so in that sense if this is seen as a meaningful obstacle to having a good implementation, were already doomed unless the 'tweaks' to the mainboard are actually them swapping the whole lot out for next gen DSDx10 technology.

[Pim]

That said all I really cared about personally was RoonReady and since it seems like that's coming, I'm done. 

We never saw SAM coming. Neither did we have any idea the we were looking at a loudspeaker when we saw the first pictures of the phantom. So it's not entirely impossible Devialet are going to pull another rabbit out of their hat with the streamer board. The streamer board might not be what you think you want but what you don't know you can't live without.

[Confused]

I'm kind of with Pim on this one. One thing we know about Devialet is that they are not a company for endlessly communicating technical information or engaging with customers on forums, I think we can agree on that one! But irrespective of this, they have still delivered some fine products and a few surprises. So the lack of communication so far does not tell us anything. Prepare for the unexpected! Meanwhile you can remain optimistic or have pessimistic moments, it will make sod all difference to what actually happens.

Picking up on Hifi_swlon's 'only 192' comment, I know that some folk get very excited about DSD files of ever increasing sample rates and file sizes. OK, I get this, and to be honest if that is your thing, then maybe the 'intrinsically PCM' Devialet is not the best choice for you? (although the DSD MAT engine does seam to work very well and may be increased to DSD128) For me, It is has a very long time since I have bought anything above 24/48, and I have never found anything 'native' DSD that I would actually want to listen to. Currently the vast majority of what I listen to is 16/44. It is what it is.

[Hifi_swlon]

I agree with that and am the same. I was just picking up on your quote which seemed to generally imply that 192 limited technology was a bit dated and wasn't going to deliver stellar performance. Personally I can barely detect upsampled 192 from 44.1 so it's not a direct concern of mine. Well see I guess.

[Music or sound]

I have shamelessly stolen this recent quote from 'barrows' on Computer Audiophile from the 'Sonore Signature Rendu SE' thread.  For those that may not know, 'barrows' works for Sonore.

"There is nothing wrong with USB when does right, in both sender and receiver.
But, I do believe that the future will be direct ethernet input DACs.  Right now it appears that only Linn and Merging are doing that very well, other DACs with Ethernet inputs appear to be compromised on their Ethernet inputs (probably using poor Ethernet input boards from relatively affordable sources, 192 only, etc...)

What occurs to me is that the new Devialet OS Core Infinity Streamer board may well off just this, although the question remains if they will do it 'very well' or not.  Some reason for optimism I would say?

USB can sound pretty good when sender and receiver are well designed but it has an inherent flaw: lack of error correction in isosynchronous transmission . See Gordon Rankins statement http://www.digitalaudioreview.net/2016/0...ty-varies/
Ethernet transmission (like Ravenna protocol) can allow for error correction (and potentially different USB protocols but it should be implemented at higher rates than USB2.0 allows)

[thumb5]

Although there is no error correction in isochronous USB transfers there is error detection, as noted in the article. Errored packets will typically be dropped by the hardware, which is usually easily audible (as glitches). So it's not likely that there is anything subtle going on due to the lack of error correction unless you can obviously hear the glitches.

[Antoine]

Devialet has "leaked" a more detailed picture of the board on their new website.  https://www.devialet.com/en-us/expert-pro-audio-system then choose Technologies -> R.A.M. but I've pasted the image below too.

Low and behold, there's a Audinate IC there! That probably means some sort of Dante/AES67 support. There was a reference to Dante/AES67 for Devialet's 'Immersive Theater System' I.T.S. using Phantoms here: https://devialetchat.com/showthread.php?tid=4111 so it could be there to integrate an Expert in such a system or as an AIR replacement of course since Dante can stream audio over IP (AoIP).

Can also notice a new powerful XMOS XU216-512-TQ128 controller (USB), WiFi board, Cyclone IV FPGA, some type of CPU probably running the new Linux based OS, some memory, etc. etc.. No time now to decipher the rest and who knows what's on the other side of the board as well.


   

[Antoine]

There already was a SRC4392 (sample rate converter) on the current DUET board http://www.ti.com/product/SRC4392 but there's an additional one of AKM now, the AK4137EQ: http://www.akm.com/akm/en/product/datash...o=AK4137EQ No idea if it's good.

Also a Marvell 88E6240 ethernet IC: (https://wenku.baidu.com/view/ce1839a4bce...6baf5.html)

The Marvell 88E6240 device is a single-chip integration of a 7-port Ethernet switch with four integrated Fast Ethernet transceivers, and one integrated Gigabit Ethernet transceiver. This device supports the latest IEEE 802.1 Audio Video Bridging (AVB) standards. The device uses these AVB technologies to identify and reserve network resources for AVB traffic streams and supports precise isochronous streaming capability.  These AVB protocols enable timing sensitive multimedia streams (such as digital video, audio, or industrial control traffic) to be sent over an Ethernet network with low latency and robust Quality of Service guarantees. The device contains four 10/100 speed transceivers and one 10/100/1000 triple speed Ethernet transceiver  (PHY), one Gigabit SERDES, and two digital interfaces that support a combination of GMII, RGMII, MII, and RMII interfaces in a 128-pin QFP package.

[Confused]

Thanks Antoine, fascinating stuff. I wonder if Devialet had actually intended to release the picture or not? As you mention, it is under the 'RAM' section. (Record Active Matching) Either way, this all looks fairly positive to me, maybe the sound quality will be very good via the new board. Time will tell.