Audio PC

Fanless CPU heatsinks are good if you have a relatively low powered processor:

https://www.fullysilentpcs.com/product-category/cpu-coolers/?v=7516fd43adaa

These are actually HUGE heatsinks.  However, they are not good enough if you're running a 12-core CPU for gaming (like me).

I tried Fidelizer a long time ago and it actually did weird things to the audio on my desktop (in other words, it did not improve it).  It did however improve audio from my crappy laptop, so verdict is out.

I have found that running Windows with stock kernal timings seem to be the best.  In some cases, the windows WASAPI drivers sounded better than kernal streaming from JRiver. lol.  If I was doing a dedicated music server, then I would look at Euphony Stylus and doing a closed silent PC with a relatively lower power cpu.  However, I have specific needs for a general use computer that has to server as a miscellanous platform, gaming platform and music playback along with data redundancy for all my personal financial data (quickbooks, etc.).  So I had to pick very specific parts when building this.

So now I focus on how good the sound card is and how good the power supply is.  The Pink Faun S/PDIF with Farad power supply is just excellent!  I will eventually add on their custom OCXO clock, which should give me significant improvements.

NOTE ON PINK FAUN:  There appears to be a compatibility issue if you run Pink Faun card on a Linux platform using Intel CPU (sound breakup problems).  This is a weird problem with the chipset used in the Pink Faun card.  Pink Faun engineers and Linux engineers have worked on this issue for many years but nobody can figure out the root cause.  Intel CPU with Windows appears fine though.  There is no issue with AMD processor and general consensus is AMD actually produces better sound as a digital transport.

I don't mean this facetiously, but why not just run USB? Big buffers. High data rate, no losing sleep over competing processes, no losing sleep over jitter? If ya got an older DAC I get it, but not much without USB now. Ya running into a processor or somethin that needs SPDIF?

@snratio - I have heard of one (sorry I cannot recall the name of it) DAC that actually uses USB with checksum, but generally DACs don't store the stream, they process it in real time, and what doesn't get accurately read in by the DAC is lost.
This is the reason AFAIK why so much emphasis on the clean linear power, signal integrity and the least amount of work, or rather unnecessary handling and processing threads that are not essential for the playback of music help the system create less distortion and loss of data packets.

SPDIF is used with claims that because it's not an electrical conductor that it doesn't transmit electrical noise. However, from many reports (not done very recently I must confess) indicate that USB as a means of data transfer to a DAC more often than not, sounds better. Then of course the latest and greatest is TCP/IP Ethernet as the means of data transfer, or I2S over HDMI.
Honestly, I can't keep up, there's so many manufacturers out there, it's like trying to keep up with who makes the best sounding speakers in certain price brackets - does anyone actually know, and then how long until that knowledge is obsolete?

I should add, and I'm sure if I say this incorrectly someone will chime in with why, and how and I'll learn something I didn't know...

In order for a DAC to store the bit stream in a buffer, the protocols required to do so have to be processed by a computer. To do a checksum for example, registers (or data storage) need to set aside to evaluate and store information about the packets in order to determine whether or not the packets of data arrived correctly. So the DAC requires an onboard computer to store the data into RAM, evaluate the incoming bit stream to ensure it's correctly read in, or request the data rate is slowed down and data resent.

The bitstream is evaluated from raising and lowering voltage, or light intensity from one device, and measured on the receiving device, if it's either not sent with a very clean signal, the measuring of that signal may be wrong, and if there's no computer to evaluate the data packets it's discarded at best, or becomes corrupted.
We can read written language and within the contxt of the sentnce detrmine the meening of what was written, a DAC cannot.

The least dropouts of data packets, and the less electrical noise, a more accurate representation of the recording, created by transforming a series of voltage changes or light pulses into an alternating electrical signal that can be transformed by a electroacoustic transducer (speaker) into sound waves, the better. (Read that aloud three times without taking a breath, I'll have you put in a circus)

Memorial Day - Lest we Forget.