Well, I’d take cash, but as a server it’s just not good for anything. It has wrong specs for a Roon server, and for a normal datacenter server (which this thing is based on) you can get a better specced one from Dell at half the price.
Use some tape, not to get dripping all over the imputs/outputs
1 Like
Given the exact same hardware (same OS), using the same music source but running different server and player software sounds very different to me on my gear, it does not surprise me at all that OS and hardware can be tuned to improve SQ of Roon. Roon is resourse hungry. Higher performance compute creates more noise. More noise requires more filtering. I’m sure there is a sweet spot. I use Antipodes gear now but previously other server and player hardware. I don’t regret the investment. We should keep an open mind
1 Like
For the sake of fun, always the USB of Windows sounds batter due to how Linux handel’s music streaming (Mac is Linux at the end). Windows ASIO drivers are speak directly to the USB port with minimum latency.
1 Like
No it’s not, not at all, it has nothing to do with Linux whatsoever
2 Likes
If interested, please read how ASIO drivers operate in comparison to how MAC built in drivers operates.
Mac still isn’t Linux, and apart from DSD limitations, Mac delivers exact same bits to the DAC.
I assume you’re comparing ASIO with Mac CoreAudio Exclusive Mode. Right?
IIUC, Windows ASIO, Mac CoreAudio Exclusive Mode, and Linux ALSA all provide bit-perfect transfer of audio data over USB.
Yes true. I did intensive tests and in every test I find ASIO is much faster and open in comparison to Core Audio on MAC. Linux ALSA provide bit perfect but it seems internally is being handeleled as floating point not integer which impacts the sound.
I do not have the time to analyze the actual science but to me ASIO always sounds batter (DAC is Chord Dave with Mscaler).
The same code runs on all three platforms. For a given bit-stream within Roon, that exact stream is written to USB without any interference or processing by the OS and it is reconstituted perfectly by any modern USB DAC.
I won’t try to dissuade you that you heard something different, but the data is the same from the DACs perspective.
1 Like
I just read quickly that MMC core audio cannot output 32bit integer in comparison to Windows ASIO. I do use something different. I use HQPlayer to output 32bit to increase the dynamic range and keep the sampling the same rate. On windows ASIO it sounds amazing on MAC mini M1, all combinations give less capable and less clear sound. Please read about Core Audio support to output native 32 bit integer.
So, it depends on use case.
I was joking, but looks there can be a discussion on this. I cannot have any opinions as I did not use windows for audio.
That is PCM, I might be wrong. Again not tried HQP with PCM
I love Handel’s music, windows Mac or linux 
Pls do not mind my bad jokes. Thanks for the discussion, there are thinks that I did not know
1 Like
Not the same code runs on all platforms. The platform APIs are very different.
For example, it is so wrong to say same code runs on Android and iOS. Platforms are completely different (APIs) and the codes are completely different. Yes designed and developed to give same functionalities and experience. Software is my profession.
It was a simplification. Of course the application may have some differences around the edges, depending on how much of the platform .NET hides/exposes. And .NET will certainly have differences to interface to the host OS. And the OS is obviously different and has differing abstractions.
But, if Roon writes a stream of bits at a supported sample/bitrate, the DAC will see the exact same bit stream after depackitization regardless of what the OS does to get it there. That is the whole point of the signal path. If the DAC does not receive what Roon’s signal path claims that it sends, then why bother with it?
Technically this is a design objective but again cannot be guaranteed. Service providers in most cases assure consumers that their software performs the same across platforms, but that is sometimes impossible to be achieved.
Same bits to the DAC yes could be but again not garanteed because PCM has no error bits with the signal like computer networks. PCM is like analog as no guarantee what comes to the DAC after the signal leaves the port.
I always appreciate scientific method for looking at matters. Unfortunately this domain is full with snake oil companies and uneducated consumers in many cases.
Regards.
Imad
1 Like
Yeah, right… but the whole point of the sweet torments of audiophilia nervosa consists of the claim that it doesn’t matter that the DAC sees the same bit stream, if you forget to take into account the means by which the stream is delivered… you have to consider if the code uses floating point or integer arithmetic to implement the relevant bits of algorithm, if the bits are delivered over 99.99% OFC or silver strands, if the moon is waxing or waning and, most importantly, if your system has taken into account the most recent recommendations by the usual suspects…audiophile press and assorted pundits. So no, my profane bits are not the same as the audiophile’s anointed bits.
3 Likes
One can rather trivially (as far as these things go) put a digital receiver at the other end instead of the DAC, which re-captures the bits, and then they can be compared to the sent samples. It would be very unusual if they differed. It’s quite common for on-board protocols not to do error checking because it is assumed that the hardware works and there would be many issues with error checking everything. And indeed computers typically do work.
2 Likes
Yes we need to take life easy and focus on what is more significant but PCM is electrical pulses at certain frequencies. This is why SPDIF cables alter the sound and any ground noise impacts the sound also. The best option is to remove this connection. A DAC should have a built in streamer like some brands do.
To share some experiences, most USB/spidif issues are based on different ground levels. Connect a laptop on battery and use it as a ference. So, at the end I agree, the bits are the same, but the communication medium created ground noise that could alter the PCM signal.
1 Like