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

I used to be a big doubter that there could possibly be any difference that you are actually hearing rather than just perceiving due to expectations, subconscious or otherwise. I now am convinced due to first hand experience that things can get weird and go wrong in all sorts of unexpected ways. Technically nothing should be perceptually changing, but that assumes everything is working correctly. Some faults are so obvious nobody would deny it if they heard them. For instance, if I use Apple Music to stream from their service, every song has a short stutter near the start of the song. This is a well known issue that Apple doesn’t seem to be able to fix, and it isn’t the same for everybody. Somewhere there’s a complex wrench in the works. Something ain’t right, and it makes you wonder what else ain’t right that might be more subtle and intermittent, thus not always showing up on single measurement tests. If I rip my own discs and don’t stream, it never stutters. Other streaming services never stutter.

I’ve got a device that runs optical audio in to my computer via a USB converter. It is unstable, and sometimes it does really weird things. If it gets bad enough, the music just stops.

On my Blu-Ray player I sometimes get audible and visible static. Just touching the HDMI cable will usually stop it for a while. Bits is bits, but they don’t always show up at the right time.

So, maybe you’re not hearing much because there wasn’t much going wrong in your particular system in the first place. I’m assuming here that in many lower end systems, like mine, many subtle things might be going wrong that happen to much less of a degree on better systems like yours, especially if you happen to have good synergy between your components, even if they aren’t at the very highest class levels available.

@mdalton 

but if it’s a decent streamer, your DAC dominates what you hear.

I'm generally in agreement with you here, at least in terms of what I'm capable of caring about, unless the streamer is glitching. All software can be glitchy. I know I can hear it. When it gets really bad my computer freezes up and I have to reboot. I just wonder if it sometimes gets just slightly bad, and I think I'm just not in the mood to listen that night when it's really my system sounding bad.

@dogearedaudio

MPD, Minimserver, BubbleUPnP, Squeezelite, etc. All of these offer somewhat different sonic presentations.

@audioman58

tip if you want your digital to sound better still Tubulus Audio cables ,they beat cables much more $$ and only true I2S -Audio cable the AQ dragon audio only exception ,and the less expensive Tubulus Concentus is much better. And Ethernet cables very good ,always put your best Ethernet cable at your end point- Streamer

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.

Maybe that’s the point here, most of the stuff that’s available isn’t actually working correctly, or even intended to sound transparent? I could believe that but I’d like to see some evidence beyond people just noticing they like one better than the other. Maybe some people like it better when they are not working correctly, or at least not trying to be transparent. How can we tell? It’s odd that these same cables can deliver huge amounts of software at blistering speed, and all of it seems to be 100 percent glitch free. The error correction works for other things. Why not audio?

I’ve seen people do experiments where they run a digital signal through a dac, and then back into an adc, and then back in to the dac over and over until obvious audible differences are apparent. If the server is doing that, running a loop to serve the data, retrieve it, and then run what’s been retrieved back through the server (and cables) should have an amplifying effect over time. Of course we could just check the first time through to see if any bits had changed. If they haven’t, then there’s no point in continuing the loop.

What timing smear? Is it correcting something that happened in the recording studio? Phase issues with the mic. and other processing done during the recording? That’s where the low hanging fruit would be. Orders of magnitude more time smear than anything happening electronically during playback - at least until your speaker and its crossover network get ahold of the signal. At that point, unless you’re fully active with FIR filters phase correcting the drivers, it’s time smear O-Rama! Better have your ears at just the right height for all the drivers to time align, or have a head tracker that adjusts the timing of the drivers as you move. Even with all that working as best it can, you'll still have vastly more time smear occurring in the speaker trying to get the sound to your ears than with the digital cables will cause. And the room, if it's not anechoic, is going to blow all of that out of the water. But we'll assume that it's only first arrival to the ear that counts for now.

With all that said, I'm not claiming the cables can't sound good. I just need further clarification on why time smear would normally be an issue, and what evidence is there for that as a typical cable issue.

I'm using HDMI out of my computer into a receiver. I'm not sure what brand it is. Some very basic HDMI cable that costs about 12 bucks and can handle 8 channels of audio at 24/96 along with a 4k HDR video signal. So I'm definitely not saying my system is one that might reveal issues with cables and streamers, although it definitely does concerning the HDMI cable from the Blu-Ray player to the TV, whe I can see and hear static sometimes when the cable needs to be jiggled to make it go away.

On better systems, where cable time smear might cause easily audible issues, what time scale are we talking about? Millionths of a second? Is this smear that would cause data errors? From what I've seen in testing it's not hard for a cable to get up to a million hertz while still maintaining excellent phase and amplitude fidelity. 

@dogearedaudio

I mean, "grooves are grooves," right? Is one person’s system "broken" because it sounds different from another person’s system?

It’s a very different situation because the groove on a record hasn’t been transferred to an abstract definition during the recording process, while the digital file has been. Because of this, each vinyl record is a one-off, with it’s own set of unique sonic characteristics that weren’t intentional, but are indelibly intermixed with the intended signal. No mechanism was used to keep them separated, and no mechanism can tease them back apart.

Each copy of a digital file, assuming no errors crept in, is identical, because it’s a definition, not a direct analog representation. The server’s job is just to serve up that definition, and there’s no excuse for a server doing anything other than that because the engineering has been solved. Any decent cable will allow a bit perfect information transfer to the DAC, as has been shown by capturing the bits into a file on the other end of the toslink, usb, or whatever, and then comparing the sent file to the received file.

Jitter artifacts are typically below the audible threshold of mere mortal humans even on $8 dacs running a synchronous signal through a cheap digital coax. 

I think it was on this forum that I read about a guy who was playing test tone LPs, and noticed that something as simple as a sine wave sounded better from LP than from CD. When he analyzed the waveform, the CD produced what looked like a nearly perfect sine wave. The LP produced something only vaguely similar to a sine wave. It added a lot of other stuff, which is why he liked it better. Pure sine tones aren’t pleasant sounding if you ask me.

So no, there’s no equivalence between saying ’bits is bits’ and ’grooves is grooves.’ Because bits is bits when comparing two copies of the same recording, while grooves ain’t. They’re just very similar. And different cartridge designs are going to result in much larger output differences than different dac designs, assuming the dacs are all reasonably competent, and the vast majority are. From what I’ve seen, the most suspect ones are the most expensive and least expensive.

This is why I would say that if server software is producing a different sound, something somewhere along the data chain is not working to specification, or there is some intentional DSP going on to add some kind of effect to make the server sound different than it would if it delivered an accurate file to the DAC.

@mdalton 

Sounds very plausible to me. I've read other accounts of EMI creating audible sound effects in susceptible DACs. 

@dogearedaudio 

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.

The server just has to get the correct series of zeros and ones fed into the DAC with timing that's accurate enough to keep jitter induced noise and distortion to a minimum, assuming it's a synchronous connection. If it's asynchronous then it just needs to get the ones and zeros into the DACs buffer in a reasonably timely manner. A lot of server software can do that, with jitter noise well under 100dB below the signal. There may be something going on that's causing it to not deliver the correct stream of ones and zeros at least some of the time, or perhaps to be accidentally sending analog signal along with the stream of ones and zeros that somehow ends up causing noise in the analog output stage of the DAC, or causes the DAC to otherwise not function correctly. So I know there are sonic differences sometimes. I've experienced it. What I'm not getting a good explanation for is how the server software can possibly make the sound coming out of the DAC sound  different if the system is actually working to spec., and not having a problem that could be readily identified with a little analysis. At least the effect could be measured on the DAC output even if the root cause might be hard to track down.

 It's much harder to bring two different cartridge designs in to uniformity of output from the same groove. It's an electro-mechanical device, and making tiny mechanical things with various connected parts vibrate identically enough to be undistinguishable in sound output is more trying. I have seen some double blind tests that showed there are some cartridge designs that people struggled to tell apart, while others were successfully distinguished. 

@dogearedaudio

I appreciate your engagement with me in this discussion. I’m not a software engineer either, and what little programming I ever did used audio APIs that were a black box that I just plugged in without understanding all the details. My basic understanding at the moment leaves just three options that I can comprehend to explain server software sounding different:

1. Not all the correct bits are reaching the DAC.

2. The bits are all correct but their timing is off enough to cause audible problems.

3. The bits are correct and the timing is also good, but other parasitic noise is coming along for the ride and somehow reaching the analog output stage of the DAC and creating audible noise and distortion there. Or it may be causing the converter stage to perform poorly even though it was getting a perfectly good and clear stream of data. The noise may be EMF transmitted from the server hardware directly to the dac regardless of cables, and of course the EMF is probably different with different music servers as they use different routines to process the data.

If none of those things are happening in any particular system, I’d propose it’s possible that the server software in such a case is working perfectly for all meaningful purposes, and there should be nothing audibly to discern between another server program or hardware bundle that is also working perfectly "to spec."

@dogearedaudio 

I think that all fits within my current understanding.