Focusrite RedNet (AOIP Nervosa)

Yes, we've had my Lynx in @drfindley 's new system. I think we're both curious about Rednet comparatively now ... if @3X0 is interested.

 

Thad & abob both had very good arguments. Being a network geek I may have gotten too excited. New toy syndrome and so forth. I love the fact that his place ain't HF. Down to earth impressions and reports all the way. Keep up the criticism!

What I'm hoping from this RN3 madness is an elevated possibility to remove obstacles between me and music by pinpointing flawed or excellent gear properties in the chain. Or maybe I should just listen to music more and wank less with the red Porker.

 

Good attitude! I get caught up in the geekdom too and remind myself often: Stop! Get beverage. Find music in library haven't heard in awhile. Enjoy.

What you are getting with the Focusrite boxes is a lot of well thought out engineering, both hardware and software, to meet very demanding studio and show environments. Imagine 64 channels in from stage and 32 out to stacks that did not deliver time synchronized samples. It is already a challenge to time align multiple sound emitting cabinets for FOH and stage monitors. Even more so if there are short stacks or monitors further out in the crowd. If the digital snake introduces unpredictable phase delays the show is never going to sound optimal and possibly abysmal.

Now anecdotally I thought I heard my U12 improve the MoMB a little more than expected lately. Noticed that after installing the Singxer SU-1 XMOS DDC driver the U12 was using the new driver as well. I have an experiment in the queue to force the U12 back to the original driver for comparison. And why would this make any difference. So I acquired a USB Trace tool for further, future examination. So many experiments. So little time. See first remark in this comment.

 

Ya know…

If you were running a Yggdrasil and were only using 44.1KB or 48KB to optimize it's performance this might be a decent way to go.

But you still have to supply 24VDC to run it.
And at $600 (plus the PSU) it would still be cheaper than an RN3.

Even so it might 'pay to wait' for what's coming down the pike, so to speak.

JJ

 

Went back to square one to catch up on some basics. Others that feel the need to do so might find this SoundOnSound article interesting and useful. Whilst focusing on the RedNet stuff, is covers some of the whys and wherefores of this technology in studio and performance sound systems.

While reading the above article, I recalled reading that, in a finely-tuned large auditorium, the timing of sound delivery is finely adjusted to arrive at delivery points distant from the stage at the right times, rather than either at the same time or... whenever it just happens to get there. In this kind of context, latency and predictability matter a lot. Whether any of that is applicable to our homes, or even our home studios, is moot, but it all looks like very groovy technology anyway.

 

Yes, the right times are tuned through the system amplifiers' (plural for large shows) crossover / processors. The digital snake should deliver same time for digital audio transport. If the snake starts mucking around with sample synchronization between channels, all hope for good sound is lost.

 

My local dealer quoted S$2200(USD1582.28). My budget can only fund either RN3 or the Schiit Freya + something else.

Guess I might have to go for the PCI-E cards or settle for Dante-AES-O format converter + freya

 

The AOIP threads I've read made me worry this was going to take hours to set up properly. I have no formal training in networking and I was able to get the RedNet integrated in about fifteen minutes. It's possible Focusrite ameliorated the setup difficulties that consumer end-users were experiencing a few months ago.

Step-by-step:

1. Install RedNet Control (this also installs Dante Controller) -- it also pushed a firmware update to my RN3
2. Install Dante Virtual Soundcard
3. Configure the Dante Controller to map the DVS channels into the RedNet (this is the part that I imagine is not intuitive for most people):
   • Expand the matrixes in Dante Controller and click where I've got the green checkmarks (inside the red box) so that your DVS 1&2 channels are the transmitters, and your RedNet 1&2 channels are the respective receivers.
4. (Optional) Install VB-Audio Virtual Hi-Fi Cable. This adds a virtual sound device through which you can route audio. This is useful for routing output of Tidal through ASIO, but you can also hook exclusively to the virtual device e.g. WASAPI through a media player. In general I elected to go through this step because my computer is general-purpose, so I'd rather not have to switch between ASIO and WDM in DVS just because I want to go from music-listening to gaming or Netflix.
   • This works perfectly with SR Follow. I was able to float between 44.1KHz through 96KHz output with no manual intervention required on my end.

Note that as mentioned earlier you need to keep RedNet Control open for SR (Sample Rate) Follow. As of right now opening RedNet Control and clicking SR Follow is the only manual thing I need to do on boot-up, and that's only to enable automatic sample rate switching.

Overall I was pleasantly surprised by the painlessness of setup and operation. I have a low tolerance for tweaking and bullshit these days but I found it relatively intuitive except for the Dante Controller routing bit.

Performance

My core system specs:
Intel Core i7-5960X @ 4.4GHz (8 cores, 16 threads)
32GB DDR4-2666
ASRock X99E-ITX/ac motherboard with Intel I218V/I211AT Gigabit LAN
Windows 10 Pro

The RedNet is plugged directly into one of my built-in LAN ports. FWIW I'm using ac Wi-Fi over an Intel 7260 mPCIe card concurrently with no issues. I'm using a DigiDesign AES/EBU snake cable to go to the Yggdrasil.

I didn't have any dropouts over the initial two hour listening period. I may have had a dropout this morning, but I'm not sure. I'll keep a careful ear out this weekend.


Need some more time to formulate a well-rounded opinion, but initial impressions are positive. Not night-and-day like going to megaburritos or the T2 were this year, but I feel there are improvements to resolution, definition, imaging and separation -- especially in pieces that sounded comparatively congested/blurry/indistinct straight over USB. Acoustic, live, classical pieces are outstanding in direct comparison. It really reveals the subtle things like room acoustics and spatial minutiae that USB smears over.

 

From @Marvey's meet impressions:

I'm getting a little curious as to how different the RedNets might be from the Lynx stuff now. I secretly hypothesized that they would sound nearly identical, but I'm not so sure now. I wonder if @drfindley or @bazelio heard differences.

I've been a little lazy/reserved about my Rednet impressions since I don't believe I'm sufficiently experienced with digital transports (AES or otherwise) to opine something of value. Direct comparison to the competition is a different story.

 

I did spend some time (not a lot - maybe 3 or 4 familiar songs) with Rednet D16 and Lynx E22 at the meet. We briefly had a Roon network with @RKML0007 's MBP and my MBP in sync and the chains were nearly mirrored aside from the Rednet vs Lynx: Lynx/Rednet -> Mutec (clocked by Live Clock) -> Yggdrasil -> ZDS (same tubes) -> Utopia. We had the ability to select one input to hear Rednet or another input to hear Lynx. Volume was nearly identical. And the result..... I'd say it was a coin flip. We still had differences in interconnects because we simply didn't have enough of one type of interconnect (see Marv's photo of the wire jungle). I do wonder if the audible differences could be explained away by the interconnect differences and I do suspect that the two systems were highly aligned. By the end, I think I slightly preferred Lynx E22 along with @shaizada ,while I think @drfindley slightly preferred the Rednet system. I heard slightly more left to right soundstage spatiousness and slightly more laid back treble with Lynx. Not sure what the observations of others were. @TMoney also briefly did the A/B and, as I recall, he simply said "If you blindfolded me, there's no way I could pick them out."

PS Those who mentioned volume differences were probably referring to the A/B of Rednet and MicroRendu. I didn't hear but the slightest volume difference with Rednet and Lynx.

PPS Standard disclaimers apply vis-à-vis the challenging "meet conditions" for evaluations. I'd also like to hear the same through speakers.

PPPS DAC connections were via AES, and we went XLR out of Yggdrasil.

 

I kinda wish I still had a Rednet 3 to compare to the Lynx E22. When I was comparing them before, I was using S/PDIF format. Switching the E22 to AES format for my current DAC gave me some perceived improvements over S/PDIF. When I brought it up to PCX, they confirmed that AES is the preferred format (Sonic Frontiers SFD-1 MKII SE+) if it's available, so I probably wasn't hearing nonsense.

At this point, I think the E22 beats the Mutec MC-3+USB alone in AES format, but I'm still not sure where it would sit against the Rednet 3.

 

I still want to hear this in my own system, I'm having a hard time feeling confident in my opinion at the moment.

 

I found myself tinkering with the implementation of my upcoming Dante network. This might be useful to someone with similar complex network needs. All unrelevant network components and services have been obfuscated from these ramblings. Basic idea in this madness is to control Roon server mainly with my iphone as the family ain't too fond of IT/audio shit lying all over the house. MacBook will be used from time to time as it has great WAF. This writeup will be updated and more specific configs will possibly follow. First phase will be mostly tech oriented, audio/dante/impression ramblings will follow later.

Objective in a nutshell

• Ditch USB once and for all
• Network controlled music server / streamer utilising Dante
• profit!

Chain:
Roon remote control for iphone -> wireless -> Roon Core -> 1GB Dante LAN -> RedNet3 -> DAC -> HP AMP -> HD800 / Speakers.

Core hardware

• pfsense fw (1GB interfaces) -> access control + DHCP server
• Cisco SG200-18 (1GB interfaces) -> L2 switching
• OpenWRT capable WLAN router TP-LINK Archer C7 (1GB interfaces) -> dual SSID 802.1q capable WLAN ap
• DSL -> porn
• Focusrite RedNet3 -> transport

My setup is overkill for the most of the people. One could just plug the DVS equipped laptop ethernet port to the RedNet3 or just use one 1GB/100Mb switch. Well life if too short for easy solutions so let's carry on.

Nerd stuff:


Client / Server hardware

• Intel i5 NUC (Roon Core) + dvs
• MacBook Pro (Roon client) + dvs
• iphone (Roon client)

Note: Audinate DVS does not work with WLAN interfaces so you'll be stuck with LAN.

Audio Software

• Roon server core over Win10 (Intel NUC)
• Roon client (Iphone, MacBook)
• Audinate Dante Virtual Soundcard (NUC, MacBook)
• RedNet Control

Basic Network prequisities

• Gigabit Ethernet
• 802.1q (vlan trunking)
• Network segmentation and firewalling
• QoS

VLANs
10 - LAN (browsing, streaming, you-name-it)
20 - DANTE (muzak only)

VLAN20 will be prioritized with QoS. One could of course use flat topology without VLAN's. Being a nerd will not accept that option.
10 min network scetch to present one possible solution:




TODO:

-Switch config finalizations -> will be uploaded later
-Buy decent DAC -> have to sell kidney
-Harass family with endless tinkering


Staff / Marvey : If this craptastic post is way out of SBAF needs feel free to shoot it down.

 

AOIP / usb nervosa threads would be correct place to discuss about these comparisons. Starting new thread with one word really ain't gonna help your quest for the answers. Or anyone for that matter. Lurk moar before causing unneeded noise.

 

So I fussed with my RN3 and replaced the SMPS and used a LPS instead.

A definite step up even this soon after letting everything settle in

I posted my mod and findings over there.
http://www.head-fi.org/t/781268/the-diyrs-cookbook/780#post_13218514

JJ

 

So I recently picked up one of these:

https://www.micromedia.ch/?portfolio=dante-aoip-stereo-maker-boards

Digimedia DIO XLR. One-in, one-out Dante to AES board. You can get one by emailing them directly. The board with a simple enclosure is 180 EUR plus 15 EUR shipping worldwide. It has 12vDC in and needs about 1amp. Plug and play.

I bought this in the hope that it might be a giant-killer for not much money. In stock form it isn't quite there. It is certainly good, but in a ranking I would say it fits in as:

AES16 > Gen 5 USB > RME 9632 (via AES) > Digimedia DIO XLR = RME 9632 (via S/PDIF) >> Gen 3 USB.

There is definitely a loss of focus and smearing of high frequency transients compared to the top league (AES16 and Gen 5). Dunno if this is jitter or noise.

What could make this interesting for a @Scott Kramer is that there are a bunch of things you can mess around with that might make this a giant-killer. There is a long thread about this at:

https://www.aktives-hoeren.de/viewtopic.php?f=30&t=7306

Things I managed to gather from that thread are:

• The 12vDC input is converted to 3.3vDC via a switching DC-DC converter which can be bypassed.
  
• You can solder in an external clock feed.
• The chip converts Dante to I2S which is then turned into AES via a TI SRC4392 chip. You can bypass that and take the I2S output direct.

At some point I might have a go at some of these myself. For the moment I am sticking with my AES16. If I do ever get round to trying these mods I will report back.

 

I would think that integrating in a network brings its own set of headaches. But straight connection to computer is not that complicated and fairly stable. Nothing a reboot cannot fix usually.

The only thing that still needs attention is the sample rate. It's supposed to work automatically now, and it mostly does, but it needs a manual intervention from time to time in RNC.