A novel way to massively improve the SQ of computer audio streaming

[jbparrish]

On 5/9/2021 at 2:35 AM, tedwoods said:

Lampizator have been using this board in select products since their 10th Anniversary last year. I believe it's pretty much standardised by now and it's in all their DACs. It uses XMOS, opto-isolators, re-clocking and very ultra low phase noise oscillators. It does PCM 768 and DSD 512 (natively).

In my DAC (Lampizator Amber 3) it was a night and day difference over the Amanero, that was installed before. So much that I'm no longer finding a use for the tx-USBUltra. I believe this looks close enough: http://www.jlsounds.com/i2soverusb.html 

I've been actually prompted to mention this because @austinpop's observations painted a familiar picture for me, as they reminded me of mine upon replacing the Amanero board.

I cannot go as far and say that this board has the same effect as the Audiowise bridge and USB isolators , since I have not tried this solution, but I can certainly confirm that the general logic behind it is very true.

 

I contacted the supplier in the link asking if their board was a plug and play replacement for the Amanero board, which I have in my DAC. Here is the response. I am interested in the Amanero styled board, because I can for sure do the replacement, not so sure that I can do the re-wiring unless someone describes exactly what must be done...

 

Hi John,

 

Our I2SoverUSB board is not directly made to swap with Amanero, so the wiring will be different :) but it's straight forward. 

We have Amanero styled board - but we need to do some more tests of it - possibly next week.

 

Best,

Lyuben 

[Nenon]

3 hours ago, jbparrish said:

We have Amanero styled board - but we need to do some more tests of it - possibly next week.

Just wait for this one. I thought it was available to the public already.

[Superdad]

4 hours ago, jbparrish said:

I contacted the supplier in the link asking if their board was a plug and play replacement for the Amanero board, which I have in my DAC. Here is the response.

 

Lyuben is a good guy and his board designs have consistently improved over the years. There are simple things he does right that so many others overlook. 

[Cedric79]

The strangest thing happened recently with my Melco S100. I was disabling the ports

of a Buffalo GS-2016, whilst the S100 was connected to the latter. Upon disabling the Buffalo’s ports, I then realised that the ports of my Melco got disabled as well. I might have accessed the wrong settings page by mistake, but I thought the Melco S100 had no accessible settings? Anyone familiar with an internal

physical reset button in the S100? My question might be OOT and in that case please accept my apologies and suppress

my post. I thought it might be of general relevance as it might happen to others :-)

[R1200CL]

Those that like to push the development of a opticalRendu in the direction of having external clock interface:

https://audiophilestyle.com/forums/topic/63082-the-opticalrendu-with-external-clock-input/

 

[lmitche]

The Noctua passive heatsink is out.

 

 

Looks like it's application will likely be limited to motherboards with 2 channels of memory.

[MarcelNL]

Nice!

 I May not be looking closely enough but I see no TDP rating, just that it's unable to handle base clock without fan on most interesting AMD CPU's.... 

I have a set of copper heatpipes and copper adapter plate on the way from China for the AM4 (Ryzen 7 2700x) that I'll try before thinking about this option. 

[Mustu]

Any audiophile grade usb female connectors one can buy to connect from a intel nuc - I use a Hdplex case 

 

cheers 

[MarcelNL]

I just upgraded the CPU from an octacore to a twelve core and once more I am baffled with the improvement in SQ more processing power brings...it has been reported previously so it's not a major surprise, but still...

 

The previous CPU load in Daphile was 0.X %, so negligable, it is now in the same range. At the same time I notice that the Solarflare card runs cooler, almost as if the CPU takes on more work the Solarflare now does not need to do?

[jelt2359]

On 1/28/2021 at 10:44 PM, ray-dude said:

I have two ERX’s. I go ISP to fiber to “modem” (ONT) to Ethernet ERX1 to SFP fiber to ERX2

My home network and wifi is all in ERX2. My audio network is isolated on ERX1 (it’s own subnet) and I use a Sonore optivalModule to connect ERX1 to my server via fiber. 
 

Getting my audio network on its own subnet was a big deal. So was optically isolating wifi on a different ERX

 

In a bid to try this without creating another VLAN or subnet (worried about control issues across networks), I went to my asus router and blocked internet access for my server (and my switch connecting that server to the router). I play only local files so this works for me.

 

REALLY nice bump in SQ!

 

[ASRMichael]

44 minutes ago, jelt2359 said:

 

In a bid to try this without creating another VLAN or subnet (worried about control issues across networks), I went to my asus router and blocked internet access for my server (and my switch connecting that server to the router). I play only local files so this works for me.

 

REALLY nice bump in SQ!

 

More for info, I use Ubiquiti gear, I have various VLAN’s. My phone (being control) is on VLAN 1, management VLAN. Management VLAN give access to all other VLAN! Across subnets! 

[jelt2359]

14 minutes ago, ASRMichael said:

More for info, I use Ubiquiti gear, I have various VLAN’s. My phone (being control) is on VLAN 1, management VLAN. Management VLAN give access to all other VLAN! Across subnets! 

That would indeed solve the problem, too! My method can work with probably all routers, though. And it’s a really simple change. Of course the downside is no internet access unlike your method...

[jelt2359]

3 hours ago, ASRMichael said:

More for info, I use Ubiquiti gear, I have various VLAN’s. My phone (being control) is on VLAN 1, management VLAN. Management VLAN give access to all other VLAN! Across subnets! 

PS- does this mean that all traffic from VLAN1 gets shared to VLAN2 (if this is where your hifi is)? May still get 'unwanted' signals going into VLAN2?

[ray-dude]

I've found in forum posts, people use VLAN and subnets interchangeably.  They are actually different (logically, and for what the NIC sees)

 

With a VLAN, each packet is tied to a VLAN ID.  At the network router level, there are different logical networks based on the VLAN ID.   For example, on my home network I use separate VLANs for guest network, IoT network, etc.  VLANs let me isolate networks from each other for security reasons, but everything is using the same wiring and WiFi access point.

 

With a subnet, it is a different IP range for the subnet.  At the router, only IP traffic for that subnet is sent to devices on the subnet (no briadcast traffic, etc).  There is no logical separation of the networks, unless you're dropping filters or firewalls on the subnet.  For the optical link from my router to my music server, the only network traffic is for the subnet I have configured for that port on my router.  In this case, I have no security isolation, but I have physical isolation of network traffic on the wire/fiber.

 

I use the later for my music server, and the former for devices I don't want on my home network (guest wifi, car, thermostats, etc).  All the devices on my home network can "see" each other, but only network traffic intended for my music server gets routed to the NIC in my my music server.

 

By not having broadcast traffic, some "discovery" features don't work, but everything I've used so far lets you explicitly configure the IP address of the music server (Roon, etc)

 

What I have found to be effective is to have the NIC on the music server "hear" as few packets as possible.  Subnets did the trick for me, VLANs did not (although when people talk about VLANs, they often are talking about subnets)

 

 

[ASRMichael]

15 minutes ago, jelt2359 said:

PS- does this mean that all traffic from VLAN1 gets shared to VLAN2 (if this is where your hifi is)? May still get 'unwanted' signals going into VLAN2?

I don’t think so, it works on establish/relate. I’m not an expert but based on that and the fact management VLAN is on separate subnet I would say no. Also I tried putting on a broadcast device on management VLAN & my Audio VLAN does not receive the broadcast. This suggests it doesn’t. Maybe we have IT experts here who can confirm. 
 

If one was concerned it does then create a firewall rule with establish/relate it then blocks traffic unless it was established on management vlan first. 

[ASRMichael]

3 minutes ago, ray-dude said:

 

What I have found to be effective is to have the NIC on the music server "hear" as few packets as possible.  Subnets did the trick for me, VLANs did not (although when people talk about VLANs, they often are talking about subnets)

 

 

My VLANs are created using different subnets? They have their own DHCP and DNS. So what’s the difference? 

[ray-dude]

1 minute ago, ASRMichael said:

My VLANs are created using different subnets? They have their own DHCP and DNS. So what’s the difference? 

 

I'm far from a networking expert (I tapped out on that stuff back in the late 90s) so your implementation may be equivalent from a packets getting to your server.  There MANY networking savvy audiophiles here, so I welcome correction/education.  

 

As a pragmatic point, when I went to subnets, I got my server isolated, and didn't have to play games with firewalls, etc (really simplified my home network setup).  I am certain there are better ways to get to the same result though...

[ASRMichael]

Here's how I have it setup. See different VLAN's with different subnets. 

 

 

[jelt2359]

Thanks for the clarification! My router can do that too, but I don't have a spare router to put my server behind- maybe I'll try that next.

[ray-dude]

I should clarify that when I was experimenting with VLANs, I did have firewall rules to connect the networks.  Moving to a subnet let me get rid of all that stuff (I have my music server as the only device on my music subnet, and it is connected to a port on my router that is dedicated to that subnet).

 

Lots of different ways to skin this cat, and I'm 100% sure that someone who actually knows what they're doing will have a better way ;)  

 

Alas, reconfiguring my network was WAY to time consuming to do a proper A/B, but I was happy enough with SQ and simplicity of subnet solution to declare victory and move back to playing with massive capacitor arrays ;)  

 

I will say that I currently have a WiFi switchable outlet on the power supply for my Sonore opticalModule that is at the head end of my music subnet.  With my current playback chain, I can not hear a difference between having the oM powered up (aka server connected to the network) and starting playback and having the oM powered down (server disconnected from the network).  No network is the ultimate reference to compare network optimizations against.  I know several here that still hear a difference between an optimized network and no network, so as usual your mileage will definitely vary.

 

 

 

[ASRMichael]

Sanity check you can absolutely no traffic on VLAN 10 (Audio VLAN). 
 

 

[jelt2359]

9 hours ago, ray-dude said:

I should clarify that when I was experimenting with VLANs, I did have firewall rules to connect the networks.  Moving to a subnet let me get rid of all that stuff (I have my music server as the only device on my music subnet, and it is connected to a port on my router that is dedicated to that subnet).

 

Lots of different ways to skin this cat, and I'm 100% sure that someone who actually knows what they're doing will have a better way ;)  

 

Alas, reconfiguring my network was WAY to time consuming to do a proper A/B, but I was happy enough with SQ and simplicity of subnet solution to declare victory and move back to playing with massive capacitor arrays ;)  

 

I will say that I currently have a WiFi switchable outlet on the power supply for my Sonore opticalModule that is at the head end of my music subnet.  With my current playback chain, I can not hear a difference between having the oM powered up (aka server connected to the network) and starting playback and having the oM powered down (server disconnected from the network).  No network is the ultimate reference to compare network optimizations against.  I know several here that still hear a difference between an optimized network and no network, so as usual your mileage will definitely vary.

 

Quick question- I'm guessing since your music server is the only device on your music subnet, that means your switches connecting to your Extreme are outside that subnet. Why did you make this choice? Why not put the switches in the same subnet too?

[ray-dude]

I have two routers (Edgerouter X SFP). On one ERX I have one Ethernet port configured as a subnet. I put my music server as the only device on that subnet

 

my second ERX is configured as a dumb switch.  It has all my home Ethernet and wifi access point connected to it. 
 

the ERX’s are connected by a short run of optical fiber, which lets me electrically isolate the router with my music subnet from all the goodies on my home network (including the wifi access point)

 

I have fiber to the home, so I go fiber to ONT (high quality power) to Ethernet to ERX router (high quality power) with my audio subnet on one Ethernet port, and the optical port connected to the second ERX configured as a switch, with everything else on home network on that switch

 

Since I put all this in place there have been a series of upgrades to the Extreme that have significantly diminished the impact of network optimizations.  I have not backed out all these tweaks to see how much they still help, but I suspect I could simplify this chain further.  It works well and is a modest investment, so I’ve been leaving well enough alone. 

[bobfa]

I am working on a simple A/B test.  I have fiber and copper to my server an Aurender.  I am currently using the Sonore opticaModule and fiber.   I also want to test using a VLAN to those ports.  

[austinpop]

First Listen: Sean Jacobs DC4 ARC6 PSU for Chord DAVE

 

For over a year now, my DAVE DAC has been powered by a Sean Jacobs DC3 PSU. I remember when I first deployed it to replace the stock PSU, the improvement was shocking, even though I have experienced the effect good PSUs can make on a DAC many times. I knew even then that Sean's DC4 was a further step up in SQ, as @ray-dude reported in his fabulous review, but I was content to wait, as I have had a Paul Hynes SR7 for DAVE on order for a very long time. I won't speculate when that unit will show up, but when @Nenon recently offered to lend me his personal pre-production DC4 ARC6 unit to audition for a couple of weeks while he's traveling, I jumped at the chance.

 

Similar to our involvement with PGGB, @ray-dude, @romaz, and I had been experimenting with @Nenon on PSU "what-if" experiments. I should say that it was those 3 doing hands-on experiments, with me participating vicariously. There is a whole story around how the ARC6 came to be, which will be told some day, but the end result is an enhancement to the DC4 that @Nenon and Sean have developed, and are now finalizing prior to release. The ARC6 version is an evolution of the entire design, but centers around a new capacitor board that goes prior to the regulators. There is more to the design than that, but a full description should wait until the design is finalized for release. The version I got is a late prototype, and is a single box incarnation with one large transformer (both box and transformer are large!). Still being considered is a no-compromise design with multiple transformers (up to 1 per rail), that will necessitate dual chassis.

 

 

Listening Impressions

 

I wanted to experience the DC4 ARC6 and compare it to my reference DC3. I know many of you also want to know how the ARC6 upgrade improves on the base DC4, but the logistics of getting me both DC4 and ARC6 units proved too complicated. But never fear — that comparison will be done soon enough by other trusted AS'ers.

 

Since swapping PSUs on the DAVE is not a simple matter, and involves opening up the DAVE, I was initially fretting about how I'd be able to switch back and forth between my DC3 and this DC4 ARC6. As it turned out, the swapping was totally unnecessary! Just a few seconds of listening to the DC4 ARC6 was enough to know this was on a totally different level of performance. If I was merely shocked going from the stock PSU to the DC3, I was flabbergasted going from the DC3 to DC4 ARC6! I'll describe the differences in terms of 4 different attributes of SQ.

 

Physicality, Density, and Dimensionality

 

One of the most striking things I heard with the DC4 ARC6 was the solidity and physicality of the sound stage. Every instrument moved up a notch in realism, sounding more  fully formed, what I have called dimensional in past reviews. You get a sense of each instrument occupying 3-dimensional space.

 

Bass

 

Nowhere was this more evident than in instruments, like the cello, basson, double bass, and tympani. Even clarinets, harps, and violas took on a weight and heft that rendered these instruments more realistically than ever. This wasn't bass for bass' sake, or exaggerated in any way. It just reminded me more of how actual live instruments sound.

 

Dynamics

 

This was no surprise, as it's often the first benefit of a better PSU. With dynamics, I find the benefit most noticeable when listening at lower volumes. Where playback at the low volume previously sounded boring and unexciting, it now took on energy, excitement, and a greater level of engagement, since crescendos appear more impactful, and diminuendos are rendered with better clarity. 

 

Resolution

 

This was an artifact of the lowered noise floor of this PSU. PGGB dovetails into this beautifully. Because PGGB reveals so much more low-level nuance in the music, the deployment of a PSU that lowers the noise floor allows that nuance to come through. I was hearing details and cues I hadn't previously. This was particularly evident in my PGGBed version of Solti's Mahler 8th, where I could hear instrument groups as a collection of individual instruments, rather than a homogeneous whole, even during full-throated orchestral passages. 

 

In Summary

 

The experience of upgrading from DC3 to the DC4 ARC6 reminded me of what I heard when I went from the dCS Bartók to the Rossini, as I described in my Bartók review. Substitute DC3 for Bartók and DC4 ARC6 for Rossini and the same sentiments apply.

 

The net impact of all these areas of improvement was that I just could not get enough of the DC4 ARC6 while it was here at chez austinpop! Every listening session has gone deep into the night, and has taken a remarkable amount of willpower to end. Suffice it to say, I'm a bit sleep-deprived and very depressed to be packing this thing up and sending it on to Roy and Ray for the remainder of its grand tour.

 

My expectation is that the final configuration(s) of the ARC6 DC4 units will be finalized in the next few weeks. It is also my understanding that Sean will provide an upgrade path for current DC4 owners, which is only fair. If you're considering a PSU upgrade for your DAVE, this one's a slam dunk!