High End Music Server (and why I've not taken the leap)

I am convinced that up stream digital source matters. That said I wanted to dive into the topic as as see it, because IMO a high end AIO server may not be the best fit for many.

So why are high end music servers like the Antipodes K50 & Kala, Aurender N30 SA, Taiko Extreme popular? Well, because they work as advertised. My point isn’t that they don’t do what they claim to. My hesitation comes from the fact that it’s an exercise in what can be done with best practice, no expense spared noise mitigation and signal path optimization and it works, but…

Why go through the exercise of all this when they’re starting from the worst possible place to start, the PC.

Once you take a look at what the money spent is going towards, my reasoning for an alternate solution, one that starts off not in a PC, not in a pace conducive to all the noise. I think in the end you can probably get more for your money. That said though, the point of an AIO music server is convenience, one chassis and control over everything in that chassis. If you are using that server as intended, meaning the housing/storing playing the digital file and rendering the digital file, and you’re connecting it directly to your downstream components, then a server may be what’s right.

My thought process however is that for someone like me, where my file storage and processing is being done away and isolated from the sensitive clocking or conversion to I2s the massive expense incurred of mitigating that noise in a high end server is not required.

In my case having a Roon Rock server withe the internal storage (very noisy environment) while isolated from the sensitive stuff (the bridge, the reclocker, the DAC) is what I mean by starting in a better position. Not having to throw a lot of money at the mitigation of noise in a PC based server means (or should mean) money better spent on the bridge, reclocking, power etc.

Different way to skin the cat, but putting it out here for discussion and simply open to information that may go against my prevailing thought here.

7 Likes

I can’t convince myself either for similar reasons.
The Taiko though is as much about the software as it is the hardware as I understand it, they use their own drivers and playback software which might be playing with the bits.
If your output device is plugged into the server, it very much matters.

I understand the reason we’ve gone with “heavyweight” computers as streamers and servers, 90% of the problem is software, most of the vendors couldn’t afford to develop and maintain their own stacks, and you can use Linux, or in Taiko’s case windows and it’s a mostly solved problem.

It’s been on my list of things to try and build an absolutely minimal streamer for years now, you could play hires audio from a 100MHz arm chip with an RTOS minimizing latency issues. You could probably get away with a < 10MHz if you limited yourself to uncompressed audio, but since you have to have an Ethernet interface on there and that will probably have a 100MHz CPU in it, you may as well use that.

The idea being lower powered lower frequency devices by their nature put out less noise.

What’s stopped me to this point has been the software stack, to keep it minimalist I think USB is probably the easiest way to go, so you need a UAC-2 driver, and I haven’t been able to find enough information to write one, and until recently I couldn’t find one.
Maybe I’ll give it another go when I get through my current projects.

5 Likes

I’ll offer a counterpoint, which is just my opinion.

Let’s say I want a really good pizza.
I could go to a restaurant. I want that Michelin star chef to use his 30+ years of experience and trade secrets to deliver me an end product, along w/ a smile and for the waiter to fold my napkin. The chef sources his cheese, flour, wood/gas/fuel, and other ingredients; he might have designed his own oven and tools. Some chefs might grow some of their ingredients, but none can do everything.

Others might say, I can buy those ingredients and make my own pizza and it tastes just as good. Sure. Some will even make their own ingredients. Go mine your own coal, or grow your own tree for the wood. Sure. I wish I had the time/bandwidth for that, but I don’t. Instead, I pay through the nose for someone else’s hard-earned expertise and in-house developments that would cost me WAY more to try and re-invent.

Same thing w/ winding my own transformers, making my own cables, writing my own streaming software. I’d rather spend my time enjoying someone else’s product than toiling to make something that (probably unsuccessfully) would sound equivalent.

Same thing w/ cars. I like knowing a company w/ 70 years of history is making my super well-tuned engine, steering, suspension and transmission. Could I cad something up and machine it / metal print? Probably. Will it be as good? Maybe. Will it be as reliable as a stamped higher volume part? Absolutely not.

Personal Bias: I used to be in a field where everything can be measured, proven, statistically modeled and re-proven. Now I’m in a field where it’s more art than science. I believe in experience/expertise and rely on such many times a day.

Personal Admission: I fear loss of time/opportunity, increasingly so as I get older and run out of it. Time is now more important than money. I’d rather buy and enjoy, than enjoy the making process. Maybe that’s because I suck at making :slight_smile:

5 Likes

I get your analogy and I’m not disagreeing with the results of that chef making that pizza. I don’t think your comparison suits what I’m trying to say though.

While using your same analogy: what would that pizza using the same ingredients and given my the same 30+ years experiences etc. But asking that chef to make you that pizza in an igloo in the arctic circle. The costs for that product would be needlessly high to attain the results we want. Much better to start where that chef has his Michelin star restaurant today.

And that’s my point, first is the use case I describe for how you’ll be using that high end server there. (Storing, distributing connecting directly to your DAC) If it is, then it makes a lot more sense to proceed. If the use case isn’t there, there it doesn’t make a lot of sense.

3 Likes

Well… I’m assuming it’s cost prohibitive for Taiko et al. to reinvent things down to the circuit layout, making their own fabs, processors, ram, SSD, etc. Aurender does some of that. They use proprietary circuitry, not sure what code they use and if everything is on some FPGA.

When I interviewed Lucas Domansky, definitely had the same questions. Why did he chose to use an intel box? What else did he try and consider? How did he design his triple-clocked SPDIF output card? It was very interesting how he approached it and how he evolved to using his current platform. He used to sell his servers w/ an Ideon USB reclocker, but now his servers are performing better than w/ the reclocker. In fact, he has his own proprietary Roon output drivers since he found sound degradation w/ Roon’s default output. Who knows, maybe in the end a very high CD transport’s way better :smiley:

Alas, I’m just a lay person who plugs in cables and prays things work :slight_smile:

2 Likes

I’ve heard this before. Imagine? lol

And yeah the SW aspect as mentioned by both @Polygonhell and @keithc is a big component but it’s not something I’m even factoring in as I’m taking the less complex argument first, cost to engineer HW to give off less noise.

I’d even go as far as taking Roon out of the picture here and just using apples to apples server with a open source transport protocol.

3 Likes

I’ve explored different combinations of digital sources from the following:

  • Straight from the PC
  • Roon through ROCK
  • Bluesound Node
  • Sonore ultraRendu, opticalRendu, ultraDigital, opticalModule
  • Denafrips Iris
  • Singxer SU-6
  • Auralic Aries G1 with the Hoer Wege PSU uprade
  • Holo Audio Red (as source and as DDC)
  • Antipodes S30/S20/S60
  • Antipodes K50

I’ve done optical networking with linear power supplies and all of that stuff too…

Nothing has come close to the sound quality of the K50 connected through I2S to my DAC. It’s now sitting after a opticalModule with a switch powered by LPSU. As long as I can, my money is with the antipodes flagship.
Can you get to that level on your own? Perhaps. I certainly couldn’t.

4 Likes

I really want to hear the Oladra. It looks slick.

2 Likes

In your case, not surprising as like I said, you’ve got the server feeding the DAC directly and via a not widely adopted i2s interface. Do you have a good sized local library and are you storing it in the K50 as well?

and I think this is a very valid statement. I’m a tinkerer by nature those people that can’t leave well enough alone :slight_smile: and there is a lot of peace of mind that comes from one stop shopping one neck to choke philosophy as well.

Out of curiosity I just looked for internal pics of the Oladra versus the K50 and there’s a hell of a lot more bespoke HW. Would love to geek at some detailed images.

images
K50


Oladra

1 Like

Very fascinating discussion!

I know @Polygonhell will check me on this :joy:, in my own limited experience I found using a CD transport to be an improvement over running a streamer. However the connivence wins out once again, what I will do is pull out some of my favorite CDs which I enjoy the whole album, as some such as Kid A by Radiohead tell a musical story, and listen through my CD transport through my DAC.

Problem is due to the decline in CD usage (some will say it’s making a comeback, I’m talking back in the heyday of CD use), there are not many options available new. With used we run into the issue of mechanical failure, and not all CD drives are created equal. I’m way off topic, my point is a CD Transport is a good alternative too.

2 Likes

Having heard both, i think the S-stack is darn close. But i agree antipodes has some sauce.

1 Like