Sound quality differences in streamers

@sns 

I’m glad you are adding your voice to those who know from experience that not only is a PC more than adequate as a streamer, it is often superior. Also (and I didn’t even think of that before) it is true that:

Cool thing with custom builds is you get to hear what individual components add to the mix, everything matters.

Very true! I must however respectfully point out that there is no such thing as a "Windows motherboard". Now, Windows has been the dominant OS for 40 years, and it follows that motherboard design certainly accommodates Windows, but motherboards are by definition agnostic.

Although Taiko successfully runs Windows in their Extreme, Linux is probably better suited to a DIY build due to its almost infinite customizability. Right out the box, Windows does not want you to customize it and it will resist. 

One last word, "Enterprise" motherboards (that is, server boards) are preferred because of their relative lack of features and absence of related componentry, thereby eliminating potential noise sources; a few of those boards also natively support SFP (the main fiber Ethernet standard), which is desirable as it provides 100% galvanic isolation between the PC and the next component in the chain, thereby ensuring that any noise that happens in the computer, stays in the computer.

Also, @sns kudos for tackling the tricky PSU problem once and for all, although you sure didn't go the budget route 

Oh and +1 for the Audiophilestyle forum

@mdalton     In your attempt to support your position, you stated that you were  guessing at an imaginary % based on a discussion thread that was in no way a credible statistical model for evaluation.  That’s a non-sequitur to say the least.

The elements for a good statistical poll include:

• A set of questions to ask (a questionnaire)¹.
• A sample of people to interview¹.
• A method of connecting the questions and respondents¹.
• Survey error reduction and positive social exchange².
• Short and clear questions, avoiding leading questions³.
• A mix of closed-ended and open-ended questions to gather diverse data³.

- - - - 

I was trying edit my previous post when I business call came in.  I deleted it to give full attention to my client.

I recently acquired a custom built server by Pareto Audio.  As alluded to above, the design goal with the Linux core is to remove all non-audio related processes, reduce the number of interrupts, speed up processing, eliminate non-essential network interfaces, and isolate with fiber optics throughout the entire network.  This has been a big step up and sounds about as good as can be done with a Sonore Signature Rendu Deluxe endpoint with USB output.  As good as this sounds, it is further improved with a SOtM USBultra USB conditioner.   To test the assumption that I've reached the limits of a USB DAC input,  the next move is to find a FO-to-I2S endpoint into an good I2S DAC TBD--perhaps the top Audio GD or PS Audio's forthcoming top model.