Sound quality of Qobuz via Roon vs Qobuz direct

Hi All,

Sorry id this is a really stupid question, I am a Roon newbie on a two week trial and am struggling to get my head around it.

I have a Cambridge Audio EVO 150 which acts as streamer, DAC, pre-amp and amp. It drives a pair of ATC SMC19 speakers and when I use its inbuilt Qobuz integration to stream HiRes audio it sound pretty damned good.

Yesterday I installed Roon on an iMac Pro in a nearby room. None of this stuff is hardwired by Ethernet but I have a good WiFi mesh and stream with very few if any dropouts.

If I play a HiRes track ‘direct’ (inbuilt Qobuz in the EVO) and then play the exact same recording via Roon, the front panel of the EVO gives me the exact same KHz and bit rate information so they should sound identical but they don’t. The Roon stream sounds less good.

Now I don’t know if, when streaming using Roon, the Roon Core is basically telling my EVO to get its stream direct from Qobuz or whether the Core is getting the stream and then casting it via WiFi to the EVO. I am guessing it is the latter.

In which case I assume that the casting is, like Chromecast or Airplay, in some way limited?

Any ideas people might have, I’d be very grateful


1 Like

The usual volume difference. Even if the difference is small enough to not being recognized, our ear fools us and we thing louder sounds better, because we hear more details. Eleminate the volume difference and both will sound the same (or in a rare case your device has a defect).


It is neither. The audio is streamed to the Roon core, processed into PCM and the PCM is sent to the endpoint.

You should make sure that you are choosing the Cambridge Audio under the Roon Ready section of Settings/Audio, not Chromecast or Airplay. You might post a pic of Settings/Audio so we can see how it is setup.

First of the Evo will offer up a few ways that Roon can stream to it. Roon Ready which uses Roons own streaming protocol called RAAT and is the method that should be chosen. It will also show up as to connect to the Evo by AirPlay and Chromecast. All music goes from Qobuz to Roon core where it then can be routed to multiple other endpoints in your case the Evo.

Ensure your level matched and are using Roon Ready and not using and DSP.

Thank you - I have the EVO connected only through Roon Ready - it is disabled via chromecast and airplay. I’ve also been as careful as humanly possible to get the volume levels matched. Is it not possible that the process of the core processing the stream to PCM and then sending that to the EVO is in some way degrading the signal? It sounds more ‘bit muddy’ than ‘bit perfect’!!!

SQ is a very personal experience. I don’t have the EVO, so can’t comment on that hardware. I assume you have made sure that all DSP and volume leveling is OFF in Roon.

Not unless you have some DSP active.

Over the years I have not found any SQ issues using Roon Ready.

Can you post a pic of your signal path?

Thank you. It is set to connect only through Roon Ready. Everything other method is disabled and all DSP is off, as is volume levelling. It’s a very neutral configuration. So I must assume that the PCM processing and streaming is to blame in some unfathomable way!

Everything is off - here’s a picture:

Perhaps it is that when listening to Qobuz directly through the EVO that the audio is being processed by the EVO or Qobuz in some way? Otherwise the two streams should sound identical.



The EVO is a streamer and DAC and pre-amp and power amp. I’m not sure what sample rate the PCM RAAT stream is but I suppose that it is as advertised on the screen of the EVO which is exactly the same as the original file on Qobuz and the same as the signal path screenshot eg:

I’ve never used the EVO but I have compared Roon RAAT against or with countless other streaming software and hardware systems.
In my experience it’s not Roon that’s altering things it’s the other systems.
It seems more likely to me it’s the stream to your EVO without Roon that is adding some colouration that you prefer.
And that’s fine. I have a tube amplifier that adds gobs of harmonic distortion and I love how it sounds.

Even more likely though, as others have mentioned, is that are minute differences in volume when you’re comparing your two setups.
The only way to be certain the volumes are identical is to measure with a microphone.
Although the volume may seem the same to your ears/brain if one is even the tiniest, imperceptibly, bit louder it can trick the brain into thinking it sounds better.


Roon isn’t the best sounding software option, just the most full featured. Audirvana sounds better and native apps can too. I use Roon due to UI and features but would prefer better sound quality from it.


I’m not going to go down the rabbit hole of explaining why any bit-perfect playback software physically can not sound anydifferent from any other bit-perfect playback software.

As a nitpick though, there is no “processing to PCM.” Anything you get from Qobuz (or any other streaming service; nobody is streaming DSD) is PCM already.


All software has bugs, including firmware in streamers. It’s quite possible that a bug or design flaw will make a streamer process different sources differently, in this case Roon RAAT and built-in Qobuz app.


That’s definitely a possibility, but it still would involve a bug that specifically takes Roon’s bit-perfect output (but not Audirvana’s or Tidal’s) and selectively runs it through some kind of processing that decreases the quality. Is it technically doable? Yes, with some effort. Likely? I doubt it.

1 Like

No, it’s enough that the realtime device driver for the internal I2S connector between the streamer module and the DAC module interacts differently with the RAAT receiver and with other sources. Back in the day that I worked with device drivers for various systems, the typical such bug involved process A writing its output data for that channel in a different way (say, different buffer sizes) than process B, causing subtle differences in driver behavior. In realtime processing, that could lead to different delays, for instance. Not saying this is the case here, but I’ve seen enough bugs of this general kind in many different systems over decades to be skeptical that streamers are immune to them.


That’s not completely “unpossible” but wouldn’t that involve some bugginess in the streamer portion of the software, where RAAT is decoded into raw PCM (or DSD) bitstream to be sent to the DA? Not to say that this can’t be the case, but this would be rather trivial to demonstrate (and catch in testing, too) - just grab the bits coming out of the streamer, and compare with what should go there. Either they are the same or they aren’t. AFAIK in any competently designed system, buffering and reclocking would happen well after any proprietary signal (be it RAAT, or a FLAC from Qobuz) has been converted to PCM…

Not to say that this absolutely can’t happen on any and all combinations of hardware and software out there but it just seems not too likely, especially when most people reporting such differences either have “highly resolving” systems consisting of either AIrpods on Bluetooth, or tube amplifiers, or also happen to believe in $800 USB cables…

But the PCM that is coming from Qobuz is in FLAC format, right? So the decompression is being done in either the ROON core or in my EVO?

For lay people like me it’s very hard to know what’s going on in which device at what stage of the process but I do know from what I’ve read that Tidal claim that their MQA is effectively perfect but that many people dispute that heavily so how do we know that RAAT actually is bit perfect? Especially over WiFi?