Inaudible on what HiFi system? Mine? Yours? Inaudible is one of those things measuring people likes to throw around with little or no proof.
There is also 2 types or jitter, one is caused by the sender of the data for SPDIF, and then there is clock jitter which all DACs have. Clock jitter is caused by inaccuracies in the oscillator, and the oscillator is extremely dependent on clean and noise-free input power. So all sorts of electronic noise (both send over inputs and generated internally) will degrade the accuracy of the oscillator, which will cause more jitter internally.
Go to a HiFi store and ask for a demonstration of a good external clock, for example Innous USB Reclocker, and judge by yourself (the effect on the reclocker is less jitter, even though it only reclocks the USB protocol).
Once you understand all sources of jitter, you will understand a lot of strangeness in HiFi, like for example why MQA in a DAC often causes more jitter (XMOS CPU works harder -> more electronic noise -> more jitter).
Weâre all anonymous on the internet. But Archimago goes to a lot of effort to post his data, explanations of his tests, etc. Thatâs way more than most folks with âopinionsâ do.
Speaking for myself, and many of those on this forum and others who grew up using computers: No.
And hereâs why: A âstreamerâ such as the one you cite from Innuos, is NOT an audio device. It is a computer, running off-the-shelf components with a bargain-basement Intel processor so as to not require a fan. It runs an open-source operating system that supports open-source protocols, such as UPnP, which, incidentally, is complete rubbish. I see these Innuos devices for exactly what they are: computers for people with 1) lots of money; 2) an interest in audio; and 3) no real understanding of computers.
Those of us who understand computers know quite well that the streamer has one function: to transfer bits encoded in files stored on hard disks to an end-point audio device. One does not need a $15,000 computer in a fancy chassis, crippled by design so the CPU doesnât melt, to accomplish this: any old hardware will do, so long as it meets Roonâs requirements. Many of us put our properly-cooled Roon Core far from our audio systems, so the chassis fans are inaudible. But thatâs alright: we know what function the computer serves without having to look at a beautifully milled aluminium chassis next to our DAC and amplifier.
And again: that function is to transport bits from the hard disk to the audio device. This is the purpose that RAAT serves, like hundreds of other file transfer protocols: encapsulating bits within a proprietary protocol and transferring them over a network. In Roonâs bit-perfect mode, this is a lossless process; how could it be otherwise? If transport errors occur, the errors are corrected using algorithms inherent within the physical, network and transport layer protocols. If the errors are unrecoverable, data transfer comes to a hard stop. How do we know this is a lossless process, free of interference from EMI, jitter, or anything else? Because RAAT could easily be modified to write the data it receives from Roon Core to a file on a hard disk, and that file would be a perfect copy of the original.
But what of EMI interference and jitter induced by running a USB cable to your DAC, you ask? Fair enough: USB never was designed solely with audio applications in mind. But there are dozens of inexpensive remedies that address substandard USB interfaces, some of which also offer electromagnetic and electrostatic isolation, just in case youâre having a really hard time finding the source of the ground loop hum in your system. Again, understanding what a computer does, we do not need to look for $15,000 solutions to problems that donât exist, because computers (as they apply to audio) just transfer bits from one destination to another. Losslessly. Bit-perfect.
Still not keen on USB? No problem! For a fraction of the cost of something from Innuos, many of us are quite thrilled that we can put together a device running Roon Bridge that includes one or more SPDIF interfaces (even AES!) which also offer electromagnetic and electrostatic isolation. For about $250, the cost of a Raspberry Pi and the add-on SPDIF card, you have a device that is inarguably better at delivering audio to a DAC than anything Innuos manufactures. No linear power-supply required.
Those of us using these low-cost solutions âget itâ: we comprehend that computers are not audio devices, and that in our audio chain they are merely there to transfer bits from one location to another. We know they do this bit-perfect, for otherwise our modern, digital world would fall apart.
We understand computers. We do not fall for the appeal of ANY computer that makes ANY claim to audible superiority, for audio is the domain of DACs and amplifiers and transducers, and not of CPUs and network interfaces. We know that even the most feeble of computers with a network interface is capable of transmitting bit-perfect data from one network address to another. Unremarkably, you will find that those of us who understand computers are the ones with the least amount of anxiety over their impact, negative or otherwise, on the audio reproduction chain.
So, referencing back to the original âversion 1.8 improved sound qualityâ, know this: The question makes no sense: Roon is software, and software runs on computers, and computers have no âsound qualityâ. Did you happen to buy an overpriced computer from Innuos or Aurender with a laughably underpowered CPU, only to wonder later why Innuos or Aurender hasnât made a full commitment to Roon? They canât: not without putting active cooling inside the chassis. But what would you make of a $15,000 computer that looks very smart next to your DAC and amplifier, yet generates noticeable amounts of fan noise?
Those who know computers would know what to call it: overpriced garbage. But as long as we keep seeing reviews for streaming devices, and advertisements are placed in audio magazines for them, which purport to be reviewing or selling an âaudio deviceâ, weâll have manufacturers churning them out to the same duped customers, over and over again: The audiophile that hasnât the faintest idea how a computer works.
Roon Community Militia (or is it the Taliban?) should change the name of Roon Community to âBits are Bitsâ and ban forever any claims on SQ improvements without the following:
Aye, and the Community Militia should be dispatched to deprive said violators of everything apart from their speakers, amplifiers and turntables. And tape deck, should they happen to have one.
You should listen more, or get a better understanding of digital audio. I have explained it all previously so you can scroll up a little and see why something like an USB reclocker (or better oscillators overall) will make a difference.
I have personal experience from this as well, when I upgrader my streamer clock to a Crystek CCHD-575 clock. The improvement in audio was bigger than I thought (and yes, it would be easy to separate them in a blind test).
That said, the OP never said anything about bit-perfectâŚit was a vague appeal to improve sound quality and that could be done, for instance, with a better implementation of EQ or convolution filters.
I think you confirmed what I said about Aurender - specifically that âthey think sound quality would sufferâ. Hardware required to run Roon would cause an âEMI problemâ, which their own software avoids.
As far as âexperimentalâ mode with Innuos - that came about because a few individuals on the Audiophile Style forum were working with Nuno to bring that about as a means to improve sound quality. This post is a good starting point. See also this and this.
Hereâs a good quote: âNuno has indicated that the reason their internal player on Zenith SE, based on the squeezelite platform, had better SQ was because they had implemented in-memory playback, and that they could not do that in Roon, since they didnât control, or alter, the code.â Auralicâs owner said something similar to me as well back when I owned the first Aries.
Iâm happy with how Roon âsoundsâ. I definitely donât want (and donât hear) any kind of Roon-specific âsound signatureâ. But I can only speak for myself - my ears, my gear, my room treatmentâŚ
And as also explained several times, while (USB) jitter is a problem for some, it is outside the domain of Roon software and is irrelevant in the question of if Roon software can have better sound quality in future releases.
A separate thread for how to get rid of jitter would be more on point.
Its more complicated though, asynchronous USB is not time bound so has no jitter in itself, but as mentioned previously the amount of clock jitter depends on a noise free environment and clean and steady power. The more the DACs USB controller (XMOS etc) has to work, the more noise will be produced.
And obviously the more the streamer has to work, the more noise will be generated and sent to the DAC:
Software can try to limit the internal jitter generated, by being as light-footed as possible (less CPU), good buffer handling, localization of data to avoid cache misses and so on. Basically try to do its job while at the same time cause a minimum of workload on both the streamer and the DAC.
This seems to be about the same as the hypothesis I tried to formulate. Iâm open for suggestions for the formulation of the hypothesis. For now, itâs just a loose idea that Roon endpoint software is so heavy on a streamer that it can affect the DAC with noise and/or jitter.
I donât agree that is obvious, as this is not a problem for all setups.
