Discerning a difference between streamers is difficult...only me or common for all?

I have no wish to add to the OP’s confusion, and I agree with many of the commenters who feel that past a certain level, streamers can vary in whatever added value they bring to a system. But we also shouldn’t overlook the differences in software. All these devices require some sort of operating system for connectivity, and most, if not all of them, are based on various versions of Linux. And within Linux there are a number of different playback and connectivity formats, among threm MPD, Minimserver, BubbleUPnP, Squeezelite, etc. All of these offer somewhat different sonic presentations. One reason I like my Pi2AES so much is that I can choose my own OS and player software. While a lot of these are free, like Ropiee, Moode or Volumio, I’m happy to pay extra for a system like Gentooplayer, which allows a wide choice of different "players" and a lot of tweaks and extras, like the Diretta ethernet protocol. Now there’s an extra layer or two of "tuneability." This isn’t for everyone, obviously, but it is worth noting that the software provided with a streamer can have a significant impact on the sound. It might consist of a basic DLNA server, or it might incorporate something more sophisticated.

@asctim 

"If the server software or the cables are delivering the digital file information to the dac in a different enough way to make the DAC sound different, then some or all of them aren’t working correctly. Or the DAC isn’t working correctly. Or none of them are working correctly. Or there’s some incompatibility. Or they’re creating different sounds on purpose through processing."

Well, that's the million-dollar question, isn't it? ;-)  In the analog realm, audiophiles spend tens of thousands of dollars on turntables, cartridges, tonearms, preamps, etc etc, all seeking "ideal" sound, when all we really should be worried about is hearing the darn record, right?  I mean, "grooves are grooves," right?  Is one person's system "broken" because it sounds different from another person's system?  Of course, in the analog realm it's generally accepted that various "macro" mechanisms will affect the transfer of sound waves from the grooves to the speakers.  When it comes to digital, there's a stubborn belief that "bits are bits."  But a lot of people would argue that the "micro" mechanisms involved in *how* those bits get turned into music make just as much of a difference.  An OS that uses a different kernel or tweaks the operation of the CPU or system clocking devices appears to have an effect on the end result.  Is that "processing," or just an attempt to improve the quality of not only the digital-to-analog conversion process, but the way the information is shuttled from one device to another?  Is it "processing" to upsample or change filter parameters to bypass a DAC's internal filters?  No one seems to dispute that in purely analog systems, there are many ways to skin a cat, so to speak. ;-)  When it comes to the digital world and it's less-readily observable processes, suddenly "bits are bits" and if one algorithm sounds different from another, it's "broken." ;-)

@asctim

I’m not talking about cables, that’s a rather different topic. I’m talking about the differences between different servers and different software. One digital copy may exactly resemble another, but you still have to get that digital file from one place to another and then convert it to a convincing analog audio presentation. Some people have latched on to SD players, which eliminate all the server protocols like DLNA and UPnP, and pretty much do away with an intervening operating system. But most digital audiophiles are stuck with a motherboard, an operating system and various forms of interconnectivity between the server and the DAC--USB, coax, I2S etc.

As for "grooves are grooves," well, take the same LP and play it on two different turntables. Heck, two different catridges! All else being the same, that LP will sound completely different.

@asctim

I’m not a software engineer so I can’t satisfy your need for an explanation. But I think if you actually look into different playback and operating systems, you will find that there are differences both in the way the operating systems control the clocks and CPUs, but also in the way the digital information is handled and moved about. Audirvana, for example, converts the file into PCM and preloads it into ram before playback. HQPlayer employs a complex variety of filters. Any serious digital audiophile will tell you that Roon sounds very different from Minimserver, even sans any direct DSP intervention. Even the more basic players, like Gmediarenderer, APlayer, Squeezelite, MPD--they all present different signatures, at least in my system. Buffers, caches, protocols--I couldn’t technically say what mechanisms are at work. But the differences are deliberate in that the designers of these various players are seeking to deliver digital music in the most convincing way. You say "working to spec." But what does that mean? Honestly, in terms of digital playback we’re waaaay past the "bits are bits" stage and well into how the subtleties of microprocesses contribute to converting bits into music.

@asctim

Okay, let’s take a very common example. Many, many digital audiophiles remark that Squeezelite, MPD and Roon all sound very different. When it comes to MPD and Squeezelite, I have the option to use either on my streamer. When adjusting settings for MPD, I can chose to bypass Linux’s internal ALSA mixer, by setting the volume control to "none." Squeezelite does not offer this option, only to set the volume at "100%". So there’s a significant difference right there. A digital algorithm is being bypassed in one player, but is still in the playback chain in another player. Both players are working to spec, that is, they’re sending "bit-perfect" information to the DAC. But one has an intervening step that the other doesn’t. Does that affect the sound quality?

As for Roon, their streaming protocols are proprietary and are obviously designed to produce the Roon "sound," which is one of their selling points.

None of this has anything to do with the source of the music file, but that’s another issue. How does the player software retrieve the music file? How much does it buffer? How does it manage the clocking systems, CPU speed, etc etc. It’s all presumably "bit-perfect," but will it sound the same?