Thread Rating:
  • 3 Vote(s) - 3 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Mutec MC-3+ USB
#21
I don't think that what's been written is correct. When we'd use the Mutec we'd connect it to a Devialet digital input, either a S/PDIF RCA or AES/EBU (I'd prefer the latter). The Devialet will extract the clock signal from the input data stream -unless- it has a ASRC (asynchronous sample rate converter) or if it buffers and reclocks the input signal. In these cases it'll throw away the embedded clock signal from it. So this ASRC/reclocking is not preferred if we'd like to profit from a high quality wordclock. After all we'd want the Devialet to use the HQ clock and not it's lesser quality own.

Since the Devialet's upsample each and every input to 384kHz/40 bit (the internal sample rate used for DSP processing) I'm not sure if any HQ clock will have any/much effect because I think it will use it's own internal masterclock for the sample rate conversion.

Those who'd want to be sure would have to ask Devialet how an incoming digital signal is processed.
PS Audio P3, Shunyata ΞTRON Alpha Digital and HC/Furutech power cables, Paul Hynes SR7EHD-MR4, DIY Roon Server & Roon Endpoint running AudioLinux Headless, Phasure Lush^2 USB cable, Audioquest Diamond RJ/E ethernet, Mutec MC-3+ USB, Shunyata ΞTRON Anaconda Digital XLR AES/EBU, Devialet Expert 250 Pro CI, Nordost Tyr Reference LS cables, Von Schweikert VR-5 SE Anniversary Edition, Anti-Mode Dual Core 2.0, JL Audio Fathom F112. More detail here.

The Netherlands


Reply
#22
(12-Feb-2016, 19:13)Antoine Wrote: I don't think that what's been written is correct. When we'd use the Mutec we'd connect it to a Devialet digital input, either a S/PDIF RCA or AES/EBU (I'd prefer the latter). The Devialet will extract the clock signal from the input data stream -unless- it has a ASRC (asynchronous sample rate converter) or if it buffers and reclocks the input signal. In these cases it'll throw away the embedded clock signal from it. So this ASRC/reclocking is not preferred if we'd like to profit from a high quality wordclock. After all we'd want the Devialet to use the HQ clock and not it's lesser quality own.

Since the Devialet's upsample each and every input to 384kHz/40 bit (the internal sample rate used for DSP processing) I'm not sure if any HQ clock will have any/much effect because I think it will use it's own internal masterclock for the sample rate conversion.

Those who'd want to be sure would have to ask Devialet how an incoming digital signal is processed.

Hi Antoine
I understand what you are suggesting, but there is no re-clocker in the Devialet. I've asked them this Q too. Everything is upsampled , yes, but not re-clocked. I'd suggest that if there was some sort of re-clocking going on inside the Devialet the impact of the Mutec would be far smaller. The difference would then be clock quality and/or re-sampler implementation. Maybe the AES input of the Devialet is "good enough" (and it certainly is very good) although I think a word clock input on a Devialet amp could improve sound quality. As would an i2s input...
*
microRendu 1.4::LPS-1::USPCB::Mutec MC-3+USB::250 Pro CI::Roon, Core on Intel NUC with Debian Linux, music on USB3 disk::Audiolense 5.5::Monitor Audio MA201 (thoroughly rebuilt, now with a SAM profile)


Reply
#23
Hi ogs, while a true reclocker may not be part of the design of the Experts an ASRC may well be. In fact the article below at the Qubuz website says Experts have a TI SRC4192 ASRC IC that upsamples all inputs to 24/192. When I look at the picture in the article it's actually a SRC4392, a higher specced ASRC.

The 24/192 sample&bit rate used to be the old official numbers spoken of in white papers, specs documents etc. I think that Devialet has later employed an additional upsampling step (to 40/384) perhaps for MAT (DSD support) or as a general improvement, in their eyes. Perhaps they've even circumvented the upsampling in the SRC4392 IC and nowadays do the upsampling inside the DSP (if that's even possible, I don't know Smile).

http://www.qobuz.com/fr-fr/info/hi-fi/ba...made175026
http://www.ti.com/product/SRC4192
http://www.ti.com/product/SRC4392

Of course the higher quality clock connected to the Mutec could help 'cleaning up' the input there locally, resulting in a higher quality output signal. As always, "garbage in, garbage out" so cleaning up as much as you can beforehand (before the actual DAC step) helps. This is also illustrated by the reports of people employing multiple reclockers in series, be it multiple Mutecs, multiple USB Regens or whatever reclocking device out there.

Nowadays I'd love to have a word clock input and I2S input as well but since it's never been part of the design I'm not sure if it could even be added afterwards by re-using a general input. I'm afraid we'd have to go back to mega buck separate DAC's and amplifiers again for that functionality. Wink

PS. This is the second discussion in a short while where I feel it would have been nice if Devialet had some representation here on this forum to clear things up. Smile
PS Audio P3, Shunyata ΞTRON Alpha Digital and HC/Furutech power cables, Paul Hynes SR7EHD-MR4, DIY Roon Server & Roon Endpoint running AudioLinux Headless, Phasure Lush^2 USB cable, Audioquest Diamond RJ/E ethernet, Mutec MC-3+ USB, Shunyata ΞTRON Anaconda Digital XLR AES/EBU, Devialet Expert 250 Pro CI, Nordost Tyr Reference LS cables, Von Schweikert VR-5 SE Anniversary Edition, Anti-Mode Dual Core 2.0, JL Audio Fathom F112. More detail here.

The Netherlands


Reply
#24
I took some internal pictures of my 250 when I got it back from upgrade. I noticed the SRC4392 so I assumed that they are doing the upsampling there. I got a somewhat confusing answer from Devialet on this:
"As you understand, a Devialet 400 is composed of two units of Devialet 200.
The Devialet 200 can upsample to 192kHz. So with a 400 (2x 200) you get a signal up sampled to 2x 192, meaning 384khz.
So, 384kHz over-sampling is not used in 120 or 200, but it is possible with Devialet 400 and 800.
I hope that will answer your question better.

Best regards,
Océane Turpin"

You may be right that Devialet does 384kHz in the DSP as my understanding is they now do 384kHz even in single amps, not just the mono blocks. Maybe the DSP takes the 24/192 output from the 4392 and deliver 40/384? But then where is DSD converted to PCM?
*
microRendu 1.4::LPS-1::USPCB::Mutec MC-3+USB::250 Pro CI::Roon, Core on Intel NUC with Debian Linux, music on USB3 disk::Audiolense 5.5::Monitor Audio MA201 (thoroughly rebuilt, now with a SAM profile)


Reply
#25
That answer by O.T. looks like it's completely made up. Smile

I'm thinking about sending Devialet a message to ask them about all steps an analog and digital signal (including USB) goes through from input to speaker terminals, including the transport protocols in between (I2S, S/PDIF or whatever) to better understand operations.

If they care to answer in detail I'll share somewhere on the forum.
PS Audio P3, Shunyata ΞTRON Alpha Digital and HC/Furutech power cables, Paul Hynes SR7EHD-MR4, DIY Roon Server & Roon Endpoint running AudioLinux Headless, Phasure Lush^2 USB cable, Audioquest Diamond RJ/E ethernet, Mutec MC-3+ USB, Shunyata ΞTRON Anaconda Digital XLR AES/EBU, Devialet Expert 250 Pro CI, Nordost Tyr Reference LS cables, Von Schweikert VR-5 SE Anniversary Edition, Anti-Mode Dual Core 2.0, JL Audio Fathom F112. More detail here.

The Netherlands


Reply
#26
It would absolutely be good to know the full signal path inside the amps. Thanks in advance for trying Antoine!
*
microRendu 1.4::LPS-1::USPCB::Mutec MC-3+USB::250 Pro CI::Roon, Core on Intel NUC with Debian Linux, music on USB3 disk::Audiolense 5.5::Monitor Audio MA201 (thoroughly rebuilt, now with a SAM profile)


Reply
#27
A few info back from our German colleagues : 

I tried to converse on the German forum, which was not that easy for me given my lack of mastering the German language ; Google Translate is your friend to a certain but limited extent :-). I am anyhow happy to share a few information I gathered there, especially from Fujak, the person who tested Mutec MC-3+ cascades.

- on the possibility to use an external 10 MHz clock ; Fujak has built his own external clock and stated that he was not able to witness any improvement with this DIY clock fed into the MC-3+ USB, compared to a cascade of 2 x MC-3+ USB devices. Fujak did not report being able to test a prototype of the Ref10 from Mutec anyhow. We will see what that brings when this new product will be available.

- on the cascading of 2 x MC-3+ USB devices ; Fujak strongly recommends to use BNC connections between the devices and to the DAC (better quality for the buck of BNC cables than AES ones). He uses Oyaide DB-510 cables to do that. Kujak made the precision that the two MC-3+ USB need to be configured with internal clock reclocking (using the clock signal from the first to synchronize the second does not help reach better SQ).

- on the upsampling question I was raising (24/192 before reclocking instead of 16/44 to the reclocker and upsampling at the DAC) : there are other reports by the German forum users that upsampling to 24/192 before a 24/192 DAC brings better SQ. There is an ongoing thread on this subject.

Hope this helps.
Reply
#28
Please find the report I filed on another blog re Mutec MC-3+ USB vs Acousence AFIS-USB
Quote:I am the happy owner of two Mutec MC-3+ USB, connected in cascade (via AES/EBU). Having read with interest the report of *progear, I approached Acousence with the intent to listen to one of their products. After giving financial guarantee to Acousence, Ralf Koschnicke, its founder, kindly sent me a test equipment that I intend to review here in comparison with the Mutec MC-3+ USB cascade.

Test configuration :
Music is streamed from Qobuz through Audirvana 2.3.3 on a Mac Mini (2.8 GHz dual-core i5, 2TB Fusion Drive). Music flow is upsampled to 24/192 on the Mac via iZotope. Mac is connected via USB to the Mutec cascade (Oyaide cable) and through Acousence supplied cable to the Acousence AFIS-USB.

Mutec cascade connected via AES/EBU to the Devialet D900 Master (Mogami Gold Pro cables). Acousence connected via AES/EBU to the D900 Companion (Acousence RJ45/AES-EBU cable). Loudspeakers = B&W 802d3 with Cardas Clearlight cables.

General impression : the musical flow seems smother to my ears with the Acousence AFIS-USB than with the Mutec cascade. Music seems even more natural, analogic. The soundstage width and depth is similar in both configurations. The localization of voices/instruments on the soundstage is at least as good with the Acousence if not slightly better, than with the Mutec cascade.

Listening to The Ballad of Bill Hubbard (Roger Waters, Amused to Death) : I am always using this track to make sure that I am sitting at the right place on my sofa, and that the loudspeakers are well in place. As expected, the dog is barking at 9:00 and the woman's voice is at 3:00. Compared to the Mutec, the width of the stage is at least as large, if not slightly larger.

Listening to Kyrie (Mercedes Soza, Missa Criolia), the separation between the guitar and the flute at 1:29 into the track is perfect ; the guitar is clearly playing from the center left, and the flute from the center right. Same remark as above ; the separation is slightly better with AFIS than with the Mutec cascade.

Listening to Dancing Barefoot (Patti Smith, Land, Digitally Remastered 1996), 
with Mutec cascade, percussions (especially the cymbals) are a little too bright, too present, making this record nearly difficult to listen to. 
With the Acousence, the percussions are still very detailled, but the music seems more natural, less surgical.

Listening to 'Un di, felice, etera - La Traviata' (Cheryl Studer, Luciano Pavaroti, Metropolitan Opera), the restitution of the voices is very good in both configurations. The position and presentation of voices appears more natural with the Acousence IMHO ; what is surprising is the position of the Tenor vs the Soprano, which seems to me better defined with the Acousence. The voice of the Tenor clearly comes from a position which seems to be on the right of the voice of the Soprano, as figured on the Album Artwork. With the Mutec, the respective positions seem more confused. Is this an artefact linked to my system or is it trully a better restitution of the sound track ?

Conclusion : as the prospect of getting huge benefit from a more precise clock seems limited in my configuration (the sound card of my Mac cannot be driven by an external clock, and my DAC is asynchronous), I tend to understand that the better isolation/signal regeneration of the Acousence is doing a better job in my configuration.

For what it's worth :-)
Reply
#29
Oh no, now I have to upgrade, again!! Wink

Thanks for sharing Our.

It's still not clear to me what this solution entails, boxes, connections/cables and how it works but I'll use google translate on the German website to make it more clear to me.

Edit; it's a shame you can't compare the Mutec to the Acousense using the same cables for both. It's not an apples vs apples comparison now.
PS Audio P3, Shunyata ΞTRON Alpha Digital and HC/Furutech power cables, Paul Hynes SR7EHD-MR4, DIY Roon Server & Roon Endpoint running AudioLinux Headless, Phasure Lush^2 USB cable, Audioquest Diamond RJ/E ethernet, Mutec MC-3+ USB, Shunyata ΞTRON Anaconda Digital XLR AES/EBU, Devialet Expert 250 Pro CI, Nordost Tyr Reference LS cables, Von Schweikert VR-5 SE Anniversary Edition, Anti-Mode Dual Core 2.0, JL Audio Fathom F112. More detail here.

The Netherlands


Reply
#30
(07-Mar-2016, 15:04)Antoine Wrote: Oh no, now I have to upgrade, again!! Wink

Thanks for sharing Our.

It's still not clear to me what this solution entails, boxes, connections/cables and how it works but I'll use google translate on the German website to make it more clear to me.

Hi Antoine,

Well, before upgrading, as Ralf Koschnicke from Acousence is a very kind man, very customer's satisfaction orientated, I suggest that you try to listen if you hear the same things than me.

Having said that, the system consists in two parts :
- a USB satellite, which is connected to the main box via 3 optical cables : one for a switch 44.1/48 kHz, one for the clock (from the main box to the satellite) and one for the sound (from the USB satellite to the main box). This satellite receives the USB signal, completely isolates it and transmits a data flow, synched with the main box clock, to the main box itself.
- a main box, which can have one input/one output (AFI) or four inputs/two outputs (AFIS).
The small box is not configurable (one USB input and one RJ45 etherconn output). You just need to chose a cable (AES/EBU or BNC) to connect to the RJ45 and you're all set.
The larger box is fully configurable. So you can chose which inputs/outputs you want.
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)