A Confused streaming system - Mutec / SOtM Ultra

[ogs]

My experience on Roon AIR vs mR/MC3 is mostly the same as Confused. As I do not own a REF10 my setup differs from both Confused's and zdenes'. I have not yet modified my MC3 for external DC either, but plan to.
Roon AIR is extremely good, but mR/MC3 is still better. There is better transparency and dynamics with the MC3 inline. AIR has a very special relaxed or calm character that (I believe) the AES/EBU input does not have. I wish I could combine AIR's calm with the MC3 transparency. This can of course be many things as the signal goes through more cables and stages with mR/MC3. AIR has the advantage of a different, and I assume shorter and more direct signal path than AES/EBU. I also believe the CI clock scheme improves the overall sound for both.
If I did not own the MC3 +USB I'd probably be quite happy with Roon AIR today and I would perhaps not upgrade. In my case there is around USD2300 extra for microRendu 1.4, LPS-1 and Mutec MC3 +USB. I am happy I own them and I will certainly continue to use them.

[zdenes]

One thing I forgot to say. I, of course, tried all combinations. No SOtM, the computer straight to the Mutec MC3+USB or no Mutec, SOtM straight into the Devialet with or w/o the tx-USB ultra. Some of these combinations don't sound any better than AIR albeit all being more problem free when streaming.

I found that only when I have all of these units connected and also the Ref 10 with good power supplies and clock cables, the magic starts. Now, I added a second MC3+USB which brought some incremental but audible improvement. It is being converted to accept DC now.

I have 5 reclocking in my system (all connected to the REF 10) and each makes things a little better. In Munich, I asked Christian Peters - head engineer for Mutec - why several reclocking matters. (I think I already talked about it earlier.) He replied that when reclocking occurs with a cleaner (i.e lower jitter) signal, the reclocking puts less stress on the unit and can improve that signal even further.
I also asked about the exact mathematical formula to justify his claims but he would give it to me. ... No, I did not.

[Vivialet]

Hi zdenes,
as I can see from your profile you still have the 800. I think the new CI board makes a big difference for AIR. I also think that an SOtM bridge could further improve the SQ of my Mutec system. But driven directly from an optimized PC (AO, Fidelizer) Roon/AIR via Wifi is better. I have already used 2 MC-3 USB clocked by the REF 10, adding 2 more boxes which might be a little bit better again does not make sense for me. I really love the minimized setup I have now (which was the reason to buy Devialet initially). It could also happen that adding SOtM is sounding 1% better and as soon as Roon RAAT is supported by a future firmware release that this sounds 1% better again. And the Mutecs tend to cut off a second of music while they are synchronizing, especially if the format is changing.

[zdenes]

Hi Vivialet,
Yes, I still have the D800 and I totally agree with you. I originally bought the Devialet 8 years ago because I wanted excellent sound quality and a simple solution.
I was a late upgrader from D-Premier to D-250 and added another 250. All just before the new pro/CI upgrade program. I thought that this time (as with the previous upgrade) I would leave to the last moment and read some feedback here on the improvement if that is the best way to spend 7K.

Initial results were quite mixed on this forum, apparently, dozens of FW problems and new ones seem to surface again and again. Most people were happier with the SQ improvement but some people were not impressed. That was the time when I looked elsewhere and found here that people were using other streaming methods. I started to build my current system based on this.

Anyway, I waited further and see what the CI board brings. It seems to have improved streaming quality but FW problems etc remained. My dealer promised to loan me an Expert Pro in mid-June for a week so I will have the ultimate test myself.

I don't know if anybody else noticed that you were "cheated" to an extent. Devialet said that you pay whatever amount for the upgrade and you'll get the CI board free that would cost 1500 euros later. It never did/does/will. Upgrades will be available for the same price that all of you paid and sent the Devs to Paris twice.

[Confused]

@Hifi_swlon :

Reading you post with a little more time, I am intrigued with what you are planning with Thierry. So the idea is that you add PEQ to your HAF convolution, then send details of the PEQ to Thierry, to either include directly, or at least to give an indication of the change you require. Out of interest, is this just your idea or have you agreed this approach with Thierry? It seams like a good idea to me, I have been describing the revisions that I want verbally, obviously sending a carefully crafted PEQ is far more precise.

@zdenes :

First to say that the time I have spent on the HAF convolutions has for me been the best thing I have done with my system for years. Plus, it has not been that time consuming once the initial measurements are made. A few minutes to send an email to Thierry, and then it takes less than five minutes to download the convolutions and load into Roon, it is incredibly quick and easy. You mentioned at the start of one of your posts that your room is acoustically treated. I think this is the key, and I think this is where you have perhaps got it right and I have not. I am at the point now where I think I could benefit enormously from some carefully considered room treatments, I also need to reposition some of my kit, this means a new rack in a new location. All of this will take time. Meanwhile, the HAF filters have provided a substantial benefit. The bottom line is that I had stopped enjoying my system, with the HAF filters I am loving it, so time well spent!

Having said that, I am defiantly considering getting some Paul Hynes SR4's for the SOtM kit. I was reading some reports of the SR4 recently and it does seam to be an exact fit for what I think would be beneficial now. As for modifying the MC3+USB for LPSU input, I did hear a rumor a while ago that Mutec would be releasing an "Empyreal Class" version of the MC3+USB, so I was thinking that I would just get one of those. I heard this rumor a long while ago and something was supposed to happen for Munich, which it did not. So this looks like duff information now. SR4's for the SOtM kit looks like it should be beneficial and easy to do. To be honest, I was trying to have a pause in my kit buying, but the SR4's are very tempting!

I am also thinking of modifying the internal clock feeds for the SOtM kit, replacing the over long SOtM internal clock cable. This one should be both cheap and beneficial. One question - You mention that you have been trying clock cables, do you have a recommendation for the cable between the REF10 and MC3+USB? Everyone is raving about the Habst, which is fair enough, but it is the same price as the MC3+USB that it connects to. Although this seams crazy, it may of course be worthwhile as this is a very critical connection. (far more critical and sensitive than normal USB or S/PDIF)

As for your surround system connecting to the Mutec kit, this is indeed very interesting. Of course, a little off topic here, but this has got me thinking!

Meanwhile...... The system is sounding great now. So all the above is just icing on the cake!

[Vivialet]

One thing I did notice, is that with Beta 14 I seem to need a lot more gain.  I have not been paying much attention to volume levels recently but did today when doing a AIR versus SOtM / Mutec trial.  Volume levels (gain) seemed much higher than I have needed to use in the past.  For good order, I will note this in the 12.2.12 thread.  I might even try reverting back to Beta 12 to compare properly, but I am out of time today.

Hi Confused,
keep in mind that a room correction always needs some headroom for the correction, even if it does only decrease the level. For Thierries filters it will be the same, so maybe this is the reason you need higher volumes now. Dirac uses -6db by default, for mathaudio I needed 2.5db headroom to avoid clipping.

[Confused]

One thing I did notice, is that with Beta 14 I seem to need a lot more gain.  I have not been paying much attention to volume levels recently but did today when doing a AIR versus SOtM / Mutec trial.  Volume levels (gain) seemed much higher than I have needed to use in the past.  For good order, I will note this in the 12.2.12 thread.  I might even try reverting back to Beta 12 to compare properly, but I am out of time today.

Hi Confused,
keep in mind that a room correction always needs some headroom for the correction, even if it does only decrease the level. For Thierries filters it will be the same, so maybe this is the reason you need higher volumes now. Dirac uses -6db by default, for mathaudio I needed 2.5db headroom to avoid clipping.

Yes indeed.  Thinking about it, I have been tweaking headroom and checking for clipping using the "Limited" count in HQPlayer, this started when I was running Beta 12.  It also occurs to me that I have seen the HQPlayer Limited count being triggered and the quick fix is to tweak the HQPlayer volume control down a touch.  Thierry had advised that using the HQPlayer volume control to avoid clipping is OK, but I can't help thinking that is best to set the correct headroom level in Roon DSP, so the convolution and headroom correction are both in the same place.  It should only be a five minute job to fine tune this, I will give it a try, it might even offer a small hike in SQ.

[Dan]

Confused - sounds like you have the HAF filters in ROON convolution. Have you tried them in HQP? I tried both a while ago and thought the SQ was best with them in the PIPLINE window of the Matrix portion of HQP. Thierry gave me filters for both ROON and HQP (since they are different for HQP). Just thought I'd throw that out there. I'll have to go back and compare them again.....

[Confused]

Confused - sounds like you have the HAF filters in ROON convolution. Have you tried them in HQP? I tried both a while ago and thought the SQ was best with them in the PIPLINE window of the Matrix portion of HQP. Thierry gave me filters for both ROON and HQP (since they are different for HQP). Just thought I'd throw that out there. I'll have to go back and compare them again.....

No, I have only tried the convolutions in Roon.  When I started my HAF journey I asked Thierry if using the filters in Roon or HQP was best, and his advice was "It is much easier to have the filters within Roon especially if you listen to files at various sampling rates." 

One thing I can say it that using the convolutions in Roon is very easy, and it is also very quick and easy to switch between different convolutions.  This is very useful to me at the moment when I have a number of slightly different convolutions, with Roon offering a quick switch for comparisons.

[Confused]

It has been three weeks now since my last major update to this thread.  Nothing much has changed in my system, I have simply been using and enjoying it.  Enjoying it quite a lot as it happens.

A couple of weeks ago I decided to revisit HQPlayer settings.  It was here I made a bit of a minor discovery.  I had been leaving my HQPlayer settings alone whilst I tweaked other things, as it happened HQPlayer was set to poly-sinc-ext NS9.  (not poly-sinc-xtr as I mentioned earlier in this thread)  It turns out that this might be significant because poly-sinc-ext upsamples to 176.4kHz.  With a bit of playing about, I found that a few of the other HQPlayer settings sounded better, specifically those that upsample to 192kHz.  Speculating a bit, there is some logic to this.  When HQPlayer is upsampling to 176.4kHz, the SOtM kit and the Mutec MC3+USB are all working their clocking magic producing near-perfect 176.4kHz feed, only for the Devialet to then have to do do the last conversion to 192kHz.  Running poly-sinc-shrt, poly-sinc-xtr and others, HQPlayer is upsampling to 192kHz, so the SOtM kit and the Mutec MC3+USB are all working their clocking magic producing near-perfect 129kHz feed, and the Devialet can simply accept it, with no conversion of the clock feed via AES3.  So that's the speculation.  Today I tried another Roon AIR versus SOtM / Mutec head to head.  This time it was far easier to decern the benefit of the SOtM / Mutec front end.  Don't get me wrong, Roon AIR is still very good and we are talking small margins here, but whereas three weeks ago I was thinking was all this worth the hassle and money, today it is a case that I would not want to part with the Mutec / SOtM kit, it is definitely the way I prefer to listen now.  It is better in all sorts of ways, subtle details, bass accuracy, punch, realism, and yes, spectacular treble.

Another curious thing is that I have on occasion been listening with Roon's DSP engine disabled, and I am now finding the HF far sweeter with the SOtM kit than I did a couple of months ago.  In fact, switching between Roon AIR and the SOtM kit, it is definitely AIR that has the fractionally harsher HF.  I am as convinced as I can be that the SOtM kit is simply improving, although it has taken a long time and a lot of hours to get here.  Yes, I know some will say it is brain burn in, I am just getting used to it.  OK, maybe, but the thing is that the comparison with AIR is now different, with AIR sounding just as bright, if not brighter, than the SOtM kit.  As an example, at the time I first had the SOtM kit someone had given me a Goldfrapp album, I recall that a couple of tracks on this album sounded dreadful, overly bright with jarring presence range / HF.  Today, with all DSP off, it sounds fine.  I know that aural memory is fallible, but the difference between playing a track and thinking "this is borderline unlistenable" and then next time (a few months later) thinking "this sounds great" has to mean something.  Curious.  To be honest, a lot of little things have changed, multiple firmware updates have certainly changed how things sound with AIR, and of course, may change how things sound with other sources.  Maybe AIR sounds more like the SOtM kit now, hence the closing of this apparent gap.  If I can ever be bothered, I might try switching the microRendu back into the system, just to see if this still sounds darker versus the SOtM kit, it would be interesting to know.  I am convinced as I can be that the SOtM kit has improved, so a revisited head to head with the mR would at least provide some interesting perspective.

Meanwhile, things are sounding great, which is good.  So has this removed the desire to tinker?  Actually no.  Three weeks ago part of me was thinking that I had had enough, sell the Mutec and SOtM kit, just enjoy the music with AIR.
Today, I was enjoying the SOtM and Mutec kit so much that it seeds the "this is great, can it be improved more?" thought. There are a few ways that I can think to improve things a touch more, but no rush though, things are back to being enjoyable now.