Hi,
Great forum. I’m a 53 year old computer scientist, who’s also done decades of sound engineering.
If the DAC is separate, the music server is simply transporting digital data, and not deliberately altering that data in any way, then it cannot make any difference whatsoever to the sound quality beyond not messing it up. The power supply doesn’t matter, the circuitry doesn’t matter, the cabling doesn’t matter (unless you’re running over 50m), oxidisation of components doesn’t matter, the shielding on the case doesn’t matter. If there’s some sort of problem with jitter then it’ll be screamingly obvious because jitter is never ‘subtle’.
Think about this: digital data has been around for decades now. Take a digital file that originated in 1980s, and has been copied from computer to computer, disk to disk, USB stick to USB stick, sent over WiFi and down a gazillions cables, losslessly compressed and decompressed… and its contents will be EXACTLY the same. The is a fact that can easily be validated by comparing the binary data in any copy with the original. This is a major feature and use case of digital encoding. We can also do run this comparison with streaming data but collating it, and comparing it to the original. Unless something has gone horribly wrong, then it will be totally and absolutely identical, no matter which transport streamed it.
The *only* way a file *could* be different is if lossy compression were allowed… but that’s up to the user. It’s the same with streaming, the data cannot change unless it undergoes lossy compression… so don’t use Bluetooth to stream music, and don’t encode music in MP3 etc.
The sound absolutely is altered by electronic considerations at each part of the analogue chain, and also in the digital to analogue conversion stage (more of this in a moment), but not within the digital realm itself, not unless it’s undergoing some sort of processing intended to colourise it.
Even the choice of speaker cable has INFINITELY more impact on the sound than digital transport does (meaning slightly less than zero). For avoidance of doubt, lossless digital transport that isn’t undergoing any sort of deliberate processing to alter its contents, has absolutely ZERO impact on sound.
The noisiest power supply in the world will not affect the fidelity of digital data, not unless it somehow reaches a point where the system’s operational integrity has been compromised. For example a computer system with a horrible power supply does not affect the reproduction of MS Excel documents in the slightest. The numbers in the spreadsheet are not impacted, and do not change because the power supply is ‘noisy’… not unless it’s so bad that the computer cannot even be booted.
Same goes for anti-vibration measures. Non-SSD hard drives certainly can be affected by vibration… should that happen then it will be hugely noticeable because the device will probably hand reset itself… but until that point it will faithfully reproduce the data with 100% accuracy.
Sometimes I read or watch hifi reviews, and start to doubt my own sanity. These people talk about digital comms as though those were a form of analogue, where components play a part in the sound. This simply isn’t true; data doesn’t have a sound in the same what that the postal worker’s handwriting is irrelevant to the contents of the mail he or she delivers.
The DAC is a different matter… but even here there’s a crucial piece of information that most people aren’t aware of: while no human has evolved who can hear beyond 24bit depth @48KHz, there is a physical issue with digital to analogue conversion that hasn’t yet been overcome by engineering. The long and short of it is that DACs function much better at higher sample rates, like 88.2K, 96KHz (and upwards into rapidly diminishing returns). So if you’re a practiced listener, who can discern a slight but noticeable difference between 96KHz audio and 48K audio, then that’s why. It’s not that your ears are picking up more frequencies or greater resolution; it’s just that DACs are not well optimised when running at 44.1KHz or 48KHz. Not even the really expensive ones. A cheap DAC running at 96K will likely work better than an expensive DAC running at 48K. This was an unforeseen flaw when the CD standard was devised as 16bit dept @44.1KHz. As an aside, only a child with exceptionally good hearing can distinguish the difference between a 48KHz sample rate vs a 44.1KHz sample rate, and only really practised listeners with good hearing can discern 24bit depth from 16bit depth. To all intents, 16bit depth @ 44.1KHz is at or beyond the limit of what *most* people can hear, due to the Nyquist rate. The advantage of using a higher sample rate is that the DAC will make a better job of the conversion.
When it comes to digital transport though, choose a transport based on how good it looks, how good its build quality is, how good its connectivity and functionality are… because the one thing it shouldn’t do is alter the sound quality. If a digital transport does affect the sound, independently of the DAC, then there are five possibilities:
- It’s actively and deliberately altering the underlying data as it streams that to the DAC, so as to colour the music. This is just sneaky software engineering, and not due to a lead encased power supply, anti-vibration measures, or cables made of gold.
- It’s communicating with the DAC at 88.2K or more.
- It’s outputting at a slightly higher volume (due to digital trim).
- There’s a short circuit between the headphones.
- The person listening to it is an ‘influencer’, and there’s an obvious reason as to why they are able to find the time to make those sorts of videos.
I’d be interested to hear any scientific explanations of how cables, power supplies, or shielding can magically alter the ‘sound’ encoded in digital data…