Has anyone heard the Devialet D-Premier Integrated Amp/DAC

[amirm]

No they were not close at all. The Davialet was sitting 3-4 feet away outside the cabinet which the Berkeley was inside on a lower shelf.

I must say, if we did not have such easy access to A/B so many amps and speakers, I am not sure I would have arrived at the same distinct conclusions.

Secondly, our frame of reference (Mark Levinson) is undoubtedly different than other comparisons people are doing. That said, what we found tracks the engineering design. Class D and its variations do excellent job of pumping out lots of power and this amp did the same thing. And switching power supply as likely used in this small package, is liable for performance degradations at the higher frequencies. As is a low switching frequency itself (this one is a guess on my part).

I have yet to hear any class D amp perform superb other than the 53 in high frequencies. Prior to this testing, we had tested another amp which also suffered in the high frequencies and losing to the Crown actually.

Again, it is entirely possible our subjective testing is invalid so I look forward to hearing feedback from others and what comparisons they have performed.

 

[amirm]

I should add that the unit is dead quiet. With my ear glued to the tweeter, I could not hear a thing. Many class D amps can be noisy. This one is not.

 

[amirm]

Did a bit of searching and seems like its switching frequency is 300 Khz. The Crown I believe is 350 Khz. The Mark Levinson 53 in contrast, runs an effective frequency of 4 Mhz. *

* The configuration is different in Davialet although I think the switching frequency still has the same effect.

 

[microstrip]

Amir,

Could you specify the system (speakers, source and cables, including power cables ) you used for this test?

 

[Orb]

....
Feed was a high-quality USB to S/PDIF converter (Audiophilleo) from a music server and analog out from our Berkeley alpha DAC. Most of the tests were done with the former, feeding the Devialet with digital signal. Speakers were Revel Salon 2s and Paradigm Signature S8s.

.....
To wit, both No 53 and Crown provided the same tight and powerful bass. The 532 while also quite good, had slightly softer impact in its bass response.

- Unfortunately, once we move above the bass, as with most class D amps, performance starts to drop. The highs were especially grungy with Devialet. Ambiance was lost and sound become flatter.

- Wanting to separate the DAC performance, we ran the unit using analog connection from our Berkeley DAC. It made very little difference in the performance of it.

To provide overall context, the 53 was hugely in front of it in performance. It was not a fair race at all between the two. The 53 performance stays at top regardless of what you throw at it, while also having the benefit of incredible bass performance.

The 532 was also beautifully nice sounding all around, only beat on tightness of bass by Davialet.

I should mention that with the right material this device can sound incredible. Indeed, that was my first experience with it: a heavy drum piece by the Marsalis family. It was everything you wanted it to be. Also, this device has a customizable feature using an SD card. We tested it with the default program delivered to us. Perhaps with a different config file it would perform better.

Again, this was all sighted testing and it is entirely possible we were biased by the much larger boxes connected to speaker wires .

One idea here might be to use this system in bi-amp configuration, with another amp driving the mid to high frequencies. Since not as much power is required there, a tube amp or small solid state amp will do the job. In that sense, having a high-pass output from the amp would come in handy.

Hehe ah well got to love how audio throws different perspectives on what we hear
Funny thing is that I thought the highs on the Devialet were one of the best digital I have heard lol.
To the point using ringradiator tweeter based speaker is crystal clear without any twinges-annoyances I usually pick up from quite a few other products.

How did the Devialet sound in comparison to the Crown Amir?
I may not be too surprised about the ML no53, considering that they are over double the price of the D-Premier and that excludes the preamp-DAC.

I was going to mention but thought you were having the D-Premier for awhile so was going to wait, blast but some things I notice.
The Devialet sounds subtly better from digital XLR, also it seems it can pickup on errors caused potentially by cabling-products.
So curious, did you manage to use the D-Premier without a USB-S/PDIF converter for the digital input?
The analogue input ideally needs to be configured as you can set sampling rate/impedances/MC-MM/can be configured to match cartridges/and more.

So it is usually simpler to stick with the XLR or S/PDIF digital (without converters), unless wanting to hear the analogue source which ideally should be configured to ideally match.
Agree this is a right pain, but most purchasing the product will be coming from digital background I feel.

I must admit I am still a bit leery about USB unless the product that interfaces to it has good isolation implementation such as Ayre - Paul Miller and others at HifiNews have identified many current USB products that do not isolate properly.
Probably why Devialet suggest at moment ideal conversion is using firewire with mac through Weiss Int202, no such issues can occur.

So some differences experienced between us but wondering if it comes down to the usb-S/PDIF converter and/or preference on the digital filter-DAC side.
One thing I have asked Devialet to do, is create more filter-DAC settings as my preference is subtly towards some of the other products I have owned - stunning measurements if you see Paul Miller's results but extended listening I have found subjectively I prefer a slight different take.
That said the D-Premier is a stunning product but does not mean it is comparable to a system over 4x its price, which some may be hinting that it is the replacement of all things and it is not IMO.

Still quite surprised about your findings on the actual amp side, as to me it is a perfect balance between the best Class A/B and Class A amps, while having nothing in comparison to other Class D amps (owned a few).
In reality the D-Premier is technically a 5w Class A amp - this is directly coupled to the speakers while it is the Class D implementation that provides the current.

Just checked if I am correct:
Paul Miller mentions this:

In practice when we measure and listen to the Devialet, it's the performance of this very linear Class A control amp that defines both the numbers and its sound

Anyway we got to smile as thats audio and why we dont buy the same products; one man's siren is another persons screechy harpy

Cheers
Orb

 

[Orb]

Ah Amir thought of another thing and this relates to its 200w capability.
Can you remember when using the digital input what the volume number went down to while listening?
I am asking because the size of your demo room may also meant that the D-Premier could had been soft clipping protection to strong clipping protection on dynamic swings for the Class D aspect.

Thanks
Orb

 

[garylkoh]

One idea here might be to use this system in bi-amp configuration, with another amp driving the mid to high frequencies. Since not as much power is required there, a tube amp or small solid state amp will do the job. In that sense, having a high-pass output from the amp would come in handy.

Unless the tube amp and the Class D amp has exactly the same group delay between the input and the output, you will likely make a mess of the phase coherence between the woofers and the midrange - and suffer the ire of the speaker designer.

 

[amirm]

How did the Devialet sound in comparison to the Crown Amir?

They were neck and neck! You would think they were built by the same company. And this was the Crown with ADC and DAC loop (there is a Crown without but we didn't have that one). To be fair, occasionally I thought the Devialet had more resolution -- the same feeling I get listening to No 53. Hard to describe but it is low level detain in lower frequencies.

I may not be too surprised about the ML no53, considering that they are over double the price of the D-Premier and that excludes the preamp-DAC.

Agreed. For us it was validation that indeed it took all that extra iron to make a class D/I amp perform uniformly well.

I was going to mention but thought you were having the D-Premier for awhile so was going to wait, blast but some things I notice.

The box is in very high demand for dealers who want to try it. So it will be leaving our shop end of this week.

The Devialet sounds subtly better from digital XLR, also it seems it can pickup on errors caused potentially by cabling-products.

We didn't have an AES source to use with it so that might be the case.

So curious, did you manage to use the D-Premier without a USB-S/PDIF converter for the digital input?

Well, kind of. We went through the Berkeley and then analog out to it. As I reported, it didn't improve the things that worried us.

The analogue input ideally needs to be configured as you can set sampling rate/impedances/MC-MM/can be configured to match cartridges/and more.

Right. Hence the reason I left the door open for maybe some improvements due to that. FYI ours shipped with an SD card that tailored that port to Phono. We took out the card to make it play normally.

I must admit I am still a bit leery about USB unless the product that interfaces to it has good isolation implementation such as Ayre - Paul Miller and others at HifiNews have identified many current USB products that do not isolate properly.

I have had superb experience with the Audiophilleo. It has improved the performance of every DAC I put in front of it. OK, that is only 4 different DACs but you get the picture .

We have the Berkeley on order. In four weeks or so, we should know where that stands. For now, I have gone from not liking USB to a total believer. Using USB, you can completely remove the clock from the source. So on paper it is superior to S/PDIF because the latter still relies on the source and cable to relay timing.

So some differences experienced between us but wondering if it comes down to the usb-S/PDIF converter and/or preference on the digital filter-DAC side.

I don't think so. The signature is that of class-D amp with switchmode power supply. It is the same problem we heard with another unit we had. And the uncanny resemblance to the sound of Crown.

if we had a different tonal difference, I would agree with you. But the effect was so similar. I think DAC performance is being held back by the amplification.

Still quite surprised about your findings on the actual amp side, as to me it is a perfect balance between the best Class A/B and Class A amps, while having nothing in comparison to other Class D amps (owned a few).

I am dubious of the use of "class A" in describing the design of this amplifier. You can have such an output stage but as long as you pump the output of a switchmode power supply into it, all bets are off. You create something that is far more akin got a standard class-D than anything else. Granted, I am just going by a sentence or two on the design of this amp but the guess seems to match what we hear.

In reality the D-Premier is technically a 5w Class A amp - this is directly coupled to the speakers while it is the Class D implementation that provides the current.

The current is going through that output stage or else you would not get 200 watts. I don't see how the combination all of a sudden makes this a different beast. Perhaps there is some improvement due to this configuration but I am not hearing it.

Paul Miller mentions this:

Funny you mention this. I forwarded my assessment to one of our guys who missed the evaluation last night. As you, he said what we heard can't be true and quoted reviews praising the high-end as Paul has. I asked him to go and listen. This morning he came and quickly walked away with the same impression. Note that I was not there and he did his own testing.

Anyway we got to smile as thats audio and why we dont buy the same products; one man's siren is another persons screechy harpy

Very well said . I wish sometimes that we could do such testing together. It would be much more productive. I usually get 4-5 people at work but getting more would be even better.

I should also say that I am learning a ton in these comparative tests. I don't know what I would have concluded if I just had one of these three amps for comparison.

All, all, let me say that this is a valiant attempt at building a new, innovative product. For that, they deserve a lot of credit.

 

[amirm]

Ah Amir thought of another thing and this relates to its 200w capability.
Can you remember when using the digital input what the volume number went down to while listening?

No. I recall. I did not pay attention to the number as it is on the top of the display and not visible sitting in the right spot. This is one usability ding against the machine. The display should be in front, not top. If I were to guess, I would say we were at -5 db or so.

I am asking because the size of your demo room may also meant that the D-Premier could had been soft clipping protection to strong clipping protection on dynamic swings for the Class D aspect.

Thanks
Orb

The amp is incredibly powerful. I doubt that it was breaking a sweat driving either one of our speakers. But I can't rule out what you say.

 

[Orb]

Thanks Amir for responding.
I did a bit of checking last night and a few differences came out, but going to start with the digital volume as some may in their large homes be hitting the soft protection, especially as the output can be configured anywhere between 165watts to 240watts (not default) these days.

From what I understand and this is using the digital inputs;
When the volume is 0db on the display, this equates to maximum output when it receives 0dbfs recorded music.
Anything exceeding this is compressed up to the 1st 15db.
After that it then is soft clipped.

One thing I feel that would also skew the quick comparison is the output impedance of the ML compared to the D-Premier, not a criticism just that on quick swaps this may show a difference and skew ones perception.
While the Devialet is near as close to 0ohm as really possible and completely linear over the power and FR (0.005ohm to 0.006ohm; 20hz to 20khz), the ML while exceptional for a Class D amp still changes and not linear (0.005ohm to 0.29ohm; 20hz to 20khz) and at higher frequencies this increases more rapidly.
This will affect the FR depending upon what speaker is used.
Of interest is also how distortion will change in a non-linear fashion based upon power and FR, however unlike normal Class D using interleaving technology it goes down.
So the ML has 0.25% distortion at 20khz with 10watts, but at 400watts its only 0.03% at 20khz.

I do want to say I am impressed by the ML as IMO it is also an engineering feat like the D-Premier, but comparing them is tricky due to the non-linear behaviour of it, which on quick comparison will skew perception (same way as quick comparing two speakers where one has more presence in mid range if played 1st or more familiar with it).
But saying this, I appreciate some of your core skills did involve listening tests so this could be me making excuses for the D-Premier

Relating to the USB, I still would had felt more comfortable with your review if you decided to use a CD transport with XLR or S/PDIF as a baseline in comparison to the USB.
This possibly leaves it impossible to know for sure what was going on, bah shame your ML SACD player had not arrived (thought I read you were possibly getting one)

Relating to the D-Premier current dumping, I will do that in a different post.
But one quick point is that the Class D must go through the Class A amp as this filters the usual Class D ripple noise-etc.
Without it being fed-controlled by the Class A stage it would need the traditional Class D filter between it and the speaker, which is part of the usual non-linear behaviour we see from Class D and seen in partial by the ML and a lot by other products such as Nad/Rotel/etc.

I still feel the possible difference we are hearing comes back to the connection into the D-Premier, the imedance non-linearity, and maybe the compression-clipping protection of the digital processor in the D-Premier.
But this is only because I am yet to hear what you did from the D-Premier - although I have only heard it on 3 speakers, 3 transports, no USB product fed into it which might be a weakness if it is compromised from upstream source.

But USB, while many feel they have good results it is still something to consider when PM can and has shown jitter fluctuations caused by various source servers/laptops/related components , and even different USB cables, while two other colleagues have noted noise can be introduced further down the chain by even accepted high end products that exist today
Would be nice if we get could Charles Hansen's view on why they went the whole hog and implemented a thorough isolation in the USB DAC, but this is probably more suited to some of the other forum threads here.
Might be for different reason, but could possibly be for the same, and why I would be more comfortable if you had the chance of a CD player-transport to be sure.

Definitely interesting anyway,
bah you really sure the Crown is comparable to the ML

Cheers
Orb

 

[Orb]

No. I recall. I did not pay attention to the number as it is on the top of the display and not visible sitting in the right spot. This is one usability ding against the machine. The display should be in front, not top. If I were to guess, I would say we were at -5 db or so.
.......

Heh I can think of a few more ergonomics myself, it looks classy but I am going to be miffed if the control becomes lost or breaks as there is only 1 button on the D-Premier, must make sure batteries are close at hand

One good thing though is that the display can automatically change orientation for when mounted on a wall,etc.
But it does look pretty I guess lol.
Still, that volume remote control is stunning to use, I was dubious of it from initial looks.

Cheers
Orb

 

[microstrip]

No. I recall. I did not pay attention to the number as it is on the top of the display and not visible sitting in the right spot. This is one usability ding against the machine. The display should be in front, not top. If I were to guess, I would say we were at -5 db or so.

I had the some reaction until I noticed that the Devialet should be stacked against a wall, as a picture ...
It looked really nice in my room, hanging near the big painting shown in my avatar.

 

[microstrip]

(...) Still, that volume remote control is stunning to use, I was dubious of it from initial looks.

I was a little disappointed when I found that the remote used a RF carrier - it can not be integrated in the usual normal universal IR remotes or duplicated to create a backup unit. But I was told by a friend that maybe some people who are used to program garage door RF remotes can duplicate the basic functions in a cheap remote. Or perhaps Devialet can supply the codes.

Just before you ask, the Devialet did not sound better than my ARC tube gear and I was just borrowing it for some days. But it sounded really good, although very dependent on speakers , signal, speaker and power cables. If I had to demonstrate power cable effects it would be one of my picks to assure success.

 

[Orb]

I was a little disappointed when I found that the remote used a RF carrier - it can not be integrated in the usual normal universal IR remotes or duplicated to create a backup unit. But I was told by a friend that maybe some people who are used to program garage door RF remotes can duplicate the basic functions in a cheap remote. Or perhaps Devialet can supply the codes.

Just before you ask, the Devialet did not sound better than my ARC tube gear and I was just borrowing it for some days. But it sounded really good, although very dependent on speakers , signal, speaker and power cables. If I had to demonstrate power cable effects it would be one of my picks to assure success.

And this makes it interesting in some ways, because with it being so linear in performance and totally agnostic compared to other audio gear it makes one wonder how it becomes so dependant upon specific speakers,cables.

I do not disagree BTW as I have experienced this, just that it is interesting how it can highlight deficiencies or irritations in some products that were not noticed before, even when it comes to speakers (I ended up changing mine lol - slightly amusing but gutting as well).
While ironically it should be the more non-linear gear that highlights dependancies as this actually affects their performance.

I know a lot is mentioned about the ADH, but from what I remember it has 4 patents relating to various technology implemented, one of those is the ADH and I think one is related to the power supply-power transformer.

Thanks
Orb

 

[amirm]

One thing I feel that would also skew the quick comparison is the output impedance of the ML compared to the D-Premier, not a criticism just that on quick swaps this may show a difference and skew ones perception.
While the Devialet is near as close to 0ohm as really possible and completely linear over the power and FR (0.005ohm to 0.006ohm; 20hz to 20khz), the ML while exceptional for a Class D amp still changes and not linear (0.005ohm to 0.29ohm; 20hz to 20khz) and at higher frequencies this increases more rapidly.

I have not seen those measurements for either product. I do have the damping factor for Mark Levinson at 20 Hz. It is stated at 8,800: http://www.marklevinson.com/downloads/ML No53 Technology Background V5 04032010.pdf

It doesn't state a load impedance for that. Assuming it is the standard 8 ohms, the output impedance would be 0.0009. At 4 ohms, it would be 0.0005. So their numbers are an order of magnitude lower than what you list. Putting that aside, as you know, output impedance is reliant on amount of negative feedback so maybe 53 has less of it.

This will affect the FR depending upon what speaker is used.

It might but I will take a level difference at high frequencies over high distortion any day of the week .

Of interest is also how distortion will change in a non-linear fashion based upon power and FR, however unlike normal Class D using interleaving technology it goes down.
So the ML has 0.25% distortion at 20khz with 10watts, but at 400watts its only 0.03% at 20khz.

Those are THD+N specs, not just THD. So the higher numbers may be noise related. Regardless, the graphs for these amps is rarely as pretty as a simple linear amp.

Relating to the USB, I still would had felt more comfortable with your review if you decided to use a CD transport with XLR or S/PDIF as a baseline in comparison to the USB.

We used another source I forgot to mention: a Kaleidescape video player with its S/PDIF output. I would say that is representative of what a typical transport would do, albeit, not an audiophile one. The results were identical. Indeed, that is where I first heard the problem and quickly swapped out sources to make sure it was not the K's quality that was causing the problem. It was not. We played the exact same content and got the exact same distortion from the USB.

Regardless, I look for uniform performance in high-end products. They need to be competent across the board. For a DAC/AMP to not like one source so much would rule it out for me completely. That is the world we live in. I have tested that converter with everything from little $1,200 PeachTree Nova to $25K Mark Levinson 512 processor and it always provides the best quality of any other connection. We also get the same superb improvement with Berkeley. So if another device can't handle the identical source well, is not good.

Without it being fed-controlled by the Class A stage it would need the traditional Class D filter between it and the speaker, which is part of the usual non-linear behaviour we see from Class D and seen in partial by the ML and a lot by other products such as Nad/Rotel/etc.

I see a lot of design improvements on paper. The NAD is a good example of using post filter negative feedback to get around deficiencies of output filter. Question is, does it really work? Mark Levinson solves the problem by pushing the effective switching frequency way, way up to 4 MHz. That makes the filter's job exceptionally simple. They have patented the process so it is hard for others to duplicate them.

I still feel the possible difference we are hearing comes back to the connection into the D-Premier, the imedance non-linearity, and maybe the compression-clipping protection of the digital processor in the D-Premier.

We did not clip the unit. As I said, we operated it slightly below reference ~(-5 dbfs) as we do with our other amps. We heard no compression at all. As I noted, that actually was one of the strength of the unit in really driving the speakers well.

But this is only because I am yet to hear what you did from the D-Premier - although I have only heard it on 3 speakers, 3 transports, no USB product fed into it which might be a weakness if it is compromised from upstream source.

Were you able to do a side-by-side amp comparison like we did? We had 30 second switch-over time which while still long, did allow us to make reasonable comparisons. Once we knew the weakness, we also picked more material like that (e.g. avoiding tracks with strong bass response). That made the issue more apparent.

bah you really sure the Crown is comparable to the ML

Cheers
Orb

In bass performance, it was. It also beat out the reference 532 in that regard. But once someone sang and the highs came, you could see that it couldn't hold its own against the much pricier equipment. That said, both use the same class I approach so putting it in the same sentence is not so bad.

We actually chatted about selling the Crown at retail. Believe it or not, today we use it to power in-ceiling speakers for background music! Their built-in DSP is good for that.

 

[Orb]

Thanks Amir.
Regarding the measurements I guess we will have to disagree
But for sure I can understand what you say, and ah cool so the bases were covered with using a transport - thats good to know so no worries there then with USB being the issue

The 0.3ohm at 20khz was an 8ohm load, and then climbs very fast after that if just looking at 8ohms and higher FR..
The purpose of my mentioning the distortion figure was well was to show that due to fluctuating with both power and FR as per other digital amps, this exacerbates the situation depending upon speakers and music as both are not consistent in terms of impedance and spectra.

As I mentioned it is far superior to any other Class D amp I have seen measured, but as an example how this causes fluctuations check out the Nad M2 output impedance measurement.
http://www.stereophile.com/content/nad-m2-direct-digital-integrated-amplifier-measurements

Sadly a really bad example for its measurement is the Audio Research DSI200 (I hope the DS400 digital power amp is very different).
Bear in mind that all digital amps will suffer some form of this, but coming back to current dumping concept similar to the Devialet it is handled in a different way.
It is the only way to truly overcome the digital issue, if going by certain papers:
Near the end and conclusion touch on the aspect in digital -
http://quad405.com/jaes.pdf
However it also highlights the challenge of a feed-forward error correction/current dumping concept, which in essense the Devialet is from what I can tell - whether it is truly successful may be what your picking up on that some of use are not *shrug*.
Is it perfect, I doubt it but it provides a "perfect" linear performance compared to what is capable from digital on its own, even in an interleaved solution.
Basically the Class A controls and feeds the Class D, so "technically" (being pedantic it is a control-measure-error correction loop with both feeding in parallel to the output amp to the speaker) it is this Class A that controls what we hear and the system design that ensures best linear measurements so far seen by any amp at HifiNews.

Regarding switching for comparisons, you got the advantage on me there for quick swaps.
So I focus on doing extended listening non-stop of 3 hours with one product every day for a week and then swap but keeping to a 3 hour listening window for the next product before changing again possibly a few days later and observe my listening habits - preferences, traits that start to show,etc.
After a certain amount of time certain habits-perceptions from specific products creep in, and might seem painful stick with the same material always that is a mix of classical and acoustic, electronic.
But then I am not looking to review these products and more interested in what fits my preference and or not.
For me the strength in the Devialet seems to be the actual amp technology, it seems it is truly a perfect balance between the usual lean/cold but linear class AB, the organic subtle warmth of Class A, and without the niggles (artifacts, grain,harshness) of Class D, with the many extended listening sessions I have done on all those types.
Before anyone comments please not I am just generalising to give a feel of what I mean.

Anyway I appreciate it seems nothing was wrong with the D-Premier/how used and for you guys at work Amir it seems to be off or weaker in the upper FR.
I might as well post this as well, below is part of the patent covering the ADH design that is similar to the current dumping concept (as seen in the earlier paper) but subtly changed to overcome some of those challenges.
http://www.patentgenius.com/patent/7545212.html

Sadly the presentations by the engineers are no longer online at a site I used - they were really interesting as well booo, well interesting if like look under the bonnet I guess

Thanks again Amir for your headsup, definitely something that others will now look out for and should try to, and IMO this is a good thing for those auditioning.
Cheers
Orb

 

[amirm]

Thanks from me also on the extra data Orb. I want to repeat that subjective evaluations are never conclusive but simply background data for others to use in their evaluations. No doubt this device measures very well. It is a quandary then. Wish we had more time with it so that I could it put through my test bench and such.

That M2 measurement is something else. I was curious how it did after hearing their marketing pitch at CES and didn't realize there was a review. Thanks for posting it.

 

[fas42]

And this makes it interesting in some ways, because with it being so linear in performance and totally agnostic compared to other audio gear it makes one wonder how it becomes so dependant upon specific speakers,cables.

I do not disagree BTW as I have experienced this, just that it is interesting how it can highlight deficiencies or irritations in some products that were not noticed before, even when it comes to speakers (I ended up changing mine lol - slightly amusing but gutting as well).
While ironically it should be the more non-linear gear that highlights dependancies as this actually affects their performance.

This is telling me there is a key weakness in the Devialet, most likely power supply related, that is making or breaking it when auditioned. I would still suspect that some other component powered up while amir was testing is the problem, if he tried it in an electrically very quiet environment I would say its performance would be much better.

This is again a good example of the closer you get to the sound being just right, that one little, tiny thing can throw it off dramatically, and cause different people to get completely different impressions ...

Frank

 

[amirm]

Addressing Frank's point, we have commercial 3-phase power. I know that power conditioning does improve amplification sound in that setting as we temporarily tested our Mark Levinson 532H that way. We did not however use that in this comparison. And of course, the other gear is performing as it did in the comparison without any help.

 

[Orb]

This is telling me there is a key weakness in the Devialet, most likely power supply related, that is making or breaking it when auditioned. I would still suspect that some other component powered up while amir was testing is the problem, if he tried it in an electrically very quiet environment I would say its performance would be much better.

This is again a good example of the closer you get to the sound being just right, that one little, tiny thing can throw it off dramatically, and cause different people to get completely different impressions ...

Frank

Possibly Frank, I know that my mains that is really dire does affect my sound in a subtle way; it even affects one of the transformers (slight buzzing) in my transport and also other none audio products, but focusing on the audio I have not tried to isolate how/which component is affecting audio.
That said it is still an excellent sound and only marginal differences to when the mains is fine, even with the issue it is as good as other SOTA high end products I have had in my house - up to 2.5x its price anyway and putting preferences aside.

But in my case when talking about cables with the D-Premier these were speaker and interconnects not power cables, so it is possibly a strength in terms of transparency rather than weakness, although I appreciate if seperating power cables from the rest it might be what you say.
Still the power supply is incredibly well regulated compared to many other amps tested, but what you speculate may have weight.

Personally I feel what I would like tweakable is the DAC-filter settings (if possible) such as offered these days by quite a few of the other high end manufacturers that has their ultimate measurement setting and then alternatives.
For me it is this that some I think will find possible preference challenges with - to a lesser extent it includes me as I prefer a slightly more "subjective" DAC presentation , even though digital measurements are again quite incredible.

In terms of digital distortion against digital signal level, again the D-Premier has stunning measurements which at -60-80dbfs is even much better than most other high end SOTA digital components, as an example quite a lot better than products like the Moon750d, the top Krell and ML SACD players,etc.
Whether this translates into better audibility it might be more academic.

Anyway I do feel in some instances the D-Premier is presenting when listening certain limitations in other products in the chain.

Thanks
Orb