Tidal Desktop application sounds better than ROON on the same box and with the same driver, why?

Then why the lightweight endpoints still do not sound as good as audiophile players in the same environment? Sorry, again, I don’t mean to troll…just want my ROON to sound better

Correct me if I’m wrong. Your software consumes a lot of resources when indexing a library and performing the audio analysis. But once all that is done, retrieving the indexed data is a very lightweight operation. I’m a software developer myself and I don’t believe that retrieving a record from an indexed 1,000,000 records database can be a challenging task for modern CPUs. I don’t take into account any DSP of course.

Hi @brian

That is a very informative post.

You will know that I’m an avid supporter of Roon across these forums - and you may not know I’m also a big supporter elsewhere too.

But just for the purposes of completeness, your post considers everything other than Roon and RAAT itself.

That’s not to say that there is an issue or compromise with RAAT that affects sound quality and at the same time if you knew of a downside of RAAT that affected sound quality I wouldn’t expect you to shoot yourself in the foot and tell the world - instead, I’m sure you are working at it.

But there are enough people (end customers) out there that have commented on Roon’s sound quality that I hope you continue to listen and continue to try and find what it may be, if it is indeed RAAT.

Again, much love but as someone who has purchased 2 lifetime subscriptions (one for myself and one I purchased for my old man, my dad) I hope you understand that sound quality is numero uno to me.

To be honest if I never tried Audirvana recently I’d be non the wiser and just as happy as my old man is. But unfortunately I did after I read the opening paragraph in John Darko’s post here:

Anyway I hope I haven’t offended. I have tremendous respect for all you guys at Roon.

Keep up the great work and please continue to have sound quality at the top of your priorities.

1 Like

That long post was trying to explain that there is not an easy, direct answer. Roon and TIDAL do many things differently. Assuming the bits are the same, and you are sure you perceive a difference, then you must believe that the effect is happening out of band of the information stream.

Those out of band effects involve more components than just Roon, but the phrasing of your question is treating Roon like a black box, as if swapping the software is a single, isolated, well controlled variable that does not have ancillary effects–which is not the case.

This, in my mind, is the real voodoo–we’re starting with a perceived (but not scientifically established) difference, assuming the bits are the same, and thus opening the door to explanations that are very indirect and complex.

Personally, I am extremely skeptical of my own listening perceptions because to my brain that is where the most dramatic variation is most likely to be. I do not have the level of certainty that you seem to have about personal listening test results, ever.

I’d still like to find out if the bits are the same. I recall TIDAL doing something with volume leveling, and that album you mentioned is poorly mastered (lots of inter-sample overs). If TIDAL is dropping the volume–even by a tiny bit–that could explain a difference.

Roon is an in-RAM object database. Big load at startup, huge footprint while it’s live. Garbage collector walking through it now and again. Not very analogous to a traditional DBMS. The object count is roughly 10x that of other music players for the same set of content because of all of the extended metadata, too.

3 Likes

Actually, any album sounds better from Tidal desktop. I spent most of the weekend listenning and comparing using different albums and settings. I want my ROON to sound as good as Tidal at least and as good as MinimServer>>JPlayStreamer at most.

USB Regen is not doing USB galvanic isolation. The replacement product ISO Regen does (make sure the ISO switch is set to 1, not ON).

It didn’t consider Roon because we’d already removed Roon from the room. It did consider RAAT (see the whole subject about data rates, PCM vs FLAC, etc).

We’re not averse to acknowledging our faults. This kind of case is a little bit tricky because all we have to go on is a loose collection of opinions, many of which are based on old versions of Roon or RAAT (anything <1.1.21 is out of date for SQ evaluation), and many of which contradict each other.

More than one time, manufacturers have come to us saying “our users say RAAT doesn’t sound as good as UPnP, but we cannot prove a difference in our listening tests”. I’ll leave unsaid what the most logical explanation for that is…

If you truly want to see progress on this, help us find repeatable examples sound quality problems that we can see and work on in-house. We will work on anything that we can reproduce conclusively.

The reason I mentioned hardware manufacturers above is this: if a manufacturer feels that RAAT is having adverse effects on their product’s sound, I’d expect them to come to us with a clear report detailing what we are doing to make their hardware not sound as good as it could so that we could address it.

If there’s a problem and no such report, we can only assume that either the manufacturer does not feel there is a difference, does not care enough to investigate/pursue the difference, or they do not understand their product well enough to make a coherent report.

Roon Ready partners receive the full source code to RAAT, and can customize it as they see fit to get the sound quality they need. Some totally re-write the “output plugin”–which handles the communication with their audio hardware. Some just do smaller tuning. Ultimately the responsibility for ensuring the sound quality of their product is with them. This idea is core to our certification: we are responsible for the user experience, our partners are responsible for the sound quality.

I would not–as you sometimes seem to–expect Roon to take ownership of the proprietary details of another manufacturer’s product and address their sound quality problems indirectly via a generalized change to RAAT or Roon.

We make generalized improvements all the time–usually by reducing resource usage–and these benefit everyone.

What we don’t do is make our whole public message about sound quality–because we have so much other stuff to talk about.

This may create the perception that we care about sound quality less than others–which is definitely not the case.

8 Likes

Fully understood Brian.

If you have a Macbook with A+ , a Chord Hugo2 and Sennheiser HD800S cans I would love for you to see/hear if you can notice the difference that I can.

Even with RoonServer in a different room, only RoonBridge running on the Mac and both the Hugo2 and Macbook both running on batteries…

Or if you’re ever in Australia (Melbourne or Sydney) please let me know and you can do a blind test on me and listen for yourself - if you’ve never heard the observed and reported differences yourself.

1 Like

I also feel that Roon offers worse sound quality than other software solutions but I fully understand that this is hard to quantify and therefor work on. But what is much easier is to consider the “drop-outs” on DSD you get with a Chord Dac, this is clearly easier to quantify. Maybe this “irregularity” in how Roon delivers data to the DAC can also be a reason to the worse sound-quality? Even if not, just fixing the drop outs would be good, no?

Also putting the sound quality issue on all DAC manufacturers is a bit strange, especially as many have old DACs that cannot be fixed, rather users would need to buy new DACs?

Dropouts on Chord DACs are not exclusive to Roon. This was one of the first thing we determined when investigating the problem. I have reproduced them with Roon, A+, HQPlayer, NAA, …

I think holding Roon responsible for out of band interactions between a computer and a DAC is a little bit strange, being that we don’t directly control what your computer puts into the environment nor what the DAC accepts from it. What we do control is whether the audio stream is delivered intact and on time. That is obviously our responsibility.

The later comments about hardware were in context of the Roon Ready program–where the responsibility lines for sound quality vs ux are clearly drawn, and there is an active collaboration.

2 Likes

Thanks, yes that I agree with.

Did you get any response from Chord then? Are they working on it or do we (i.e. the end-users) need to bug Chord about this.

We are expecting to receive a few more hardware pieces from them soon…once that happens I was going to sit down with all of the products at once and see if we can learn anything about how to modulate the problem and make a solid report to them.

1 Like

Sorry for more off topic, but I really hope you can help Chord get to the bottom with this, Roon + Chord is a killer combination for discovering music through tidal. Great that you are getting more hardware to enable better validation of Chord+Roon. :slight_smile:

I run JRIVER from a Windows10 Laptop to my DragonFly RED (and formerly exasound e22 and Chord 2Qute) DAC. I too thought I was ‘done’ for SQ by adding Intona Galvanic isolation and a battery for the USB +5V. However, I would just like to add that by clamping ferrites(WURTH ELECTRONIK 74271131) on my USB cable I have put my playback on another level. That RF noise from my Laptop makes its way to my DAC (and perhaps my AMP) to affect sound quality is plainly obvious to me. Even across the air-gap of my Intona …near GHz noise got thru …at least until i installed my ferrites (10 of them). I get incredulity from DAC engineering people that this matters …so I don’t expect any resolution from them.

Did you measure this “near GHz noise” or is it just an asumption ? Did you evaluate the various connections in your signal path for Rf noise or did you just throw in 10 ferrite rings ? Is the perceived difference really an improvement or is it just a difference ? Most people don’t have a clue what RFI means and how it works and I read way too many completely false assumptions everywhere. Just randomly throwing in filtering of any kind (ferrites are filter after all) is not allways the best solution.

I know a lot of people who fooled around with way too many ferrit rings everywhere in their system and claiming it was a noticable upgrade in sound. Most of them removed all of them within a year because in the end it rubbed too much life out of the sound. Ferrites on asymetric digital cables (spdif for instance) can create more jitter because you are literally inserting an inductance in the signal path wich can cause phase shifts, transmission line ghosting and such. Believe it or not, most of times introducing some extra jitter in a system is perceived as more “air” more “depth” at first but in the end it’s nothing more then enhanced treble glare.

Back on topic on the perceived differences in playback software. First of all let may state that the differences are allways minor, allways and I have tried about any player that has ever been written. Way way too many audiophiles are shouting of the roof when they have perceived a tiny difference and will let you believe it is a difference of night and day. If that is really the case, and please be truthfull to yourself here, then there is something terribly wrong somewhere in the signal path but it is a hard assumption to blame the software for it. Yes I also do hear atiny difference between Roon and the Tidal app but as stated before it is minor and it is impossible for me to tell wich one is better then the other. On my headphones Roon is defenitly better because of convolution filters and crossfeed filters, for the rest the difference is so small I don’t even care. What I do care about is that the Tidal app interface plainly su…c…k…s big time, it’s completely terrible to use at least for me. Roon makes it so much better to use.

2 Likes

I am actually not interested in improving perceived sound quality. That is all in the listener’s head.

1 Like

This is just your opinion. I actually find that only poorly implemented players sound the same…bad.

What is a good sounding player in your ears if I might ask ? I’d like to try it. Maybe there is one I don’t know yet.

Have you ever tried removing the jitterbug out of the chain. I had one and it did more harm then good, in fact some software players had more trouble with then others. Don’t ask me why, I don’t know but it just did. On three computers it sounded actually worse with the jitterbug in place, dac’s where not recognized, some players didn’t even play at all, some distorted. It’s most defenitly NOT a plugin and forget device.

I’m interested in all of…

I would like people to perceive Roon’s sound quality as excellent.
I would like Roon to measure as well as possible.
I would like Roon to perform as well as possible under listening tests that are conducted scientifically.

No-one in this thread is backing up their opinions with measurements, and no-one in this thread is running listening tests that would meet scientific standards (double blind, multiple subjects, multiple trials, …). Non-scientific, non-blind or single-blind/informal listening tests fall under the first item, not the third.

What I said there is very direct (if a little bit nihilistic)–since I don’t expect rigorous, objective definition of what “sound quality” is to come along and rock the Audiophile universe anytime soon, perceived sound quality is the thing we’re actually discussing here.

There’s nothing wrong with the idea of making concrete, measurable changes–that is most of what we have done up until now.

Many of our partners perform end-to-end measurements on devices before releasing their integrations. We measure all of our digital signal processing components whenever they are changed. Roon has been measured by unrelated third parties.

Early in Roon’s life, one of our partners took it upon themselves to iteratively make measurements and give feedback about them. We fixed about half a dozen issues (mostly prior to 1.2) and by improving the measurements, perceptions of Roon’s sound quality also improved. Shortly after 1.3, we went back and did a revision on the up-sampling filters oriented towards improving their measured performance…these too created positive perceptual reports.

“It sounds bad to be, sorry can’t elaborate further” is a valid viewpoint, but not a very constructive one. In the real world, we have to understand a report before fixing it. Drive-by opinions that don’t go deeper than “I don’t like it” aren’t going to make anything better. Efforts to help show us repeatable issues are always welcome.

10 Likes
  1. MinimServer -> JPlayStreamer running in KS mode on Windows
  2. Audirvana 3 on MAC
  3. BugHead - sounds really good but has an awful UI and is very difficult to setup right.

In both cases you need to run the player on a carefully optimized computer that has nothing running on it except of essential system and sound related processes. I personally started to hear the difference between different players once I created a dedicated environment.

My windows machine is Win 2016 in Core Mode, AO 2.20, Process Lasso, JPlay 6.2, Fidelizer Pro 7.10. The hardware was also carefully selected and optimized. My windows computer has only 26 processes running.

My MacMini running Audirvana in the second system is also optimized with CAD script. Audirvana 3 is simply phenomenal. The sound is very smooth and detailed at the same time.