I have currently 2 licenses for Roon (one is usually located at my office, and one at home - but because of these COVID-19 times; I have both at home). I had someone start playback on one or the other (random) on the same track on Roon 1.7 and Roon 1.8 against the same Roon Ready endpoint. I could in no way tell the difference. During this exercise I could not know which Roon core were currently active (I could not see the screen controlling them).
Edit:
Tested with Trinnov Altitude32, Trinnov Amplitude8 and B&W 802 D3 + Matrix Audio Element X with Audeze LCD-4
I disagree.
Fair to say that these tests are not always rigorous hence can be misleading: for example, plugging in new cables cleans/refreshes the electrical contacts and that alone can make a change, if people do not regularly clean the traces of oxydation that appears. Metal oxydes are far more resistive than metal itself and can have unwanted dielectric properties.
But the sound of cables, as well as of passive components, which of course do not entirely conform to their first order mathematical model of being a pure resistor, capacitor or inductance, is sufficiently well documented. Discerning measurements are not obvious, because the audio band effects are subtle.
And our audition is also subtle. Behind our ears you find a powerful neural network. All has been designed by million of years of evolution, where sine waves were not common place, but hearing weak signals like those of a future predator or prey, were a survival thing. And we do train our ears to discriminate subtleties that matter to us.
It’s very simple. When you know there is a big release coming like 1.7 to 1.8 record the output of your DAC while you are still on 1.7. Once updated to 1.8 do the same process. Load both of the files in to a DAW. Invert the phase of one of the tracks and play them both at the same time.
Silence means they are the same. Sound means you are listening to the change in sound quality.
The change notes mentioned an optimising the efficiency of the DSP. Not that there was a change to the core sound engine.
Put aside DSP and just look at “bit-perfect” which is viewed as a “gold standard” for digital audio transport. How can Roon announce an improvement? They’ve been selling a “bit-perfect” player for a few releases. To suggest this has been improved in 1.8 is to suggest users of previous versions have been “sold a pup”.
@Adam_Arczynski simple objective test confirms this view. Unconscious/expectation bias can be a powerful thing.
Well, that’s not possible technically with bit-perfect transport - which has been proven - because this background noise must be part of the encoded signal.
But if you are inferring that the noise is being generated by the conversion process, then again we’re looking at something being defective in your DAC.
That this tiny a difference in the lowest octave makes or brakes your listening experience lets me think that a very large amount of the recorded music catalogue won’t ever make into your listening room, since that part of the spectrum is virtually MIA.
I’d really like to see a properly performed measurement of your system in-room…
Hello,
Not this time. For me it was clear since the first seconds. Emotion was gone but I agree it is not something you can find on many systems. Using the Antipodes music server squeeze player allows to recover emotions so it is not psychological.
Regards,
It’s not so complicated if you lay out the permutations.
There are 2 noted changes that might affect SQ in this release:
change to audio pipe cpu/memory utilization or any other electric/noise impacting changes
changes to DSP dithering.
This effectively generates 4 different use-cases in order of overall potential SQ change:
A. core is well isolated - via wifi/fibre/clean ethernet via endpoint (built in to DAC or standalone) - and you do not use DSP i.e. you run “bit-perfect”
→ you shouldn’t really perceive any SQ changes
B. core is not isolated from DAC and you do not use DSP i.e. “bit-perfect”
→ you might perceive minor SQ difference due to electrical noise leakage from core server
C. core is isolated from DAC and you do use DSP
→ you might perceive minor SQ differences due to DSP output …
D. core is not isolated and you also use DSP
→ you might perceive more SQ differences due to BOTH Noise and DSP
I basically agree with you, just a little comment about your A option:
A friend of mine is running an endpoint (SotM Ultra + Farad LPS) with a PC as dedicated Roon core. He hears difference when changing frequency of the CPU in the Roon core PC. So a normal ethernet network is not enough for total isolation. He does have a high-end system though.
Fiber on the other hand tends to completely remove the impact of up-streams equipment, as long as they work as they should (i.e. stable transmissions etc).
Yes there is a chance for noise to travel across ethernet. I in fact use an OpticalRendu with FMC to isolate the audio system as much as possible. But I am not sure how much of that can be ascribed to a core more than 1 hop away. If it was direct ethernet connection then that’s not really a ‘network’ of course. One would think that routers/switches in between would contribute as much if not more noise to the end signal but I could be wrong there.
Nonetheless, I will edit my post.
Yes, forgot to mention that but he has his PC directly connected to the streamer (which according to him sounds better than to hop via a switch). Btw, I also use fiber (ethernet → MFC → fiber → opticalModule → streamer). I can play games on my Roon core computer without it affecting the sound quality
I’m pretty sure the sound changes are related to the smaller waveform indicator. It’s reasonable less energy is going into the waveform and more to the DAC. On a highly resolving system you can even hear the difference tilting the iPad from Landscape to Portrait as the waveform gets smaller.
Scott, I agree 100%. I figured out in addition the following, once you hold your iPad in an exact 37 degree angle you can hear the leading bit of each note dropping into the dac, but I have to admit this works only when you have switched the dark mode on!
Read the introduction from pros of digital measurements.
Viewing the Analog Origins of Digital Signals
What do all these characteristics have in common? They are classic analog phenomena. To solve signal integrity problems, digital designers need to step into the analog domain. And to take that step, they need tools that can show them how digital and analog signals interact.
Digital errors often have their roots in analog signal integrity problems. To track down the cause of the digital fault, it’s often necessary to turn to an oscilloscope, which can display waveform details, edges and noise; can detect and display transients; and can help you precisely measure timing relationships such as setup and hold times. Modern oscillo- scopes can help to simplify the troubleshooting process by triggering on specific patterns in serial data streams and displaying the analog signal that corresponds in time with a specified event.
…May be he guys that claim there is nothing that can be measured or detected, just did not invest in the right equipment but fail to acknowledge it.
