Sound quality of Roon versus JRIVER

I’ve participated in a few double-blind tests conducted at shows organized by Stereophile, before they took the cop-out editorial position that getting meaningful results from blind tests is essentially as fundamentally impossible as determining through which slit of a diffraction grating a single photon passes. Some of the things they claim to hear today has destroyed much of their credibility with me. I have found that really honestly evaluating sonic differences that are very small to non-existent is very hard work. At some point, if the music is good, I just want to sit back and listen (and save the hard work for the piano).

The panels I was in were conducted shortly after Gordon Holt proved to be the only listener who could get statistically significant results blind listening between a pricey tube amp and Bob Carver’s transfer function clone thereof. And these were things that could theoretically make a difference, not boutique USB or power line cords.

I haven’t compared Roon versus JRiver sonically because given they are both configured correctly (no unwanted DSP) I don’t expect to hear any differences, and don’t go looking for trouble if the sound is good. Also, as a longtime audiophile my views have evolved to be diametrically opposed to pat_bannon’s: given a high quality system built up through the years according to my personal tastes, and recordings that can produce magic, I’d be willing to give up half an iota of sound quality (assuming there were evidence it’s even necessary) for a superior user interface like Roon versus one that’s a constant aggravation. The whole music experience matters to me. Also, given the tasks of editing music file tags (which JRiver does better than anything else IMO) and cleaning and cueing up LP’s, my personal preference is for the latter.



I use Jriver 23 streaming DLNA to a Sonore Ultrarendu (bitperfect FLAC). I can also stream the same FLAC file to the same Ultrarendu using RAAT and my Roon server (bitperfect). The DAC connected to the Ultrarendu is my Chord Hugo 2 and I listen through HD800 headphones. I am confident that it is a resolving setup.

The sound quality is very good through both Jriver and Roon setups and I am happy with either / both. But there are very small audio differences between RAAT streaming and DLNA bitperfect stream rendering through the systems. These differences are what most would consider to be minute.

I was interested in your description of the Jriver method sounding slightly more liquid. This is probably the closest to what I am hearing. It generally makes Jriver DLNA a little more musically engaging (listen to the choir at the start of John Rutter’s For The Beauty of the Earth) but sometimes feels like Roon may be more accurate (the choir voices sound slightly less like angels and more like people). Both sound amazing.

Anyway I am interested as supposedly they are playing the same file through exactly the same hardware and both in bitperfect mode. Maybe there are some rendering / DAC timing differences at the roon endpoint in the Ultrarendu?

Sorry for the long post.


Before Roon come along, I used JRiver for ripping, editing as well as playing back. The SQ for both are virtually indistinguishable other than features.

Now I use Roon exclusively for playing back from NAS and Tidal streaming.

1 Like

Hi MusicEar,

I agree that the differences are very small and that roon UI and cataloguing features are substantially better than Jriver. However there are small audio differences (not saying one is better than another) to my ears which I find interesting as It challenges my understanding of bit perfect.


What’s happening inside the UltraRendu is radically different when you play a FLAC via UPnP vs playing via RAAT – RAAT streams the PCM, so no FLAC decode is done… this FLAC decode in the UPnP case is very slightly adding CPU usage to the stream processing. That might be causing unintended effects to the analog processing stages later down the audio path.

With the right equipment, you should be able to tap the USB output of the UltraRendu and capture the stream of USB Audio data – I’m sure it’s the same. What won’t be the same is the EMI/RFI – but it shouldn’t make a real difference.

If you really want to know if you are not having a error via bias, try getting someone to help you to do a blind test.


Bit perfect will assure you that the playback is as transparent as possible. However, if there’s a overhead processing such as using DSP, it may introduce noise. If these noise is not properly taken care of and gets into the DAC path, then it will affect the SQ.

There’s are many ways to reduce this, one such example is to isolate the noisy PC that does the DSP processing and stream to a ultra-low noise endpoint then to a DAC.


… but before deciding on a solution to a problem, do try to make sure it really exists, in other words, do the blind test like Danny said, and don’t simply assign this to some semi-mystical ‘noise gremlins’ and buy more stuff… I mean, wasn’t the microrendu supposed to be exactly the cure for all that ails USB playback, at least compared to a regular SBMC (like a Cubox)?.. Work from verifiable facts, such as your ability to tell the difference between the two streams with very high precision, if that should indeed be the case.
Good luck, and let us know how it works out!


that’s what the UltraRendu is for in this case, which is the unit in question.

1 Like

It is indeed which is why I raised the point. Again I don’t want to get carried away with the differences but it is very easy to switch the Rendu between uPnP and RAAT and to remotely access both Rriver and Roon servers so i have been able to listen to a lot of comparisons in a timely fashion.

I cannot rule out bias. I have a friend whose hearing I trust coming to visit in December and I will do a blind test with him. Any advice on best practice? Should I even tell him what I am testing?

@danny thanks for the insight into upnp v RAAT. I hadn’t realised that the rendering of FLAC to PCM was carried out in a different location for each. I am very interested to understand why I am hearing a difference - I am constantly surprised how listening to music can be affected by extremely small changes in timing and accuracy and don’t rule out the impact that bias can have either.


I agree. I am reasonably confident that there is a perceptable difference but it will be interesting to test.

My friend is visiting in early December. He is a long standing acquaintance and I respect his hearing. I’ll report back after conducting sone blind tests.

It doesn’t have to be him doing the listening.
Here’s how I would do it - although nobody asked, but what else is there to do on a Saturday afternoon, right?
You would listen to the same first e.g. 30-60 seconds of a song, ten times in a row.
Each time, one of the two sources will be selected by someone else (an assistant) whose moves you cannot see.
Your assistant first draws 10 numbers (e.g. using and writes them down, without you seeing these numbers.
Then, for each number, if it is odd they use the first source to play the song, and if even they use the other source. This makes it likely to listen to the same source twice or more in a row.
You write down which source you think you are listening to.
If you get at least 8 out of 10 correctly, you can probably assume that there is indeed a difference.
I would repeat the experiment at least another time.
Have fun! :slight_smile:


An excellant methodology for any A/B testing. Including the often threaten testing of ‘audiophile’ cables.

And to the extent anyone complains about the “30 to 60 seconds” of a song (with the usual anti-double blind testing comment that one needs longer time with a song rather than a snippit), then remember that there is nothing about double-blind testing (other than the availability of enough time) that keeps you from using longer test items (entire songs, or entire albums for that matter).

Exactly, I wrote 30-60 merely as an example, based on Adrian’s assessment that he knows the song intimately, so a short sample could suffice. Of course, the actual parameters may vary, while I think the most important point is that there should not be a simple switch between the two sources back and forth (A/B/A/B/A…), as that reduces the test accuracy dramatically. Cheers guys!

One clear area where Roon is lacking in SQ -

After a quick calibration, the SQ improvement to me is astonishing and, well, scientifically backed so no surprise really.

More people should be aware of that, if listening at <75dB.

I need to provide you with an update on my observations wrt Roon / Jriver sound quality.

My planned test did not happen in December. I have continued to compare Room / Ultrarendu vs Jriver / Ultrarendu and there is an absolute difference in the sound quality between these two software players (& netwoking transports). My initial observation that Roon seemed more detailed and truthful vs Jriver seeming more musical actually caused me to reflect on what it was that made me feel that the sound from Jiver was more musical. I believe that familiarity played a part together with a slight prominence in the bass that the Jriver pipeline delivered over Roon.

However after many Months of comparing and also after optimising my PC OS (using ProcessLasso) I am convinced that Roon / RAAT audio quality is superior in every way compared to Jriver / DLNA. I ended up upgrading my mains power supply (AC regenerator) for the network end to ensure the Ultrarendu / LPS1.2 and Chord Hugo 2 / Sbooster LPS had the best environment to operate in - overkill perhaps? .

Apologies for the delay in responding to you all after my initial observations in December. I am a convert to Roon and have no doubt of its ability to deliver audio content - I am enjoying both the interface which allows me to broaden my listening and most importantly the music itself.


1 Like

I agree with your observations, even though I don’t use JRiver. It still matches my observations with other method of playbacks. I tried DLNA via Audirvana briefly as well. Roon sounds more accurate to me, even if at first it can sound a bit too dry.

I did spend quite some time tweaking on the network side of things though to make the sound as fluid as possible to my ears.


I’m not impressed with Audirvana via Dlna .
Direct from the computer it is on , it is fantastic and is easily on par with Roon audio quality .

Did you serio Audiophile ptimizer un minimal server mode And Roon as Shell replacement on Windows server 2012 r2 ? Ir Even with Windows 10 pro? The differences between bith oplayer área muchas more audible when you hace ypur PC absolutly optimized forma HighEnd Audio transporte.plse ley me know your setup

[quote="Radu_Popescu, p