A novel way to massively improve the SQ of computer audio streaming

[asdf1000]

4 minutes ago, the_doc735 said:

In terms of sound quality with my subjective impressions (on topic), I find that all DSD512 makes my tinnitus worse. So for now I have to stick with PCM. Needed - tinnitus friendly DSD (LOL).

 

LOL sorry to read that doc. Surprised that this variable (DSD512 vs PCM768) with the same HQP filter (?) is enough to affect your tinnitus. We can ask Jussi in a HQP thread, to ask why that may be.

 

[asdf1000]

4 minutes ago, Forehaven said:

I still have a pop when going from DSD to PCM in AL.  I play everything native.

 

Noted - but this is something very different to what I'm describing.

 

See here:

https://www.computeraudiophile.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/?do=findComment&comment=903916

 

[asdf1000]

53 minutes ago, Miska said:

 

It is actually not very hard, thanks to (at least uRendu) being hardware-wise similar to CuBox-i.

 

I just uploaded updated image for CuBox-i and µRendu, tested briefly on µRendu too and for me it is working fine. No web interface or such though.

 

 

Wow! Thanks Jussi !

[asdf1000]

8 hours ago, Miska said:

 

It is actually not very hard, thanks to (at least uRendu) being hardware-wise similar to CuBox-i.

 

I just uploaded updated image for CuBox-i and µRendu, tested briefly on µRendu too and for me it is working fine. No web interface or such though.

 

 

Installed and working brilliantly. Seems to sound a little better too.. for HQP NAA, I think your image on the Rendu is as good as it gets (to my ears, my experience only)...

 

Remember the saga we had a long time ago with the below track at DSD512... just tested it with your Rendu NAA image (just to double check) and all is crystal clear! ?

 

Is this Rendu image also ramroot? Not that I really care for SQ reasons but it should make pulling power on the Rendu out a bit safer if i need to poweroff, if nothing is being written to the Rendu's SD card?

 

Thanks again for this! You need to be inducted into the Computer Hi-Fi/Head-Fi Hall of Fame ! 

 

[asdf1000]

1 minute ago, Miska said:

 

No, it is not. Usually these devices have so little RAM that I don't see a point in wasting large percentage of it to hold the image. Most of the image content are things like device firmwares and drivers for WiFi interfaces and such. It would be holding a lot of things in RAM that end up being never used. On a PC it is less problem to hold 600+ MB of unused data in RAM.

 

 

No problem.

 

Nothing is being written to the SD card, right? So like your other non-ramroot NAA bootable images, there's no issues in pulling the power of the Rendu right?

 

There is no safe way to shutdown anyway (no power on/off button) so I guess there's no choice but to just yank the power cable.

 

[asdf1000]

1 minute ago, Miska said:

 

Nothing much, the OS itself may write some small amounts of runtime information, but nothing critical. It is journaling filesystem so if you pull the plug it will fix up unfinished things on next boot up. And if it ever goes bad for what ever reason, you can just dump the image on card again and that's it.

 

 

Thanks again Jussi. Happy days

[asdf1000]

41 minutes ago, Bricki said:

Of course there are better dacs than a mojo but I would rather spend my money to improve things upstream because that is where I believe the more significant bottle necks are. 

 

I've had the Mojo. In my experiences (not expertise), it's USB input is much more sensitive to upstream changes than say, it's bigger brother the Hugo2...

 

With Hugo2, Rob Watts added a lot of focus on RF filtering on the USB input, something which Mojo only has on it's power (microUSB) input, not on the USB (data) input... and it helped Hugo2 (and Qutest) a lot...

 

Hugo2 is a little more sensitive to upstream changes than it's biggest brother, Dave... and Dave is not insensitive to upstream changes as romaz and others will tell you too...

 

Mojo to Hugo2 is a big improvement though (to my ears). Based on that, I would easily recommend putting more money to upgrading to Hugo2/Qutest before some of these tweaks that aren't cheap... initially anyway. The fun in tweaking and tinkering will always be there - we all enjoy it to different extents.

 

Of course if you love Mojo, ignore all the above. This is just a hobby after all, nothing life and death.

 

All the above is to my ears and experiences, not any expertise... and only talking about USB input...

 

 

[asdf1000]

48 minutes ago, Bricki said:

Interesting, I didn't know about the extra filtering on the Hugo 2 usb. 

 

https://www.head-fi.org/threads/chord-electronics-hugo-2-the-official-thread.831345/page-568#post-13795767

 

and

 

https://www.head-fi.org/threads/chord-electronics-hugo-2-the-official-thread.831345/page-568#post-13797577

 

and

 

https://www.head-fi.org/threads/chord-electronics-hugo-2-the-official-thread.831345/page-970#post-14539500

 

and 

 

https://www.head-fi.org/threads/chord-electronics-hugo-2-the-official-thread.831345/page-611#post-13888314

 

 

 

[asdf1000]

1 hour ago, The Computer Audiophile said:

Has anyone tried the LattePanda Alpha yet?

 

https://www.dfrobot.com/product-1728.html

 

Yup, had one for over a year. Too many droputs with DSD512/PCM768KHZ.

 

Jussi’s recommended UpBoard Gateway (I’ve linked a few times in this thread) is much better.

 

Very low power consumption too. Easily boots and works from a USB3.0 port - for no screen/keyboard/mouse connected and DAC that isn’t 5V bus powered.

 

Edit: my LattePanda isn’t the big Alpha that you linked but much smaller and cheaper Intel Atom model

[asdf1000]

1 minute ago, jbparrish said:

 

Hi Em2016

 

How is the UpSquared Gateway very different in terms of power to run HQPlayer than the NUC7i7DNKE? They seem very similar except the DNK has a higher burst speed if I am following this right.

 

Thanks John

 

Hi John

 

I have the UpBoard Gateway, not UpSquared.

 

UpBoard is recommended by Jussi for HQP NAA duties and works fantastically well at PCM768kHz/DSD512x48.

 

As mentioned, very very low power consumption too. I can easily power it by a USB3.0 port (900mA).

 

The UpSquared Gateway can run HQPE but with limited filter options, due to it's limited CPU. But the NUC7i7DNHE is much more capable with running HQPE.

 

My NUC7i7DNHE (in fanless case) can run Roon Server and HQP Embedded with poly-sinc-lp-2s at DSD512. It can't do xtr-2s at DSD512. Maybe the newer NUC's can though.

 

 

[asdf1000]

3 hours ago, jbparrish said:

 

Hi Em2016

 

I forgot to ask how is your DNHE performing? I mean how does it sound with HQPE? Is it better than other options in its price range and/or performance range?

 

Thanks

 

Hi jbparrish

 

Working really well, alongside Roon Server on the same machine for me.

 

 I haven’t compared with other options because I initially built it inside a fanless case to run RoonOS (ROCK). This model NUC is officially supported by Roon to run ROCK, so that’s why I selected this model.

 

But I was later surprised I could install Debian Stretch Server and run both Roon Server and HQP Embedded on the same machine and have no drop in performance. It can’t do poly-sinc-xtr-2s at DSD512 though... not enough grunt.

 

[asdf1000]

On 4/3/2019 at 6:52 AM, austinpop said:

Most important: please realize this thread is about bleeding edge experimentation and discovery. No one has The Answer™. If you are not into tweaking, just know that you can have a musically satisfying system without doing any of the nutty things we do here.

 

Unfortunately, this very message at the top of every page of this (your own word is "nutty") thread may be exactly why Roon's CTO is unlikely to spend resources (time and money) to try this stuff..

 

Respectfully, I think something that may get more attention (not just Roon's attention) is if you can demonstrate your USB source is actually a lower RF generating source than say, a Roon Ready Sonore ultraRendu powered by Uptone LPS-1.2... and that it isn't actually generating more RF? I know you've had chats with Rob Watts on this stuff (and you're a big fan of Hugo TT2 + M-Scaler).

 

Also respectfully, In addition, if Roon's CTO looked at your system chain, I'm not sure how/where he would start to be able to reproduce this at his end... there are so many variables at play with so many components in the system chain, he would need to reproduce your system identically?

 

The reason I mention this specifically is from some experience: I raised a SQ related issue some time ago with the CTO. I won't go into details about the issue or the outcome because it was a private discussion. But he did confirm he did attempt to replicate my system  setup to try and hear what I was hearing with certain tracks. My system at the time was very simple: Macbook on battery > USB to Chord Hugo2 on battery > headphones. No ground loop / current loops involved. No room effects involved. But with the complexity of your chain, it may make things tricky for Roon's CTO if he were to attempt to reproduce... I've tried this AL and RAM thing with NUC7i7DNHE to Hugo 2 to headphones and heard no difference to just Roon Bridge on Debian Stretch (minimal headless server version)... does it require a more complex setup (more PSU's in the chain, more components, more RF sources?) to hear what you are hearing?

 

I'm asking these questions respectfully. If you look at the number of feature requests on the Roon forum (Roon management have confirmed they read ALL by the way, regardless of number of likes and number of comments that follow...) I can understand why they prioritise other things much higher...

 

https://community.roonlabs.com/c/roon/feature-requests

 

[asdf1000]

17 minutes ago, Dev said:

I am not sure what you are exactly asking ? Why does Roon CTO even care about anybody's chain ?

 

Noted but respectfully, I was replying to Austinpop, not you, and I’m sure Austinpop understands where I was coming from, even if he may not agree (which is fine).

 

Spend enough time discussing with Rob Watts and Ted Smith and others about how/why things may sound “different” (maybe not technically “better”) and it becomes clearer.

[asdf1000]

1 minute ago, Dev said:

Really ? Owning the DS, I am quiet familiar with Ted Smith's thought and I frequent their forum as well. I am curious, did TS explain this in their forum anywhere ?

 

Sure Ted has discussed the concept that the more complex a system (more PSU’s in the system), the more potential one MAY have for ground loop / leakage current loop associated issues or things (sources, cables, etc) potentially sounding “different”... Rob Watts too. That’s what I meant regarding comparing with more complex chains and being able to reproduce something someone else is hearing... 

 

I wasn’t making any comment about Ted or Rob having tried these AudioLinux and RAM experiments...  Anyway as mentioned, I was replying to Austinpop...

 

 

6 minutes ago, Dev said:

Oh, I didn't release that.

 

Not a problem. I had quoted him directly... so it was a reply to him...

[asdf1000]

1 hour ago, Dev said:

Why does Roon CTO even care about anybody's chain ?

 

You asked this question. On a similar topic, Roon's CTO has publicly stated the below. With this background info, have another more careful read of my post and hopefully it becomes a bit clearer what I'm asking and why... If someone wants to question the approach the Roon CTO takes to solving issues, best you take that up with him but it sounds quite reasonable to me... especially when you consider all the current Roon feature requests to get through (see link I posted earlier) as well as existing stuff on their roadmap...

 

 

"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."

 

 

 

[asdf1000]

46 minutes ago, austinpop said:

Brian also reached out to me

 

Excellent. If the CTO has reached out to you then I don't follow the issue "but the issue with Roon is there is no way to directly communicate with their product management or development folks to log feature requests or requirements other than through the forum. It's like walking through a picket line to get some place!"

 

What more do you need than the CTO reaching out to you?

 

It's just a matter of waiting patiently for a response now, no?

[asdf1000]

2 hours ago, mourip said:

Basically this is a remarkably friendly open forum

 

This particular thread is not an 'open' thread 😉

Posts are deleted and people have been banned? That's fine of course. Anyone has the option to start a new thread of course.

 

2 hours ago, mourip said:

It is not for everyone but great discoveries are found here. Do you have any specific ideas to contribute concerning advancing sound quality?

 

I've followed this thread for a long time and tried many things. Some worked nicely, some no different in my system. I have no problems with the thread existing and trying things at all and you seem to think I'm one of those that does? Not correct though. 

 

2 hours ago, mourip said:

BTW. If you really did not want others to comment then perhaps a PM would work better for that goal. 

 

Of course anyone can comment on my post but if my post is misunderstood or misinterpreted, I am allowed to clarify that my post wasn't directly at everyone as a general comment but as a reply to a specific post by someone else... it's just to add context to my post... I hope that clarifies things...

 

I'll be Captain Obvious here but if the Roon CTO can reproduce what is being heard by others here, and Roon can find the root cause of the difference and this results in a change that improves SQ, then it's a win for all Roon users (including myself). So we just wait I guess.

 

 

[asdf1000]

1 minute ago, BigAlMc said:

Hi @Em2016,

 

You raise some valid points but as Austinpop said it's not a case of asking Roon to replicate our 'nutty' chains.

 

Brian at Roon is pretty reasonable when he eventually engages but in my mind there are two issues here.

 

1. Lack of Roons focus on SQ as they keep chasing gimmicks like streaming your screen to a chromecast enabled TV.

 

2. The sheer vitriol and dismissive posts received from the hostile masses on Roon Communities. I added my +1 to Roon please investigating RAM playback. In return I received:

 

* Prove it

* You're deluded

* Speak to Innuos about their shoddy gear

* You have a good imagination

 

All this because I politely requested Roon investigated something, where if I happened to be correct and the dissenters were wrong, might just improve their SQ regardless of whether they understood or believed in the mechanism.

 

I've said this before but audiophile music software that dismisses SQ in favour of gimmicks is running a risky strategy in my book.

 

Cheers,

Alan

 

Hi Alan,

 

I kindly appreciate your polite reply. 

 

As mentioned above, I'll be Captain Obvious but if the Roon CTO can reproduce what is being heard by others here, and Roon can find the root cause of the difference and this results in a change that improves SQ, then it's a win for all Roon users (including myself).

 

So we just wait I guess.