I have to say Taiko came a long way with the Extreme I wasn't expecting to hear what I heard . The Extreme is a masterpiece and I can say today the N30 isn't even close to the performance of the Extreme, I'm glad I gave the Extreme another chance .
the greatest value in the Extreme is that it‘s a constantly improving platform. Or as Steve W likes to put it: You hop on the Taiko Train.
I'm hooking into this as this is a good example of a changing environment affecting baseline performance. Our networking environments and usage have changed over the past 5 years since we designed and launched the Extreme. Back then network utilisation was a fraction of what it is today, we have much higher performance home networks and infrastructure, but also a rapidly increasing number of devices and applications creating activity, often even when not in use.
Network "tweaks" back then could give you a small performance boost, or introduce a small amount of colouration you might like, but in todays environment these tweaks and your home network setup have become mayorly influential, up to the point where they can even make or break your performance.
5 years ago the playing field was different to what it is today, and this affects your designs choices, in both hardware and software.
As everybody actively following this thread from the beginning knows, we started working on network solutions at around the start of 2020. It took 3 years to design and launch the switch, network card and router. Why this took so long is 1) the environment changing while we were designing, in a few occasions obsoleting certain design choices in about the same timeframe it takes to design something, and 2) gaining knowledge while designing, which is actually how we gain the most knowledge by changing parameters and listening to their effect as, especially 3 years ago, there was very little common knowledge already available on how network behaviour affects sound quality now, and perhaps even more important, how it may in the future. We went through 4 different switch designs for example, and then I'm not counting all the modifications we made to those designs before deciding to enter a redesign stage.
I deliberately mentioned the future in the previous paragraph as like the environment changed over the past years, I'm pretty certain it will change even more over the next few years. In a competitive electronic world, both in hardware and software, it may become a very realistic scenario where we are streaming 8K video and uncompressed high res audio, perhaps we'll even have 10Gb home networks, and of course even faster Wi-Fi. Hence we have tested our current designs in much more heavily utilised 10Gb networks running multiple heavy tasks, a stress test of sorts. This, perhaps unsurprisingly changes SQ performance again which we have hardened our currently launching solutions for as these are harder to upgrade and evolve in hardware with user installable cards, unlike the Extreme.
Circling back to the specifically mentioned Aurender N30SA, first of all, I actually admire Aurender for being a consistent performer, you could even call them pioneers in audio streaming. They early on developed their own music playback app to run on self designed dedicated hardware, with solid performance, basically they did everything right from the get go. This hardware platform, with (relatively) low processing power does not allow it to run Roon, however, though obviously lacking all the functionality Roon offers, in recent years this has actually been a strength from a SQ POV, not needing nearly as many adjustments to compensate for a changing (arguably evolving) software environment. This however has not been immune to the evolution of home networking like all other solutions, hence the launch of the N30SA which does a good job at addressing these changes. They "beat us to it" in this department, it simply us more time to launch our solution.
With the router we are launching our first, unique, designed from the ground up motherboard, with our perspective on how this should be designed, highly efficient, using Gallium Nitrate transistor based switching regulators at very high frequencies from which we get superior, lower noise, performance over "old fashioned" linear regulators powering digital circuits. Do note that this took considerable effort, we're now at the 8th incarnation of our GaN regulators and it took more then 4 years to reach these performance levels. Furthermore this motherboard integrates a "mil spec" hardened CPU, system memory, flash storage for the operating system, 2 M.2 PCIe slots, USB, and several configurable network links with recently added integrated Wi-Fi, while consuming only around 6 watts in total.
This motherboard, combined with our own software playback software, XDMS, could serve as a direct competitor to integrated devices, with a low support intensity, like Aurender. However if we are actually going to launch a product like this is very unsure for several reasons. First of all, this would be a streamer model which will just have to be replaced, like an Aurender or similar type of device, when technology moves forward. Of course software can be upgraded as it evolves, which it will, but you cannot supply user installable hardware upgrades while our current R&D drive is largely focused on that, and we have quite a few of these under development right now which would then be out of reach for customers having purchased this product, unless we subsequently keep launching new models. Although this could generate a lucrative revenue stream, this is in direct conflict on what we are trying to achieve with an evolving platform, attempting to minimize our customers expenditure for staying current. Nevertheless, perhaps there's a future for this kind of application, and we do now have the technology to do so. Should the router become obsolete for whatever reason, it could have a second life being a good streamer with just a software update. Anyway, for the foreseeable near future, we'll most likely just leave this market segment to others.
There is a reason I'm mentioning this bit, being the second reason, which is with everything we've learned over the years, with everything we tried, including designing this motherboard, we just cannot get the same comparable performance with low CPU power solutions as we can from high processing power CPUs, assuming comparable external environmental influences. That is not saying there's no place for them, but it's just not on our current path. Even when used as an endpoint, as a dual server setup, CPU power still matters, even on an endpoint.
Now there are several benefits to a
dual server setup, and it has spawned another R&D effort on our part which I'll elaborate on in a follow up post.
