Building a DIY Music Server

[ray-dude]

1 hour ago, matthias said:

 

Congratulations,

you are in good company with others like @romaz who found the same.

 

When you have an excellent server this "endpoint" concept comes to an end. It is very enlightening to read the Taiko Extreme thread on WBF. When you combine the Extreme with an endpoint the SQ is inferior to the Extreme alone. The same when the Extreme itself is an endpoint and has another server as source. 

 

For me endpoints do not make any sense at all, the best is to have an UHQ server with excellent USB output into an DAC.

 

Enjoy:-)

 

Matt

 

I am continually being surprised by the impact of network traffic (and network topology) on SQ.  The EtherREGEN started me down this path (surprise -> attention and study) but the rabbit hole is much vaster than signal regeneration.  Honestly, I thought this would be at the minor tweak level, but there seems to be something fairly fundamental going on here.

 

Having just switched from a 2 box i7 NUC setup to an Extreme, the impact of network traffic and topology has become at least an order of magnitude more obvious (more surprise -> more attention and study   I'll have more to share next week, but when running the Extreme as an end point (Roon, HQP, etc) on the native Windows or booting to Linux from a USB stick, you are giving up a lot compared to running everything on Extreme in a one box solution.

 

 

 

 

[ray-dude]

I did similar back of the envelope math before I purchased my Extreme. Of course, DIY allows you to calibrate how far you travel into the crazy pit, and to spread the pain over some period of time.  With Extreme, you are all in and liberated from decision making, but definitely at a price.  If you're willing and able to go all in, I think there is tremendous value in the Extreme (much more on that next week...).  I also tremendously appreciate the amazing work and insights folks are sharing here...there is so much to learn and to figure out!

[ray-dude]

1 hour ago, Dutch said:

Roon works as long as endpoint(s) and server are in the same VLAN (or subnet in a non VLAN’ed network). Control devices can be in a different VLAN but one has to manually enter the server’s IP address once in the Roon app.

 

@Dutch can you please speak more about this?  I'm experimenting with this currently (my Roon server on 192.168.4.x, my laptop and the rest of my home network on 192.168.3.x).  I ended up having to setup port forwarding for UDP 9003 to see the Roon server from my home network, and TCP 9100-9200 to actually be able to connect to it.

 

Looking at the Roon config screens, I'm not seeing anywhere obvious where one can manually enter a server's IP address?

 

I have not looked to setting up routing between networks (I have an EdgeRouter X).  I have a similar goal is to isolate the Roon server as much as possible from other network traffic

 

[ray-dude]

10 minutes ago, Nenon said:

This is an Extreme, cost no object, version of the Audio Mirror Reflection 45W Set monoblocks with custom transformers, Mundorf Mtube and Mlytic capacitors in the power supply, Z-foil resistors all over, Duelund CAST PIO tinned copper capacitors, exotic wiring, WBT Nextgen connectors, Furutech NCF inlet, etc. The list goes on. I use four of these to bi-amp my high efficient speakers, which have 18'' woofers. I tried Jeff Rowland 625 S2 amps for the woofers thinking that might be the ultimate bass amplification. But those SET monoblocks outperformed the Jeff Rowland by a huge margin in the bass section. It's unbelievable how much more detailed and vibrant bass I got with those. That was one of those shocking and completely unexpected moments. I was so biased towards the Jeff Rowland that it took me quite some time to believe it. But it sounds like this is a small toy compared the stuff you are talking about :). 

 

...and this is EXACTLY why I'm in complete denial about the existence of SET monoblocks...something like this in the house would become a full time job (obsession?)  Gorgeous work Nenon!

[ray-dude]

Maybe I'll get a nice Class D with switching supply and have that remind me what I don't pay attention to amps any more   If I waver, I need you guys (as my sponsors) to call me out and help me hold the line...

 

That being said, I would be delighted to foster any SETs that are looking for a home during these troubled times

 

 

 

 

[ray-dude]

16 hours ago, Nenon said:

One of these DIY computers would cost you anywhere from $4,500 to over $12,000. Here is a breakdown:

 

 

Fantastic Nenon!  I did similar back of the envelope math earlier this year, then added in a fudge factor because while I knew I would eventually get to the more premium end of the spectrum, I also knew I would feel compelled to travel the intermediate steps to get there.  That's how I got to spitting distance of jumping on the Extreme train and skipping the intermediate 5 steps.  It's amazing how much power (and to a lesser extent clock) dominates the calculus.  We're basically building power supplies and clock generators with some electronics and storage attached.

 

I continue to have a nagging tingle from my spidey sense that Emile's "mother of all capacitor banks" for filtering will have a LOT more running room on the power side of the equation than the incredible supplies from Sean and Paul et al.  If one were to have (theoretically) perfect filtering capacity with an arbitrarily large and fast filter bank, your options on the power side of the bank REALLY open up (who cares if you have a noisy supply if the filter bank takes care of business?).  From a DIY perspective, Sean certainly leans heavy into crazy large filters, but even he is an order of magnitude off what Emile is doing.  Has anyone seen another power supply maker take a similar strategy?

 

 

[ray-dude]

 

The other wrinkle is Emile sharing that high quality clocks actually take away SQ in the Extreme (although he is understandably proprietary about his clock strategy).  He is definitely taking the road less traveled when it comes to power supplies and clocking.  

 

That being said, that many rails of Sean Jacobs DC4 rails would be something to behold!  I only have one Sean Jacobs supply (DC3'ish) but it was custom built/configured for my DAC, so I haven't had the opportunity to hear this bespoke supplies further back up the digital chain.  Alas, I suspect that by the time my SR7 custom build gets here, we will have all moved on to Mr. Fusion power supplies

 

 

[ray-dude]

34 minutes ago, Nenon said:

And clocks work hand in hand with vibration treatment and power supplies.

 

A thousand times this!  When we finally unravel all the secret sauce, I suspect a huge part of it will stem from conceptualizing all of these as aspects of a single component in the design.  

 

I keep relearning basic metrology - what you measure/observe is only as good as the integrity of your weakest reference.  Especially for digital systems, perfect reference power and timing, and a lot of "problems" just go away!

 

And since we're sharing untold stories, somewhere in the garage I "unofficially" have a memory board from a Cray 1S. As a whipper snapper engineer, I learned so much from studying that board (and for programming for the Cray).  The entire board design was about clock and power distribution, with a twist of cooling mixed in to keep it from (literally) melting. Complete brute force design, like a lot of Cray designs back in the day, but overwhelming brute force focused like a laser on the key things that mattered most.  I was thinking a lot about that board as I studied Emile's design.

[ray-dude]

5 hours ago, Nenon said:

I wish we could all have a well burned-in Taiko Extreme server and a top class DAC in our systems for a few days just to hear what's possible with digital and decide for ourselves if it's worth to take that journey.

 

Outstanding post @Nenon.  I wish above were true as well.  There are rewards in the journey, but the journey is much easier to tolerate (and motivate) when you have a sense of whether the rewards are worth it for you.  

 

Each of our brains may hear the same things, but how we respond to these things is VASTLY different.  At this level of investment and refinement and nuance, indeed it is different courses for different horses.  Best you can do is hear a potential end game, then decide if it is a game you want to be playing.

 

For example, listening to my current chain with local content (network disconnected) vs network content is so danger close that it has definitely soothed my FOMO when I read of all the exciting things folks are doing with switches and their network chain.  Yes there is a difference, but not enough to motivate insane focus vs other things that have my attention right now.  I will get back to that lever, but I am delighted to be pressing other "happy" buttons right now.

 

Looking forward to post Covid times to open up for listening parties again...as much as I appreciate forums and shared experiences here, I miss sharing this journey with others in the room hearing and reacting to things at the same time.

 

 

[ray-dude]

52 minutes ago, Nenon said:

You should try Windows LTSC and ProcessLasso. 

 

I have been having a LOT of fun with this combo on my Extreme.  My recommendation is to carefully study the block diagram.  All the answers are there, and become VERY obvious as you play around with CPU affinity, memory affinity, I/O affinity, etc, and balance priorities.  Stunning what an impact even the smallest tweaks make to SQ. As a hint, pay very close attention to which PCIe slots have affinity to which CPU, the UPI channel between CPUs, and the affinity of built in I/O to CPU1.  As you get to next level, pay attention to the affinity of memory to each CPU.  This block diagram is the roadmap to design and balance processing and playback.

 

 

[ray-dude]

34 minutes ago, OAudio said:

Hours and hours of fun working through all this

 

But SO satisfying then it all clicks, you look at the diagram, and you know exactly where to make tweaks (and then they work!)

 

(and spot on insights, thank you @OAudio!)

[ray-dude]

I can't help myself!  ;)

 

I'm on the alpha for Taiko Audio Server (the Taiko music player) and I'm doing a lot of experiments with upsampling right now.  Learning a lot about how to tweak up the Extreme to different configs.  Looking forward to the Taiko USB card at some point, which should be a game changer (the block diagram will show why ;).  I suspect there will be a part 6 to my Extreme review at some point (one year in type of thing), once the tweaks and updates accumulate a bit more.

[ray-dude]

30 minutes ago, Dev said:

the question wasn't so much about who can buy what. Some folks seems to just think that the source is more important than the DAC itself. Take a hypothetical example, if you have $40k for your digital rig, would you spend $30k for the Extreme and $10k for the DAC or vice-versa ?

 

I have your hypothetical ratio (Extreme to Chord DAVE) and am delighted.  That being said, if someone offered a trade for any of the more upscale DACs (1 for 1, have to keep it) I wouldn't take it.  DAVE hits WAY above its weight class for me, esp. with the extra attention I've been able to give it.

 

I tend to not view things from a more important lens though.  What is most important depends on where you system is as a starting point, and how balanced your system is.  In my Extreme review, I tried to distill my experiences into a couple succinct ideas:

• First Law: Invest in cabling, power, and digital hygiene to do as little harm as possible (directly or indirectly) to the DAC’s clock, ground plane, or reference voltage plane
• Second Law: Invest in a digital endpoint that moves bit perfect digital data from ethernet to a USB DAC with as little variability and as much timing and signal integrity as possible (aka, as close to precision real time data streaming to DAC as possible)

The DAC is the core of the First Law. Every DAC has an intrinsic noise floor limit and timing accuracy limit and distortion limit.  That is the best that DAC can ever do. Everything else is about letting the DAC perform to its full potential. Power optimizations, mechanical optimization, RF, cabling, etc. all help get you there, but so does signal optimization to the input of the DAC.  As I detail in my just posted DC4 review, the DAVE has such amazing intrinsic performance for noise/timing accuracy/distortion that it scales incredibly well with better power and better mastered content and better input signal quality.

 

That leads to the source, which for me is the Second Law.  If your DAC is already able to operate at its intrinsic limits, the source won't help as much.  If the I/O of your DAC is being held back by your source, it will have a huge impact.  In my case, the intrinsic noise/timing accuracy/distortion limits of the Chord DAVE are so otherworldly, the Extreme (which is the state of the art for the Second Law, IMO) had a staggering impact.

 

I should note that the general audiophile community takes things like DACs more seriously than it takes sources (bits are bits, etc)  Given that bias, in general an investment in sources may have a bigger impact for most, but mainly because of where folks are typically starting from.  All things being equal, I would much prefer to run my DAVE with a RPi or NUC than a lesser DAC with the Extreme, but I am an absolute addict when it comes to the timing accuracy and transparency of DAVE. For others, the calculus is sure to be different.

 

That being said, I think there is part of the calculus that should be the same for everyone. Although not published quite yet, in part 2 of my current review I looked at the impact of the Sound Application PGI TT-7 power line conditioner. It had a massive impact on everyone of my components (Extreme, DAVE, bass units to my speakers, and even my TV).  That has lead to a new Third Law:

• Third Law: Invest in the cleanest and fastest and highest capacity power delivery to all your audio (and video) components

So which is most important?  I think the question is which combination gives the most balanced investment to get to the best Sound Quality (alas, everything still matters).  Start with awesome power and awesome content, then build sources and DACs that reveal everything that is there to be revealed.

 

[ray-dude]

I would love to hear it, but I suspect 8fs input may be the achilles heel of this DAC.  That puts a fundamental limiting on transient timing accuracy, unless they have a MASSIVE amount of compute to push to higher rates with idealized filters.

 

The power, construction and clock all look first rate.  I'm skeptical about the synthetic feedback scheme and what that does for noise shaping, but they may inject it prior to noise shaping?

 

Lots to learn!

[ray-dude]

To add to that Ted, TAS allows you to choose your playback engine: the proprietary Taiko one or HQP (in case you want upsampling), similar to what you have with Roon. TAS is tuned and optimized to how Taiko has provisioned and configured the Extreme (and soon, their Evo server).  I'm running the TAS alpha, and using HQP+NAA for the playback engine, so I can take advantage of upsampling for streaming content.

[ray-dude]

3 hours ago, Nenon said:

 

I think you read a bit more than what was written in that post.

Emile did not refer to fixing corrupt files. Whether the file is stored on a NAS, locally, or on a USB drive, or somewhere else, it would still be exactly the same file with the same md5 hash for example, but it sounds different. 

He is not saying that Taiko has found an algorithm to fix corrupt files. What he is saying is that they are into something with the TAS software that makes the difference of playing the same file from different locations vanish. 

 

From my experience using the alpha release of TAS, above is spot on.  Very clever and very thoughtful implementation, with meticulous and systematic attention to every detail.  Emile and team are bringing the same focus to software as they do to hardware and system configuration.  

[ray-dude]

1 hour ago, Marcin_gps said:

I will wait for @JayM reply to the question in the post above before I descrbe other technological advantages in the OPTIMO ATX.     

 

Perhaps best to wait until both products are in the hands of real people?   With listening impressions, a discussion of technological advantages and how they impact listening impressions would make a lot more sense.

 

I'm very much looking forward to both products being in the wild, and hearing what people hear.  

[ray-dude]

VERY excited this project is finally public.  Over 6+ months of helping tune these filters, the progress has been remarkable, and the results extraordinary.  We're really approaching some fundamental limits here on digital processing, and it has been humbling to hear what was always there in digital content, just waiting to be revealed.

 

Well done @Zaphod Beeblebrox!

 

(Disclaimer: although I'm delighted to have been helping on the project since basically day one, I have no financial interest in the project, I paid full price for my license, etc.)

 

EDIT: to not derail this topic, I started a new one for PGGB:

 

 

 

 

[ray-dude]

8 minutes ago, Marcin_gps said:

Are you OK to put it up against the OPTIMO 3 DUO? 

 

FWIW, when I wanted to do a review of the DC4, I approached Sean Jacobs and Sean was very supportive. Perhaps suggest to Jay that he do the same?

Nenon was kind enough to build the demo unit, based on Sean's request, but the DC4 is Sean's product.

[ray-dude]

1 minute ago, Marcin_gps said:

Regardless of the outcome, I’m looking forward to the shoutout.

 

 

Me as well.  Very much looking forward to reading Jay's review of the Optimo 3 Duo, and definite icing on the cake to read how he thinks it compares to the DC4

[ray-dude]

The star scheme made a big difference for me in my experiments with energizing DC supplies (pretty big impact actually).  In the end, I got the best result with (I kid you not) 1/8" thick copper electrical bars.  That was the lowest I was ever able to push output impedance (make sure to ultra sound clean those screws and washers in acetone!  The impact of reducing output impedance goes way up as you go from 100mOhm to 2mOhm to <1mOhm).  I should note that this was an experiment in extreme wiring.  In the real world you definitely want insulation around the exposed parts of the busbars (with those caps charged up, that is a massive amount of energy available to instantly discharge)

 

[ray-dude]

1 hour ago, oneguy said:

Also like to point out that Aluminum was just fine for the Taiko extreme transformer wall. 

 

Don't forget the impact of the bath towel underneath ;)

[ray-dude]

58 minutes ago, Nenon said:

Thank you, John. It was more like 50%-60% the first days!

BTW, give me a call if you want to come and grab my prototype ULPS to play with. It will be sitting unused after next week.

 

Yeah, I saw your post on WBF and was planning to send you a PM. I was surprised you ordered an Extreme as you had a similar server to my DIY server. Well, congratulations! Get some help when it arrives - it's heavier than I expected. 

500lb hydraulic lifts are your friend ;)

 

when I first got mine I had to enlist the civilians in the house to help lift it. They were beyond terrified that they might drop it

[ray-dude]

There is a reason it never gets turned off Larry ;)

 

[ray-dude]

My other vice these days is big ugly 300lb boxes that make a lot of noise and have a lot of flashing lights (aka pinball machines). Movers like this (and hydraulic lifts) are quite common.  It is also unfortunately common to run out of friends to help move things by the time you're moving your 3rd or 4th machine from floor to floor.

 

At the risk of streams crossing, I am in the process of upgrading the speakers/crossover/amp in one of my machines.  Needless to say, given what happened with my music setup,  this is a very dangerous path to start down.