Does Roon download entire track into RAM? [Memory Playback Discussion]

Ok it feels like you’re in the wrong thread, but taking your comment at face value; Roon team have already responded to this request, in this thread.

I would like to be able to follow up on your comment please? (Even though we are here to talk about roon).

About a year ago they ‘discovered’ that a tiny NUC with AL was the best thing since sliced bread! I challenged this and told them that I was using a DIY PC with humongous LPSU and a CPU with max. cores and threads, and that I had good results. They all mocked/scorned & ridiculed this, saying that PC’s with enormous power was totally wrong for audiophiles and that it created too much noise and interference with all the switching circuits embedded in the mainboard and was asked to try out the tiny NUC as an endpoint, with a second box as a server (2 box solution). I haven’t noticed any improvement with those kinds of setup! Fast forward a year and guess what(?) many of them are now using single box solutions with massive memory and Xeon processors (or the most powerful NUC’s they can get their hands on!). Seems like low power small devices are now out and huge high power devices are in (again?). My point being that they are not consciously moving forward with new discoveries and pioneering experimentation, but merely going round in circles and they now praise the type of equipment they ridiculed me for using only a year ago! As for ‘small is best’, Sonore have done this with the Microrendu sometime BEFORE the tiny NUC’s ever came on the radar.

Have Fun! (LOL).

1 Like

the only time austinpop on AS/CA was asked to do a blind test, he got it wrong and supposedly picked the inferior device as ‘the best’ of those on test - LOL!

1 Like

I reactivated this conversation 6 days ago here. It seems that my main point eluded some of the folks who’ve commented since then.

This isn’t about believing the reports or not, but about doubting the conclusion that the SQ issue on their streamer must be a problem with roon lacking memory playback and so roonlabs should investigate. Why doubt this analysis? It was reported that roon sounds like LMS+squeezeplayer without memory playback on that particular streamer. IIRC, the OPs were satisfied with both, then were shocked how the LMS SQ improved when memory playback was turned on in squeezeplay. This should have lead the OPs to investigate if there’s some platform (or other) issue affecting both software stacks that memory playback gets around. Instead, they suggest roon SQ should be “fixed” with RAM playback like LMS+squeezeplayer SQ was fixed. This seems to me to be an invalid conclusion.

I quoted some of austinpop’s initial positioning about the CA thread in my first post too, @James_I. He indicated they want to report observations and don’t want to get bogged down if there are no readily available analytical explanations. A doubtful explanation that would have bogged down their own exploration is thought to be sufficient here to get roonlabs to investigate their theory, which, as roonlabs explained, had already considered and reasoned through. The proponents haven’t explored more likely explanations, and by their own account aren’t looking for explanations anyways. It doesn’t help their case that they haven’t provided any rigorous/repeatable observations.

1 Like

I really wanted to stay out of this thread… but your post kind of brings up points of some things I thought about… so why not :grin:

There is an assumption here that because LMS+squeezeplayer with memory playback sounds different to Roon, that different means technically better…

Even if @brian were to nail double blind testing and correctly identify that Roon and LMS+squeezeplayer with memory playback sound different… can different actually be technically worse?

For anyone that follows Rob Watts (Chord DAC designer) posts on Head-Fi or has the chance to chat directly with him:

"As you know, RF noise creates noise floor modulation, as the intermodulation distortion from random RF noise is a white noise modulated by the wanted signal. This then results in noise floor modulation, and is very very audible. It accounts for the things sounding brighter and less smooth; additionally, when you reduce RF noise, things sound considerably warmer and darker, and one consequence of this is perception of tempo - more midrange gives the impression of a slower tempo, as individual instruments have much more body.

Now if somebody prefers the brighter sound from more noise floor modulation, then fine - that’s their taste and preference. But it’s not accurate."

Rob has said publicly (many times on Head-Fi) that he often sees people incorrectly thinking one digital source is “better” than another (yes we are assuming bit perfect playback here) but they may actually be enjoying something that is technically worse… mistaking more ‘detail’ for more RF…

So is it possible that those reporting more detail, more dynamics with LMS+squeezeplayer with memory playback , are simply enjoying the results of increased RF, causing increased IM distortion and noise floor modulation…

Hard to know for sure unless someone measures?

So is this thread asking @brian to do noise/RF measurements?

Or should the people making the claim be doing this and bringing results to Roon?

I’m just asking questions! I’m not brave (or smart) enough to make any claims.


I have tried many software and hardware over the years and recently used a standalone battery operated player that reads wav files from a usb stick: no network, no streaming, very low power processor dedicated to playing audio. The company that manufactures this player (ECDesigns in holland) are well respected engineers.

The difference in SQ is significant (silence and transparency). Its use case is obviously limited (no streaming, no Roon).

The point is: in my experience, you either have to live with streaming and accept less than perfect quality, or do without streaming.

To think that one can take a general purpose PC, tweak it with software and some added components, and achieve miraculous results is a fallacy.

The various “experiments” being done on CA, such as using RAM playback, are just shots in the dark, performed without either a reference to benchmark the results or any measurements to support the claims.

I see no reason for Roon to invest time and effort on these unsubstantiated claims. Today it is RAM playback, tomorrow it could be some other trick…


I would imagine sound quality is system dependant (How good it is). Currently, with streaming at CD quality and MQA, I am experiencing the best audio I have ever heard and see no loss of quality between streamed files and my local library. A golden age indeed.

Meridian DSP SE


Chris, have you heard MQA 2.0 will be exclusive to Meridian’s new line of DSP speakers? But not backwards compatible with your model sadly… And they’re saying MQA 2.0 is heaps better too…

What are ya gonna do mate !?

I better tell you I’m joking before you speed out of the driveway to your nearest Meridian dealer :grin:

:joy:That did make me smile … :joy:

You are right though, there is always something new. You get to an age where you have to buy something…

1 Like

I figured ‘I better tell him it’s a joke’ otherwise this would be you, on the way the nearest dealer!


This thread reminds me of a ridiculous thread on where punters have written tombs of bull about building an ideal music server, every tweak yielding orgasmic levels of increased sonic pleasure.


Likewise, I’ve been following a thread on the Linn owners section of Hi-Fi Wigwam that enthusiastically extols the amazing SQ improvement of using Cat8 Ethernet cable over Cat7. Predictably, the thread turned into a farce! :sunglasses:


I’ve read threads extolling the virtues of good quality power supplies, I’ve bought same with very noteworthy increases in SQ. Same with a managed Network switch.

If I read a reliable source writing about a demonstrable improvement in SQ, I’ll listen and then maybe try.

I guess it gets down to the credibility of the writers. I’ve seen and heard enough to have an open mind. I’ve been repaid for keeping it open.

1 Like

I was about to post something similar. That I would rather hear austinpop out over just ridiculing the entire concept of computer tweaks. There are knowledgeable folks over there at AS/CA.

The issue I see is that the AS/CA forum thread that austinpop oversees is very deliberate about excluding anyone who wants to think through the logic, science, or engineering of the tweaks. The premise of the thread is that if you don’t believe with your heart don’t come with your brain. (This is an interesting metaphor for many different groups in the world).

I would be interested to see if austinpop can adjust to the different tweaker culture here in the Roon forum – that is, where what he calls “opinions” or “the noise” in his “signal to noise ratio” are unwelcome in the AS/CA thread, that type of questioning and intellectual rigor are more the norm here. The question is, can austinpop come in here, lay out the information and experience he has or that they have, and be patient enough to understand he has to address rather than ban posts from those who want to explore the science or engineering. And that includes delivering data.

Look at it with this analogy. Austinpop is a doctor who volunteered to treat sick people in some developing country. He comes in with his box of western meds but in the course of the year finds out that the patients respond best to some local weed that gets boiled and fed to the patient. After a year of experience with this, he’s convinced the weed is the true medicine of the future. Then he brings it back to the western university and is presenting it to a bunch of scientists and doctors, who are interested in his empirical evidence but insist on seeing the science. They want to know why. They want to understand and discuss side effects. They want a true set of trials and studies. Austinpop gets impatient and calls the scientists and doctors a bunch of flat-earthers and walks out. Because of that, no one benefits - the potential miracle drug is never looked into, and austinpop is seen as a crackpot. Both sides needed to find a way to communicate with the other for everyone’s benefit, but they were too stuck in their own mindset to do so.

So, the question is, if the doctors and scientists are willing to attend the meeting, can austinpop adjust his sensitivities to rigorous query that will naturally include both skepticism and some ridicule, and open a true dialog within the culture here at Roon.

Yep, this is part of my point. I think we ought to be open to non-absurd ideas and tweaks, but tweakers have to be open to other explanations AND to those who want to explore their ideas using western scientific thought rather than techno-mysticism. The scientific thinkers need to be open to the idea they don’t know everything, but the shaman have to be open to the idea that what they know CAN be reduced to science.

Sorry…maybe I just haven’t had enough coffee yet this morning…


How about this nonsense? Point #2 is my favorite.

I suppose you don’t believe the Roon Server partner who underclocks the CPU either?

Ha, more bullsh*t that can’t possibly make any audible difference.

My re-post of your post wasn’t directed at you. I know you don’t believe such nonsense, at least that’s the way it seems.

As far as those who want the fastest CPU and those who underclock the CPU, I say we put them in a torch lit arena and let them fight to the death. It’s the only way to resolve it.

1 Like

Torches throw off a lot of flicker which interferes with the true assessment of battle-worthiness.

1 Like

I would agree except when you start reading all the the stuff that he does and how every single time it is a “massive improvement” in sound quality. First it was with the super expensive word clocks. Then with the low latency OS on otherwise noisy hardware.

All credibility is lost because everything sounds better and, no, it cannot be confirmation bias. The fact that he is unwilling to discuss the technical merits of any given new super improvement just makes it worse. That thread on AS is just about the most laughable audio thread on the Internet.

1 Like

All I’m saying is don’t throw the baby out with the bathwater. Whether some of the claims are totally ridiculous, there is still a lot of knowledge there.

1 Like