ECdesigns

[Qhwoeprktiyns]

18 minutes ago, bodiebill said:

 

I was not talking about the difference between ethernet and wifi, but about audible difference between two audio players that run on the server. I cannot imagine that this difference would disappear by switching from ethernet to wifi. And if it would it would mean that wifi is not as good as ethernet 🙂

 

Two bit-perfect players can sound diffeent because they are designed differently, and when run generate different activity on the computer and therefore different interference "spectrum". If that computer is connected somehow to the U192, through a "wire" - the interferences may have an incidence. If the connection is wireless, you don't have that problem (but you may have others).

[bodiebill]

20 minutes ago, hopkins said:

 

Two bit-perfect players can sound diffeent because they are designed differently, and when run generate different activity on the computer and therefore different interference "spectrum". If that computer is connected somehow to the U192, through a "wire" - the interferences may have an incidence. If the connection is wireless, you don't have that problem (but you may have others).

 

What do you mean by 'incidence' in this context?

 

The question is whether the audible difference between the players is due to the source or to the whole topology.

For instance, we would agree that a 24/96 version and an mp3 version of the same material will always sound different, whether via ethernet or wifi. We do not even want them to sound the same or we would suspect the audio system to be faulty or low-res.

For now I am assuming that the players difference is of this type. But I am open to other explanations.

 

PS: Sorry, I read to fast... You assume both players (JRiver and Upplay) are bit perfect. Maybe that was not the case when I compared.

[Qhwoeprktiyns]

Yes, I was not referring to differences in the source format, but differences in noise "spectrum" linked to the configuration (hardware, OS, application) of the server/player.  

 

[bodiebill]

6 hours ago, hopkins said:

Yes, I was not referring to differences in the source format, but differences in noise "spectrum" linked to the configuration (hardware, OS, application) of the server/player.  

 

Interesting. So one question is: what is the definition of 'source' in the context of 'source independence'. Apparently the 'source format' is not part of it.

 

Or another question: will two audio players on the server in a server-client setup always sound the same if both spit out bit perfect streams and supposing the ECD combo is used?

(Which could be a different way of asking whether such an audio player belongs to the 'source' in the above sense.)

 

Not saying that I know the answers, but I like these simple questions because the whole subject matter -- and I agree with @hopkins there -- is too complex to jump to hasty conclusions.

 

May the ... be with you...

 

[Qhwoeprktiyns]

Bit-perfect playback is simply a pre-requisite. 

 

Let's assume that whatever you use is bit-perfect: Roon, UPNP, LMS, etc... With the U192, you have two things being input through the USB cable. 

 

- power (5v)

- data (the music bits) 

 

The power is supplied by whatever device you are using may be of varying quality, but let's assume the U192 does a good job of cleaning that up. 

 

Now the data is communicated with a signal that has a bandwidth of 5Ghz, which is then "reduced" to 20Mhz in the U192. That 20Mhz bandwidth is the minimum bandwidth, ECD ecplains, to let the "music data" pass through. Within that 20Mhz bandwidth, they explain that you can have some noise from the source, and that can adversly affect the clocks in the U192 (from what I understand) and deteriorate the quality of the signal that is then passed on to the DAC. 

 

Well the noise in that 20Mhz bandwidth originates from the source, and everything that is connected to the source. 

 

So the less "noisy" your source, the better chances you have of getting good results. 

 

Using the U192 does not mean that everything upstream will sound exactly the same. But as you yourself found out, the "effect" the source has with the ECD combo is probably much lower than with other systems (and even with a bad source, the ECD set sounds good). 

 

How different will these various sources sound? What is a "low noise" source? I don't have the answers. I can give some answers based on what I have tried, that's it. 

 

I don't think a "high end" solution is needed. But you can't just put anything either... 

 

 

[Huubster]

Great to see all of you going to great lengths to find the best computer source, I have respect for the patience and time you put in to find your optimal solution.

 

Personally, I'm a completely different type of guy I don't want to mess arround with all that and therefore I use a dedicated streamer server which enables me, out of the box, to easily control my music using an app on my phone. 

 

Might not be of interest to all of you, but today I did a first comparison of my Innnuos Zen Mini as a source for the U192 against my laptop, being a basic Windows laptop, nothing special. 

 

I did absolutely nothing to tune my laptop for audio, no special software, no nothing, connected via wifi, using the Spotify app installed.

My Zen Mini uses a dedicated LPS, using Spotify steaming, cable connection.

 

Long story short:  The difference was actually far bigger than I anticipated. A lot of dynamic energy gone, compact soundstage, smallish sounding. Still very pleasant to listen to though, no harshness at all and still a very clean sound, but the energy and dynamics I'm used to were gone. 

 

My 2 cents, back to my Innuos now

 

[bodiebill]

41 minutes ago, hopkins said:

Bit-perfect playback is simply a pre-requisite. 

 

I know that this is an axiom in the ECD system of thinking. However I have heard some report that they prefer to upsample redbook to 24/96, for instance offline, or on the fly using minimserver. They find this makes a positive difference for SQ. Is that really impossible?

[Qhwoeprktiyns]

16 minutes ago, bodiebill said:

 

I know that this is an axiom in the ECD system of thinking. However I have heard some report that they prefer to upsample redbook to 24/96, for instance using minimserver. They find this makes a positive difference for SQ. Is that really impossible?

 

I have never tried. If the upsampling algorithm is good I don't see why it could not work. I have a few 192 tracks that I downsampled to 96 and they sound great. 

 

I think they insist on that because left alone some applications or OS (Windows direct sound, for example) may give pretty poor results. I was personally interested in their bit perfect test because I thought that my LMS server may be doing some funny stuff on windows, or that network playback could screw things up in some way, but I don't think that's actually the case.

 

When you have a complex  problem to solve its good to break it down  and be able to discard some of the variables involved to pinpoint more precisely the real issues - avoid comparing apples and oranges. That's what I meant by "prerequisite". 

 

If you are not satisfied with your DIY computer, maybe just do like Huubster and get something like an Innuous and call it a day? It's the "throw money at the problem" approach (what I did with the power supply) ☺️ Does not always work in audio as we know. 30-day money back guarantee is a good thing... 

[bodiebill]

2 minutes ago, hopkins said:

If you are not satisfied with you DIY computer, maybe just do like Huubster and get something like an Innuous and call it a day? 

 

I was satisfied with it, but with the ECD combo I hardly hear a difference with my new NUC (untweaked other than PS). so I am now using the latter. In that sense, I did call it a day 🙂

 

(HDPlex 800W DC-ATX and Apacer ECC wide temperature RAM already have other happy owners thanks to ECD.)

[bodiebill]

10 minutes ago, hopkins said:

When you have a complex  problem to solve its good to break it down  and be able to discard some of the variables involved to pinpoint more precisely the real issues - avoid comparing apples and oranges. That's what I meant by "prerequisite". 

 

That makes sense. It is one way to get rid of one of the variables.

 

By the way, I am trying wifi on my NUC, but so far not successfully with GentooPlayer...

[matthias]

1 hour ago, Huubster said:

Great to see all of you going to great lengths to find the best computer source, I have respect for the patience and time you put in to find your optimal solution.

 

Personally, I'm a completely different type of guy I don't want to mess arround with all that and therefore I use a dedicated streamer server which enables me, out of the box, to easily control my music using an app on my phone. 

 

Might not be of interest to all of you, but today I did a first comparison of my Innnuos Zen Mini as a source for the U192 against my laptop, being a basic Windows laptop, nothing special. 

 

I did absolutely nothing to tune my laptop for audio, no special software, no nothing, connected via wifi, using the Spotify app installed.

My Zen Mini uses a dedicated LPS, using Spotify steaming, cable connection.

 

Long story short:  The difference was actually far bigger than I anticipated. A lot of dynamic energy gone, compact soundstage, smallish sounding. Still very pleasant to listen to though, no harshness at all and still a very clean sound, but the energy and dynamics I'm used to were gone. 

 

My 2 cents, back to my Innuos now

 

 

Interesting,

now the Zen Mini is the entry level in the Innuos range.

I am wondering what would happen with an Innuos Statement.

Do you have an Innuos dealer where you can try the ECD combo with the Statement or a Zenith Mk3?

 

Matt

[Norton]

1 hour ago, bodiebill said:

 

I know that this is an axiom in the ECD system of thinking. However I have heard some report that they prefer to upsample redbook to 24/96, for instance offline, or on the fly using minimserver. They find this makes a positive difference for SQ. Is that really impossible?

As NOS there is some logic in feeding the DAC96 the highest input rate possible. To my ears the current and previous gen ECD DACs  (even the 16 bit MOS) sound best with 24/88 or 96 input, whether  “native”, or upsampled RBCD on the fly via Roon, HQP or XXHE plus some limited try outs offline using the TEAC tool.

[Qhwoeprktiyns]

29 minutes ago, matthias said:

 

Interesting,

now the Zen Mini is the entry level in the Innuos range.

I am wondering what would happen with an Innuos Statement.

Do you have an Innuos dealer where you can try the ECD combo with the Statement or a Zenith Mk3?

 

Matt

 

I'll do it but not for free, and my fees are very expensive 😂

[matthias]

5 minutes ago, hopkins said:

 

I'll do it but not for free, and my fees are very expensive 😂

 

Do you want to buy a Taiko Extreme for your ECD combo? 😁

Matt

[Huubster]

32 minutes ago, matthias said:

Do you have an Innuos dealer where you can try the ECD combo with the Statement or a Zenith Mk3?

 

Hi Matt, in what way would it benefit you if I do this? Lots of impressions are already being shared here, very informative and with a bit of deduction you can already draw your own conclusions. You keep asking people to try more and different things on their sets, this is not going to give you the answers you need, every set is different. 

 

If you are curious, just buy the set and try it yourself, if you are not in NL I'm sure you can arrange something with John regarding a trial period.

 

But FWIW, I bought my Zen Mini from a local dealer and he adviced me the Zen Mini, regardless of price. His words: "Do yourself a favour and buy the Mini, add a good power supply like a Keces P8, use proper AC and DC power cables and it will beat the Statement hands down". 

[matthias]

3 minutes ago, Huubster said:

You keep asking people to try more and different things on their sets, this is not going to give you the answers you need, every set is different. 

 

Please do not care for me, it would be just an interesting data point for the community here in the search for source Independence.

 

Matt

[Qhwoeprktiyns]

Maybe you are curious to know how far  the U192 can be taken in terms of SQ ? 

 

I can't answer for others, but I can assure you that in my case I don't really feel like it would be worthwhile to spend more on the source at this point, other than perhaps for "practical" aspects, like getting an "all in one" server or something more "robust" than a Pi (and in a nice and cool aluminum case). 

 

That is based on listening comparison between the U192 and UPL, which I hold as a reference, and based on the fact that listening to music with the U192 is now completely satisfying, whatever it is I am playing (Casals playing Bach, the Rolling Stone"s "Get Yer Ya Yas Out", or a 1928 Bessie Smith recording, to give some recent examples from my playlist). In fact, i was really worried at first that what I heard was some kind of fluke (or lets say a lucky combination of uncontrollable factors, like air humidity, power grid supply, or whatever). With each day that passes I am reassured! 

 

On top of it I get the convenience of LMS with a nice web app (subsitute that for Roon, or whatever you prefer) and a fairly simple system with not too many cables and boxes. I love it. 

 

More money invested in my system would either go to speakers, or to improving my vinyl setup and collection (which is basic, I just got started). 

[Rexp]

10 hours ago, Huubster said:

Great to see all of you going to great lengths to find the best computer source, I have respect for the patience and time you put in to find your optimal solution.

 

Personally, I'm a completely different type of guy I don't want to mess arround with all that and therefore I use a dedicated streamer server which enables me, out of the box, to easily control my music using an app on my phone. 

 

Might not be of interest to all of you, but today I did a first comparison of my Innnuos Zen Mini as a source for the U192 against my laptop, being a basic Windows laptop, nothing special. 

 

I did absolutely nothing to tune my laptop for audio, no special software, no nothing, connected via wifi, using the Spotify app installed.

My Zen Mini uses a dedicated LPS, using Spotify steaming, cable connection.

 

Long story short:  The difference was actually far bigger than I anticipated. A lot of dynamic energy gone, compact soundstage, smallish sounding. Still very pleasant to listen to though, no harshness at all and still a very clean sound, but the energy and dynamics I'm used to were gone. 

 

My 2 cents, back to my Innuos now

 

I tried Spotify again as a result of one of your earlier posts and it does sound better on my PC than Tidal. However when I use my phone as the source using the same USB DAC, Tidal HIFI wins easily. I wondered whether you've tried Tidal with the ECD gear? Thanks! 

[tapatrick]

7 hours ago, hopkins said:

I can't answer for others, but I can assure you that in my case I don't really feel like it would be worthwhile to spend more on the source at this point, other than perhaps for "practical" aspects, like getting an "all in one" server or something more "robust" than a Pi (and in a nice and cool aluminum case). 

I agree wholeheartedly. Apart from trying a Pi when I come back from being away, I have found very minor differences with different sources feeding the EC combo in my system. It has amazing SQ regardless of input - but that is in my setup and others may differ but I doubt it. This is very important news! 

 

I know how hard it is to give up endlessly tweaking, trying new things and reading about it, as this is a fun and big part of this hobby.

[bodiebill]

1 hour ago, tapatrick said:

I know how hard it is to give up endlessly tweaking, trying new things and reading about it, as this is a fun and big part of this hobby.

 

Actually for me this was a relief, especially as I suspected some (certainly not all!) of the perceived results of PC tweaking to be wishful thinking. Of course I wanted to hear the benefits of Apacer ECC wide temperature RAM, but did I ever really?

[Qhwoeprktiyns]

In that line of thinking,  @austinpop who has "come out of the closet" to reveal his distaste of spaghetti is now a good candidate for trying the U192 + Fractal DAC and getting rid of a few more boxes 😁

 

All said in good spirit and respect for everyone's efforts and contributions! 

 

 

 

[Huubster]

11 hours ago, hopkins said:

I can't answer for others, but I can assure you that in my case I don't really feel like it would be worthwhile to spend more on the source at this point, other than perhaps for "practical" aspects, like getting an "all in one" server or something more "robust" than a Pi (and in a nice and cool aluminum case). 

 

Don't forget that hopkins feeds the RPi with a 700 euro power supply..

 

@hopkins: I assume that contributes to your sound quality big time right?

[Huubster]

3 hours ago, Rexp said:

I tried Spotify again as a result of one of your earlier posts and it does sound better on my PC than Tidal. However when I use my phone as the source using the same USB DAC, Tidal HIFI wins easily. I wondered whether you've tried Tidal with the ECD gear? Thanks! 

 

I have the same experience, but with Qobuz instead of Tidal, and also with using ECDesigns stuff. Actually I took another month of Qobuz Studio after receiving the ECDesings stuff to re-confirm my former findings. 

 

My theory is that my current system is so resolving that more information is being 'displayed'. And Spotify has simply more information, that is something I never would have believed before, it was simply impossible, mp3 vs uncompressed, no discussion... Well, wrong!

 

In my former setups (and that are many different setups actually), which were far less resolving, it was Qobuz that sounded 'better' due to more perceived dynamics. But now I can clearly hear that the perceived dynamics is actually a very compressed and unnatural sound, like pushing too much water through a narrow straw. Spotify is far less compressed, has more natural accoustic mass and as a result has far better soundstaging, in hight and especially in depth, Qobuz is as flat as a pancake..

 

And using a phone brings back the perceived better performance of Qobuz. 

 

PS: I confirmed the above with muliple streaming sources, so in my case it's not source dependent.

[Qhwoeprktiyns]

Vis a vis the Keces, yes I think it definitely adds to that extra relaxed sound (low noise, increased transparency). Unfortunately this morning the power switch of the Keces failed (bad contact - you have to keep your finger pushed on the switch for power to stay on) 😭

 

I am going to return it and try a few other options in the meantime. Will update the thread... 

 

Concerning the pi 4, use of a small USB-C port to power it is a really bad design choice. But apart for that it runs flawlessly. 

[Ben75]

1 hour ago, Huubster said:

 

Don't forget that hopkins feeds the RPi with a 700 euro power supply..

 

@hopkins: I assume that contributes to your sound quality big time right?

Just a message from John who can be trusted to have tested almost all of the main PS solution out there so far for his own RPI4.

 

"I use:

 

Macbook air 1.8GHz Dual Core i5 mid2012

Macbook Pro Retina 2.4 GHz Quad Core i7, early 2013.

 

Apple makes relatively quiet (interference) laptops because of high efficiency (optimise battery life). This greatly reduces interference power levels compared to other brands.

The macbooks are battery powered (clean power supply) and the battery charger outputs the lowest noise of all chargers we tested.

 

RPI4 is quite powerful and with a clean power supply it is comparable with a macbook running on battery power supply. PCs usually produce way more interference (mains power, no need to conserve power) so perhaps a PC or server (even running on Linux) is just not not optimal for audio.

So it is simply a matter of using a relatively clean bit-perfect source.

 

For RPI4 you need a -linear- 5V / 2.5A power supply like this one for example:

https://www.audiophonics.fr/en/hifi-power-supply/linear-power-supply-5v-25a-13w-usb-lt1084-p-10532.html

https://www.audiophonics.fr/en/hifi-power-supply/audiophonics-lpsu15-linear-regulated-power-supply-5v-25a-15va-with-usb-p-10255.html

 

Quality doesn't matter that much as long as it is linear and it can provide 5V / 2.5A. You need 2.5A for the RPI4, 2A is not enough. Wiring with low DC resistance is also very important at 2.5A peak current,  use thick wiring + suitable adapter plug."

 

So, in conclusion, a simple but reliable LPSU would do the job on par with the Kaces PSU. So yes : the EC Designs combo has never been so cheap and performant in terms of SQ (if John is right). I can't wait to get the confirmation of that with Hopkins another day - and hope he will get the chance to send back his Kaces P8 in time ... for refunding ! 😉

 

Beyond the conveniency of Streaming Services (Qobuz etc.), John told me that Streaming was not as great as local files, ripped from a CD and should be abandoned for serious listenning sessions (because of DSPs, online masters tweaked etc.)... Perhaps there is an answer there to the question : "What is the best streaming services platform"? If you have a fine tuned High-end hifi setup, perhaps this is not ideal to use streaming services playback at all.

 

Anyway, the few albums I downloaded from Qobuz were always worse than extracting the masters myself from XCDs. 🙂