Just read the way Magnus put it out some days ago in the same thread…
In other words, simplistic explanations not always sufficient - especially to track indirect causality.
…Obviously too simple to say that, combining what we know from Roon team and various comments, it is affected by:
- various sensitivity to jitter (e:g Ethernet > USB > SPDIF)
- various sensitivity to electric noise contamination typically from computer domain to fine audio domain via power lines or ground lines (Wif, optical > Ethernet, USB, SPDIF)
+(side remark: unfortunately neither Wifi or optical are perfect w/r to jitter or bandwidth) - depends directly on whether people use Roon DSP or not
- might indirectly depend on whether people use other DSP or not
- depends on level of subtlety of the audio part. For example, many rooms and system trade in balance between bass, medium and treble, by not managing to get read of (or keeping) a bump in the 60-80 Hz region. That bump alone is sufficient to mask many details both below and above.
These are the factors I would put on the list, please complete with other reasonable ones I did not think of.
If you have jitter in your HiFi from Roon, that’s a system issue… Design.
If you use DSP, just make adjustments.
1 Like
Simplistic argument again.
Jitter is a time misalignment. Time misalignments are a continuous variable, with physical causes - delays in signal propagation in components, circuits, wires, cables. However small (w/r to signal band) jitter tend to accumulate, rarely cancel out. As Tektronik rightly points out in the course intro about digital signals being some kind of physical analog signals in physical circuits, higher sampling of digital signal creates more jitter for the same physical configurations.
And as explained by others above in this thread, various transfer protocols (USB1, 2 or 3, Ethernet, etc…) address these issues differently.
The best strategy is of course to remain as asynchronous as can be. But in the end, the conversion itself must be synchronous anyway. Anything that can perturb this ultimate synchronicity can be labelled a jitter.
Again, who has never heard sound degradation upon increased computer load (on a sufficiently high quality system of course) ?
Jitter is known about for years and good hi fi should be designed to eliminate Jitter.
This is why I chose my Meridian system with active DSP speakers. So many of these problems just go away as the design keeps everything in the digital domain until the last moment when dedicated amp/driver designs deliver the sound.
Yes and no. Let’s say it considers the destination but not the path to reach it.
I like to understand why things work better or worse. In the analog and physical domain of setting up an audio system that is not a fully preset system, it is important at least as it helps to be able to progress rigourously in experimenting what works best - and selecting better components.
I think it makes sense to keep the same attitude towards “digital” audio (that comprises tons of very high frequency analog signal paths).
Known for years, yes. If it should be 100% solved for years, why are all these discussions about people hearing differences in things like imaging accuracy (directly linked to acute timing differences between left and right channels) ?
I don’t understand why people claim to hear a difference, I can only think it’s noise pollution in analog circuits which is a poor design and implementation issue.
Unfortunately no brand of hifi has a silver bullet for every component of an audio system otherwise the choice would be so easy… I heard in Scotland a high-end Meridian system in a shop with this setup. It sounded well: natural, detailed, ample and smooth, but not, by a fair distance, as timbrically accurate, dynamic or natural as more esoteric / more ambitious systems I had heard before, heard later and have later built, that might have other issues such as speaker size for example. For many reasons, systems are always compromises in one way or another.
So I believe it is better to avoid arguments of the type “I have equipment of this brand and you guys are messing around problem solved”, even though that problem happens to be of second order in your system and be inaudible.
There are multiple problems to solve in audio. Acknowledging this fact is also a good way to avoid any commercial bias - not pretending your specific argument is commercial, but as a general remark.
Life is always a compromise, we all have to find our own balance. With my set up I don’t have issue over changing sound quality. That’s important to me. It all just works and sounds great, as great as it needs to for my pleasure.
You can always spend more and do more. The upgrade train never stops, it’s such a shame it never arrives at the destination.
Note to self: never admit you are happy otherwise someone will try and convince you that you aren’t.
13 Likes
Your speakers have other problems…
The digital input of your speakers f.i. is limited to 96kHz…
Modern DACs produce 705.6kHz…
Those Meridian speakers need mqa for hires, they can’t handle real hires.
One would start to believe they want you to like mqa 
And as if by magic…
2 Likes
Hi Christian , Could I ask what your technical background is?
The original data is not 705, it is rarely above 96.
It’s true, modern DACs upsample to 705.
And so do the Meridian DACs.
They can’t handle 192k qobuz.
For spending, I try to stay away from commercial realizations as much as I can - that for me is far easier on the analog side. To be relatively successful, it requires some knowledge, some time and a taste for rigorous experimentation, but it is an interesting trip. I like to test ideas as well as I love music. For me the trip is also interesting as the destination evolves. For example I am now into Opera music, that to my perception requires a fairly demanding quality of reproduction to be enjoyed, owing to its multiplicity of instruments and voices. It took me some years of patient efforts, admittedly. Alternately, one can buy an expensive setup in one go to pass that threshold.
Now I perfectly understand your way, I have several friends who are on the same line. In particular, it takes a bit of a change in mindset, for sources.
We have been used for years to buying a cartridge or a CD player “for its sound” - that did not change unless we opened it to change the decoupling capacitors, or output integrated circuits for example. With dematerialized music, computers, streamers, suddenly software matters, and the sound of a given device is no longer warranted to stay the same over time…
As clients we have to endorse this idea and pray that in the long run at least, and hopefully at every step, updates actually improve the sound, especially as they are not always reversible ! So, a bit of change in mindset, yes, and it can feel strange.
Another aspect is that our own audio software, the one between our ears, is dynamically adaptative. That is why we might bias ourselves: when you visit a new friend who makes you listen to music, we need some adaptation time and some varied pieces to be able to say something about his system, to get used to the room (unless the room is really better than the system). The same happens for us at home. We get used to the sound of our system, we tend to eliminate its biases, just because we focus on the music, or on some aspects of the music.
Indeed our senses evolved to detect changes. Who would like his brain to constantly remind him minute after minute, hour after hour, day after day, that we are wearing socks or shoes…
For that reason, it is always refreshing to (play) or listen to live, acoustic music, once in a while…
3 Likes
Yes, and it is argued that time misalignment is important for SQ, so therefore…?
But these are different things.
The audio time misalignment people talk about is in the range of microseconds.
The jitter measurements are in picoseconds (a picosecond is a millionth of a microsecond).
Jitter does not cause time misalignment, it causes distortion.
4 Likes
This is perfectly true.
Well said.
This is the peak frequency spectrogram of a 24/192 file. ZZ Top, Gimme all you Lovin. Downloaded from Qobuz.
The top frequency is 96 kHz.
Look into any HiRes file. All of them look similar. Most of them have a noise line at 27 kHz.
Above 25kHz there is nothing. Why do we need more then 96 kHz sampling rate (48 kHz frequency)?
This spectrogram I created this morning.
1 Like