The new version’s sound quality has worsened, especially with the lack of power in the low-frequency submersion.
@junbiao_jiang your view is very similar to mine a few months ago. I noticed the midrange and upper bass was thin sounding. I don’t anymore 
Psychoacoustic maybe
My only differences before upgrading my Server hardware was changing from DietPi running Roon Server to Roon Rock. Some will chime in saying there isn’t a difference as it’s all in the digital domain. I agree partly, but “I” did notice the difference. Whether that was just in my head is unclear.
Since installing my new Roon Rock machine I’m back to loving the SQ of Roon. I’m not saying this is the reason, just coincidence. Might be the recent super-moon we experienced.
My advice is to reboot everything, unplug some stuff, and plug it back in 
. Move something from one location to another and listen again. You’d be surprised how your brain may hear a difference after that.
Whilst it’s unlikely your audio equipment has developed a fault, but not impossible, maybe check everything.
Does another option sound better?
2 Likes
15 Likes
There’s absolutely nothing in the release notes to hint that anything has changed regarding transport, so I’d be very surprised if there’s any difference. Take a deep breath, think happy thoughts, and play your favorites loud! Enjoy the music 
5 Likes
Great video share 
Outside of the use of the DSP engine (volume leveling, crossfade, EQ, etc.) Roon is not changing your music in any way shape or form. In fact, Roon doesn’t deal with music, it deals with data. That data is turned into music much later in the system and totally outside or Roon’s domain.
Consider Roon as a very nice and featureful interface with the sole goal of setting up plumbing between a track file on some disk and a DAC.
It will access the file, make sure that the data is appropriately decoded into PCM or DSD, then create all the pipes to take that data to the digital side of your DAC. This task, from an IT point of view, is dead simple and really not heavy on resources. A lot of those tasks are also not Roon tasks but operating system tasks. That data is untouched outside of standard decoding (with only one way of decoding).
Now, if you have any use of Roon’s DSP engine, that data is changed. There again, it’s maths on data, not on music. There aren’t multiple ways of using maths to do EQ or whatever the engine allows, there is only one.
That makes it quite impossible for Roon to have variance in sound quality between versions.
Any variance of SQ will happen beyond data delivery to your DAC and only if you change anything from there on.
If, for any reason, there was an SQ variance, that would mean that Roon is not just arranging plumbing, but changing the data in the middle, when not asked to do it. That would be a serious bug in the very core of the product, making the probability of its occurrence close to zero.
11 Likes
It seems that there are at least two calculation ways for DSP: the correct one and the Roon one. 
Look at this bug, which is over 2 years old and still not fixed…
1 Like
OK, there may be one element of the DSP engine that may have a problem.
That doesn’t render the rest of my argumentation invalid at all, though. Especially since I would suppose that DSP users are not necessarily that common.
1 Like
I have no statistics on that point but would strongly disagree. Personally I use DSP on four of my endpoints using convolution to make up for non optimal speaker placement. I works wonders and is simple to do using something like the Housecurve app, or as complicated as you like using REW/Accourate etc. It costs nothing to do but your time so why wouldn’t you?
1 Like
Yours is advanced behaviour. I’m sticking to my belief (I don’t have any stats either, granted) that most user don’t have the same approach as your.
Anyway, this is far from the original topic.
I use DSP in some way in all zones. Remember even something as simple as volume levelling is DSP.
1 Like