ECdesigns

[Huubster]

2 hours ago, Norton said:

I have used Qobuz with the ECD kit, but with an unmodified  U192 (no prospect yet of getting updated it seems)

 

After the initial reports on the upgraded U192 I contacted John, I couldn't wait, and he agreed to send mine in right away. John received my unit today 11AM, upgraded it and returned it around 3PM, I can expect mine back tomorrow. So you could contact John and ask him for the same. Little sidenote, these timelines are for local postage, within NL.

 

I'll be waiting for the mailman in person and report back with my findings. I will not be testing as extensively as bodiebill and tapatrick. I will be able to compare the UPL playing though the Fractal DAC, which is reference for me, with the U192 connected to my Innuos Zen Mini, with Innuos' own LPS. I will compare with local file play from it's internal HDD, WAV only.

 

 

[Qhwoeprktiyns]

So I had a friend over who has never heard ECD's products, but is a passionate audiophile, like all of us. It was valuable having an outsider's point of view (expectations can taint our judgement). 

 

The UPL is really an outstanding source - there is "magic" to it (my friend's own words). Concerning the U192, it "elevates" subtantially the quality of any source, and I think everyone agrees as well that differences in sources are less significant due to its "interference blocking techniqie" (interferences outside the 20Mhz bandwidth required to pass the data). 

 

I found  that  my simple raspberryPi network player was not quite up to the level of the UPL, and I was able to identify precisely segments of tracks in which these differences were obvious, and important. I used the track "No Moon at All", by Diane Krall - I know the track well as it has often be used when testing systems. I noticed two things when comparing the RaspberryPi/U192 and the UPL: the voice, with less "texture" in the first configuration (probably missing some micro-details), and the piano solo (starting around 2:30) with some harshness in the high notes, and less clarity  again.

 

While very good, the RaspberryPi/U192  did not offer the completely relaxing and "realistic" sound of the UPL. I guess I was a little too optimistic ...

 

So back to basics... I trust John Brown when he states that the U192 and the UPL are (or can be) on par - but in what configuration ? Going back to his post on DIYAudio: playback of local files on a MacBook. He told me after that battery vs power supply made no difference. No wifi, bluetooth..

 

So the first thing I tried was to switch to local playback, but for practical purposes on my NUC running Windows 10 (this cheap model: https://www.intel.com/content/www/us/en/products/boards-kits/nuc/kits/nuc5cpyh.html). I compared local playback using Foobar and network playback using LMS (squeezelite). The sound of local playback seemed promising, so I also tried it completely disconnected from the network (used a keyboard and hdmi screen to launch a playlist) shutting down some processes in Windows 10 (nothing too extreme, just uninstalling a bunch of useless apps), and disconnecting the keyboard/screen. I also added some ferrite beads to the extremeties of the USB cables, and placed the NUC farther away from the DAC. I tested playback from a USB key. I will try also an extermal HD and SSD. The NUC was powered with its own cheap SMPS, but plugged in on a seperate outlet far away from my system. 

 

In this  configuration, after careful comparison with the UPL, I reached the conclusion that the sound quality was in fact identical to the UPL, and just as satisfying. So that is now a good "reference", and I am very happy with the sound quality. 

 

Next steps is to experiment further and try to identify which changes are important:

- how much of an impact does the network connection have, if any, when using it only to control the playback (not to stream the files from one computer to another)? 

- is the placement/shielding important? 

- does the software have an impact? 

- storage and storage power supply

... 

 

Given the fact that direct interferences from the source to the U192 (going through the USB cable) are limited to that 20Mhz bandwidth, I don't think you can necessarily apply the same recipes as before. 

 

John also suggested placing the U192 at a distance from the DAC, and I will try that. I may also try to provide some shielding around the DAC.

 

Quoting him: "wireless interference EMI  spreads similar to radio waves and can travel long distances and reach each and every circuit, luckily related energy level drops rapidly with increased distance....  When combining digital audio with  conventional analogue signal processing (interlinks, pre amplifiers, amplifiers) these problems are likely to get worse". 

 

I will also try to test other streamers (a friend has a SOTM). Good servers like the Antipodes or the Innuous will certainly provide at least lower RFI/EMI, and may be better at handling a network connection to control the playback. I am not convinced that software optimisations are going to change anything, but will see. 

 

John also hinted to the possibility of further limiting  the bandwidth for data transmission within the U192, something already implemented in the Fractal DAC (where the bandwidth used is only 100Khz), but is more expensive and complex to implement.  I think we can get very good results as is. 

[tapatrick]

1 hour ago, hopkins said:

John also suggested placing the U192 at a distance from the DAC, and I will try that. I may also try to provide some shielding around the DAC.

Thanks for the great report! This is interesting info.

 

I did play around with RPis for a while with lots of isolation, reclockers, LifePo4 power etc but in the end found the low powered ARM processor is a limitation. For sure I found the basic RPi for playing music is one of the worst sounding sources.

[bodiebill]

Regarding the optimal source I am still toying with the idea to try one of the (rarely found) boxes that converts SPDIF to USB, thanks to @pm325 's suggestion. This would enable, for instance, the SDTrans384 to work with the ECD combo.

 

Anyone knows whether the Mutec MC-3+ USB could do that?

https://www.mutec-net.com/product_mc-3-plus-usb.php

It does look like it as its USB connection says 'I/O', but I am a little confused as this would require a (nonexistent?) USB cable with USB B connectors at both ends. 

[tapatrick]

29 minutes ago, bodiebill said:

Regarding the optimal source I am still toying with the idea to try one of the (rarely found) boxes that converts SPDIF to USB, thanks to @pm325 's suggestion. This would enable, for instance, the SDTrans384 to work with the ECD combo.

 

Anyone knows whether the Mutec MC-3+ USB could do that?

https://www.mutec-net.com/product_mc-3-plus-usb.php

It does look like it as its USB connection says 'I/O', but I am a little confused as this would require a (nonexistent?) USB cable with USB B connectors at both ends. 

Ghent Cables would be able to help you with a custom B to B cable. They are very helpful and reasonably priced.

[email protected]

https://www.ghentaudio.com/usb/index.html

[seeteeyou]

Quote

Any input presented at the AES3, S/P-DIF or optical can be recorded over the USB.

 

https://audiophilestyle.com/forums/topic/16801-mutec-mc-3/page/42/?tab=comments#comment-611944  

On 12/28/2016 at 11:22 PM, julian.david said:

This is unfortunately not possible despite the bidirectional USB interface of the MC-3+USB. The reason is that both the MC-3+USB and the DAC are USB slave devices expecting a USB master (i.e. a computer) on the other end of the line.

 

Recording should be supported (i.e. saving to a specific file that's stored on local drive) by Mutec while we're talking about SDTrans384 → Mutec → Computer → U192ETL in real time here.

 

What kinda software could actually accomplish that particular task mentioned above?

[bodiebill]

25 minutes ago, seeteeyou said:

Recording should be supported (i.e. saving to a specific file that's stored on local drive) by Mutec while we're talking about SDTrans384 → Mutec → Computer → U192ETL in real time here.

 

So that does not look promising. I have no interest in recording as the music files are already available for the computer, for playing to the U192ETL.

 

I was actually thinking of SDTrans384 → Mutec → U192ETL, i.e. bypassing the computer.

 

Mutec's manual nor reviews are explicit about this, but it seems what I want is not possible given master-slave roles?

 

I will ask Mutec for confirmation.

[pm325]

4 minutes ago, bodiebill said:

 

So that does not look promising. I have no interest in recording as the music files are already available for the computer, for playing to the U192ETL.

 

I was actually thinking of SDTrans384 → Mutec → U192ETL, i.e. bypassing the computer.

 

Mutec's manual nor reviews are explicit about this, but it seems what I want is not possible given master-slave roles?

 

Maybe @pm325 can confirm or refute.

Maybe via OTG cable? Need to confirm by Mutec.

[seeteeyou]

AFAIK Julian himself is indeed an employee of Mutec

 

https://www.facebook.com/mutecpro/posts/2038324139582039

Quote

Producer/engineer and MUTEC associate Julian David

 

Did anyone actually own any USB audio devices that don't require a computer / mobile device of some sort to operate? In other words, the communications between the USB hosts (i.e. computers) and the USB slaves (i.e. USB audio devices) should rely on the device driver to work properly.

[Ben75]

17 hours ago, matthias said:

 

OK, I have Qobuz Studio free trial, will try Spotify as well.

 

Any listening impressions from others?

 

Thanks

 

Matt

Hello Matt : according to John own words... streaming is not great at all and he would recommend you to rip yourself the CDs in Flac or Wave to get he best sound quality.

 

The apparently limitations with streaming providers are they use a DSP to process music playback so it not transparent anymore.

[Qhwoeprktiyns]

I think John's comments may have been speculative, at a time when he was trying to figure out differences between the UPL and U192, and check for "bit-perfectness" in various operating systems and applications.

 

However, I did read about Qobuz applying dsp to upsample, but only when using Qobuz's app.

 

Spotify may be a different case altogether as they offer downsampling from 320 kbs for performance reason (for ex: cell phone usage). 

 

Streaming services rely on files that are provided to them by distributors. 

[matthias]

8 minutes ago, hopkins said:

I think John's comments may have been speculative, at a time when he was trying to figure out differences between the UPL and U192, and check for "bit-perfectness" in various operating systems and applications.

However, I did read about Qobuz applying dsp to upsample, but only when using Qobuz's app. 

Streaming services rely on files that are provided to them by distributors. 

 

Agree,

experiences with an UHQ server show that streaming from Qobuz does sound inferior to stored music in this server but superior to music stored in a NAS.

 

Matt

[matthias]

A just released answer of ECD on diyaudio about the PowerDAC:

 

https://www.diyaudio.com/forums/digital-line-level/79452-building-ultimate-nos-dac-using-tda1541a-769.html#post6293442

 

Very interesting 😃

 

Matt

[tapatrick]

3 hours ago, Ben75 said:

Hello Matt : according to John own words... streaming is not great at all and he would recommend you to rip yourself the CDs in Flac or Wave to get he best sound quality.

 

The apparently limitations with streaming providers are they use a DSP to process music playback so it not transparent anymore.

This is definitely not true in my neck of the woods 

 

[Qhwoeprktiyns]

I played around with various configurations this afternoon. My NUC is still giving me a superb sound with Foobar, and I have no idea why! 

 

Playing music on an external SSD, and connecting it to the network to control playback remotely does not degrade the sound.

 

Everything else I tried has inferior sound, with those same caracteristics I was describing earlier:

- LMS and squeezelite-x on the NUC

- wtfplay on the NUC (which is really odd) 

- LMS on my raspberry pi

- USB Audio Pro on my Samsung tablet

 

Its a little frustrating not knowing why something sounds better. If it all has to do with "interferences" there are certainly many possible causes. Perhaps these different players/configurations generate not only varying levels of noise, but in different "spectrum"? 

 

I would be curious to try "audiophile" players, but I don't have any. 

 

We should organize a seminar sonewhere in Europe where everyone brings their player of choice and we can compare 😊 More seriously, I would welcome some kind of comparative evaluation of sources. One way would be  to find an audio dealer willing to benchmark various sources (high end or not) against the UPL. Not sure anyone is going to be too interested in doing that. 

 

 

[Qhwoeprktiyns]

14 hours ago, tapatrick said:

I did play around with RPis for a while with lots of isolation, reclockers, LifePo4 power etc but in the end found the low powered ARM processor is a limitation. For sure I found the basic RPi for playing music is one of the worst sounding sources.

 

I googled "USB noise spectrum", and looked at a few white papers (from TI, etc...). Noise radiation from USB connectors and cables is powerful stuff - it is easy to check for yourself by hooking up various USB devices to a computer and observe how they interfere with a bluetooth connected mouse, for example. I would not be suprised that a 40$ RaspberryPi is simply not as well designed, in this respect, than an Intel NUC or MacBook, not to mention a Taiko Extreme

[One and a half]

1 hour ago, hopkins said:

 

I googled "USB noise spectrum", and looked at a few white papers (from TI, etc...). Noise radiation from USB connectors and cables is powerful stuff - it is easy to check for yourself by hooking up various USB devices to a computer and observe how they interfere with a bluetooth connected mouse, for example. I would not be suprised that a 40$ RaspberryPi is simply not as well designed, in this respect, than an Intel NUC or MacBook, not to mention a Taiko Extreme

You would have read about radiated and conducted noise in among the papers.

 

Conducted noise is the worry for USB interfaces, since it has sufficient strength to transmit to a DAC for example via the USB cable.There exists a direct conflict of high speed digital data operating at MHz/GHz speeds radiating noise spectra on printed circuit boards, and the smallest of capacitance acting as antennas is enough to convert the radiated noise into conducted noise. That's just one component, like a digital gate, so multiply this several (hundreds) times, creating a cocktail of noise. 

 

Suppressing this noise is extremely difficult, since it depends on the construction of the PCB, the type shielding used on which parts radiate the worst noise. Extensive testing is necessary to weed out even the simplest, and that of course adds up to the design cost of the computer/DAC.  

 

It's no wonder USB cables (for one) can and do sound different, since their construction acts like a filter at certain frequencies which either reduce or let through noise, the main effect is listener fatigue. 

[seeteeyou]

FYI - Greg found that Allo USBridge Sig as well as properly modified + powered RPi 2B could yield excellent results, about the same league as SDTrans384 with a decent (but not his best) PSU

 

http://www.tirnahifi.org/forum/viewtopic.php?f=15&t=4541&start=330#p87998

http://www.tirnahifi.org/forum/viewtopic.php?f=15&t=4541&start=350#p89678

https://www.diyaudio.com/forums/vendor-s-bazaar/337859-shanti-dual-lps-5v-3a-5v-1-5a-10.html#post5809606

https://www.diyaudio.com/forums/vendor-s-bazaar/337859-shanti-dual-lps-5v-3a-5v-1-5a-34.html#post5887722

https://www.diyaudio.com/forums/pc-based/329911-getting-allo-coms-katana-dac-40.html#post5945670

 

Positive impressions of USBridge Sig

 

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/page/613/?tab=comments#comment-1009197

 

I guess that maybe SDTrans384 could still be fed to I2S input of ReceiverPi while connecting U192ETL to USBridge Sig then? 

 

---

 

When it comes to storage, there's always this wide-temperature + SLC option from Apacer but not everyone would like to pay so much for 4GB these days

 

http://eflash.apacerus.com/spec/USB/Drive/EH353/Industrial-USB_EH353.pdf

https://www.rutronik24.com/product/apacer/apha004gak0eg-2t/7710798.html

https://www.soselectronic.com/products/apacer/apha004gak0eg-2t-307137

 

Though here's something used from China

 

https://www.ebay.com/itm/Apacer-8GB-Industrial-CFIII-8GB-Extenden-Temperature-CF-card/293621554873

 

Patrick and others on Tír Na HiFi tried SLC cards and they seemed to be beneficial, others were actually getting larger sizes for installing Windows

 

http://www.tirnahifi.org/forum/viewtopic.php?f=15&t=4541&start=390#p90199

http://jplay.eu/forum/index.php?/topic/2413-apacer-cf-card-extended-range-temperature-as-ssd/

[tapatrick]

6 hours ago, hopkins said:

I would not be suprised that a 40$ RaspberryPi is simply not as well designed, in this respect, than an Intel NUC or MacBook, not to mention a Taiko Extreme

The Pi is a wonderful little gem and Ian Canada has and continues to produce amazing add ons to elevate the SQ to loftier heights. I did follow this for a while and achieved satisfying results - I might hook mine up and see what it does with the EC Combo. @gstew could weigh in more fully with some of the latest Pi developments as he has taken this quite far with great reports.

 

[tapatrick]

5 hours ago, One and a half said:

Suppressing this noise is extremely difficult, since it depends on the construction of the PCB, the type shielding used on which parts radiate the worst noise. Extensive testing is necessary to weed out even the simplest, and that of course adds up to the design cost of the computer/DAC.  

 

It's no wonder USB cables (for one) can and do sound different, since their construction acts like a filter at certain frequencies which either reduce or let through noise, the main effect is listener fatigue. 

This would explain why the EC components sound so good as John Brown seems to have spent a lot of time dealing with these sources of noise and using his ElectroTos interface over USB plus band limiter filters

[bodiebill]

@hopkins

Do I remember well that you also had an Allo USBridge? Was it not used in your comparison?

[tapatrick]

3 hours ago, seeteeyou said:

Patrick and others on Tír Na HiFi tried SLC cards and they seemed to be beneficial, others were actually getting larger sizes for installing Windows

 

http://www.tirnahifi.org/forum/viewtopic.php?f=15&t=4541&start=390#p90199

http://jplay.eu/forum/index.php?/topic/2413-apacer-cf-card-extended-range-temperature-as-ssd/

Yes I found that the SLC cards were one of the main causes of uplift in SQ when using the SD card player

[Qhwoeprktiyns]

I sold my USBridge a while ago (when I started using the UPL exclusively).  In addition to a couple Raspberrypi I also have a NanoPi Neo. Will give it a try sonetime. 

 

P. S. In the tests I conducted yesterday I placed the U192 as far away as possible from the DAC, and my NUC (and other sources) as far away as possible from the U192. FYI John makes 3 meter versions of his USB cables. Mine must be 1.5 meters.  The SSD I used is a Samsung T5 (1Tb) which has a nice metal enclosure. Will try an external HD as well, some other time. I put ferrites on all USB and power cables. 

 

 

[bodiebill]

10 hours ago, hopkins said:

I played around with various configurations this afternoon. My NUC is still giving me a superb sound with Foobar, and I have no idea why! 

 

Interesting. John Brown said that according to their own tester Foobar + 2017 WASAPI driver (if properly configured) and JRiver + WASAPI deliver bit perfect, and "that is about it" (John's words).

 

Quote

I would be curious to try "audiophile" players, but I don't have any. 

 

Did you try Upplay?

https://www.lesbonscomptes.com/upplay/

There is a Windows version that I use with minimserver. The latter has the advantage that it can convert to wav on the fly. To my ears it sounds better than JRiver.

I must say I am using these on a separate computer that acts as control point.

[Qhwoeprktiyns]

33 minutes ago, bodiebill said:

 

Interesting. John Brown said that according to their own tester Foobar + 2017 WASAPI driver (if properly configured) and JRiver + WASAPI deliver bit perfect, and "that is about it" (John's words).

 

 

Did you try Upplay?

https://www.lesbonscomptes.com/upplay/

There is a Windows version that I use with minimserver. The latter has the advantage that it can convert to wav on the fly. To my ears it sounds better than JRiver.

I must say I am using these on a separate computer that acts as control point.

 

All the software I tested is bit perfect, and I checked that using ECD's Mosaic UV, which is how John does it as well. I also made sure that the bit-perfect test failed in some cased (when using Windows direct sound or playing a Flac file in the Chrome browser). I did not try UPNP, and will give it a shot. 

 

LMS+squeezelite is bit perfect, but for some reason it just does not sound as good. No idea why. Have not tried Roon, as I do not have a subscription (and don't want to bother). 

 

I am convinced that there are probably many ways to get good results with the U192, and I can only urge everyone to experiment with an open mind. 

 

If the key point is to minimize noise under that 200 Mhz bandwidth (and avoid EMI/RFI spreading to the DAC + analog equipment), there is really no way of knowing what is going to work without actually testing it...