This is irrelevant to Roon. Roon, unless one is using DSP, is passing the signal unadulterated to one’s DAC. The DAC is doing the processing, where Reichert’s comments have merit.
1 Like
If I remember rightly there have been comments on here of people using DSP having to alter it again following 1.8.
As said previously, if it’s unadulterated it shouldn’t sound any different.
There are certainly accounts of people believing they had to.
Now, this is the interesting bit. From an objective point of view I can’t come up with any explanation as to why 1.8 would sound different to 1.7, so I believe it won’t sound different, and - lo and behold - I don’t perceive any difference.
Does this mean that everybody will think it sounds the same? No, clearly not.
If I was in the ‘bits are bits, but …’ camp - where all sorts of weird and wonderful things can cause differences in SQ - I might be more inclined to a) believe that a version change can cause a change in SQ, b) be very much on the look out for it, and c) end up perceiving a difference because I’m now listening very, very intently for any possible change. Expectation/cognitive bias - it’s by far the simplest and most plausible explanation, but some people seem unwilling to accept that their own senses are demonstrably fallible, despite reams of data, practice and theory to the contrary.
1 Like
Which component is “degraded” and could that be the issue, not Roon?
What did you use as a reference for the comparison, a running version of 1.7 or something else?
If Roon were to use listening tests:
- should the results of those tests persuade them that they shouldn’t implement a bit-perfect transport?
- what file formats (MQA?), DAC, amp, speakers, etc. should they use to set the standard?
- what steps can they take to fix bit-perfect?
The first of those is likely to prove particularly controversial. BTW I’m NOT saying that the Roon folk don’t listen at all, but the questions above are tough ones. If you think the bit-perfect needs fixing, there’s DSP after that its all about the DAC and the downstream kit.
1 Like
No sound change here I’ve noticed, on either of my setups, and I’m one of those ‘audiofools’ who can hear the difference between switches, cables, fiber, power supplies, etc.
1 Like
For me the 1.8 was much better sounding than the previous releases.
But after the 1.8 778 build I am back to the old sound.
I am not sure about why the sound has changed to the worse but Roon is the only explanation I have right now because nothing has changed in my system.
I will try to find another way to stream so I can compare SQ.
1 Like
Has anyone tried replacing their “tired” ethernet cables with fresh ones?
So, I did. And I really like this track. It’s got a very nice jazz bop feel to it, not at all the pop/metal label Qobuz/Roon give it. But I gotta tell you, while this track has some nice mid-bass, with some good texture, this is not a bass showpiece. Not trying to pick a fight, but there is a lot of stuff out there with much deeper, textured, articulate bass. (FWIW, my speakers are Legacy Valor, which generally have no trouble playing bass if it’s there.)
As long as you enjoyed the music, that’s all that matters. My deep bass is curtailed due to neighbours but I like this kind of natural bass that you would enjoy if the instruments were in the room.
I hoped we were past the discussion that bits are bits.
There are no zeros and ones running through your cable, it’s modulated current (square waves) that represents digital information. But along with that current there may be RFI and EMI distorting the signal, thus degrading SQ as error correction chimes in when bits are detected wrong or in the wrong time frame. Minimizing electrical noise helps the DAC doing its job right.
Roon reported that they optimized memory management (less traffic, less CPU usage) in release 1.8, so theoretically this would help improving SQ.
However the better your setup is optimized regarding avoiding electrical noise the less you will hear any differences in Roon software changes.
1 Like
This simply isn’t correct. Sound quality does not degrade if bits are “detected wrong” or in the “wrong time frame”. To take the latter first: TCP is asynchronous, and the data is buffered at the DAC, so ‘wrong time frame’ isn’t a thing. Second, error correction in TCP transmissions is extremely robust. If the wrong data turns up, or fails to turn up, it’s resent. None of this has anything to do with sound quality.
How? Which theory or theories are you referring to? If the data doesn’t turn up at all, then sure, this will have a marked effect on sound quality, but how does it have an incremental effect? I’m assuming we’re talking about the usual stuff here - deeper blacks, wider soundstage, etc. How can less CPU usage improve sound quality?
8 Likes
You are treating an digital signal as if it were analog, A signal represents a 0 or 1 and can deviate within certain limits and not change what is being represented.
If a signal is degraded so bad as to change a 0 representation to a 1 (or vice versa), then some part of the system is not up to spec. Being up to spec is all that is needed, more than that is snake oil.
4 Likes
41.5Hz - the lowest note on a four string bass. Not much goes lower apart from electronic music. although there are two cathedral organs with 64 ft pipes that go down to 8Hz! Or try this for even lower…
The 64ft and 128ft pedal organ stop experiment - YouTube
Can you tell I’m playing with a subwoofer at the moment?!
2 Likes
I used a 5 Hz signal to break in my par of ML Dynamo x1100 subs. I may not have heard it, but I sure felt it, enough that it felt like it should be audible . I would have used a pipe organ, but I didn’t have the space for one. But I love them when they are recorded well. I want my E. Power Biggs!! (or Virgil Fox)
That weariness cuts both ways, I’m in the other camp, but I feel the pain 
But it is theoretical and close to impossible over “to specification” commodity ethernet. I suspect you’d really have to work very hard to build a real-life core → ethernet → endpoint setup that demonstrably transmitted core noise to the endpoint via ethernet.
If the DAC’s not well designed/implemented, then it’s possible, USB is theoretically most vulnerable, but this is just accompanying electrical noise. It hasn’t polluted the binary “signal” any, so the DAC needs to reject accompanying noise at the input. When the signal’s digital, this isn’t the most difficult engineering issue. DAC designers/builders who can’t address this will find trickier noise problems further down the chain. Put slightly bluntly; it’s a noisy world electrical wise, don’t buy a Goldilocks DAC.
If you really think this is an issue, use an optical cable to the DAC and bask in isolation.
2 Likes
Not trying to quibble, because I generally agree, but several keys on a piano go lower than 41.5Hz. And there are ample recordings with decent bass; as you note, organ and electronics. Like I said, I enjoyed the track he posted, and it’s got some good electric bass-work; I just don’t think it will really show off a system or be much of a test.
There is nothing simple in digital audio. And there is more than TCP. Streaming services more often than not use other protocols than TCP.
F.i QUIC (a Google-modified version of UDP) or HLS. And as you know Roon uses RAAT. But even if we talk about Spotify, which uses TCP, this protocol effects would only go up to the streamer but it’s the connection between the streamer and the DAC that counts. There are different interfaces available to connect your streamer to the DAC (USB / SPDIF / Optical / i2S / Ethernet) with different protocols (each with strengths and weaknesses regarding Jitter).
Some easy to understand video:
No! It is an analog signal with digital information inside.
This hair splitting doesn’t change my original point, but I’ll rephrase.
The information is being treated as though it were analog rather than digital.
The signal can deviate with certain limits without changing what is being represented.
2 Likes
What’s not easy to understand is the level of FUD he raises about the transmission of audio data. There’s nothing remarkable about the bit-rates of hi-res audio formats yet he talks about CD rates as though they’re cutting edge. The problems he describes should all be addressed by competent DAC design.
1 Like