Hi Anthony,


Well of course - this is what I'm explicitly saying. But the how of it is beyond me, unless it can be related to "speed". And that the speed increased is overly clear.
How *that* could happen is indeed the big riddle. Make that a Quest.

By now I think two years ago, I bought the "cable analyser" (LCR) which is still brand new in its box. It could be time to unwrap it.
Still, my guts tell me that this is all not related to the USB protocol, but merely to how it works out at the DAC's end (probably even via the mains). Point is : I work with the shield and not with the data integrity or something, plus that the USB errors are zero always anyway (and e.g. a retransmission *if* it would exist (it does not) is not in order.

The only thing I might get from this myself is that the shielding (at least) as how it is now, protects the environment from what the cable itself does (emits). Thus, this should follow from the impossibility (that I can see) that the (protocol-)data is influenced by the environment.

If this makes sense at all, it should imply that people / we should focus on shielding from the outside.
(and now you ask How, right ?)

Regards,
Peter

Here's the topic about Phase Alignment.

All,

Because it was asked elsewhere, here a brief outlay of what Phase Alignment is/does in XXHighEnd.
It should not be used any more.
Right.
(plus I'd think I disabled it per 2.11 - not sure though)

Phase Alignment is dangerous.
Right again.

Phase Alignment is the most wild trick to let your electronics behave better.
Back at the time everybody used it (this is an XXHighEnd-only), BUT it was under control because it was a hot subject. Thus, everybody grew into it, and only one person blew all of his speaker drivers. So that was quite OK.
Right (not).
(and Phasure paid him the new drivers)

Those with a NOS1 were in luck because they could "trim" the Phase Alignment with their DC-Offset meters. That is, in real time. Not doable without such meters, but it could be pre-set as well with a multi-meter.

It is the most complex piece of software in there, and of course only a phool like me could make it.
So ...

So Phase Alignment plays with DC offset.
It anticipates electronics to behave worse with negative voltage than the positive voltage does. N.b.: Negative voltage is always a derivative from the positive "rail"; it is an extra step (with always additional THD). Btw, don't ask me whether some theory exists that negative can be as good as positive after all, but even if it can, nothing tells that this has been applied in ALL of the circuits in pre-amp, power-amp, and in the end also the DAC of course.

The best part of this whole trip is the avoiding of zero crossing.
Now imagine that each self-respecting PCM DAC (in general these are R/2R DACs) has inherent relatively "huge" problems with the voltage crossing zero, impeded by the "digital code" which causes this. Actually the biggest current (!) step is implied because ALL the bit but one, flip. Thus, those poor D/A chips must apply this when zero is crossed :
00000000 00000000 00000001
11111111 11111111 11111111
... thus because all the bits (but one) need to change state, this draws a relative huge amount of current, but only when the voltage crosses zero.
N.b.: The PCM1704 is quite OK with this zero crossing distortion, but still it measured significantly better on the THD. It also sounded better - and still should;

---

If it is not active any more, I could reactivate it. But the messages will be larger again. Point is : everybody should measure the DC Offset at the speaker input (hence power-amp output). Speakers won't like this ...
But also : I can imagine that not everything is compatrible with it, like for example the new Pause feature (in Unattended). And when something is not compatible, what happens is that the DC jumps (a big bang). This is not nice for your heart ...

Peter

Hi Anthony,

IIRC this was the very very first control I made in XXHighEnd (besides mentioned Q slider, which is today's Q1).
Btw, Q1 *is* speed (because buffer size) related, so ... who knows.

haha

But it never really did something for me (with someone guiding me who was sensitive to it !). Maybe I will give it a try again !

Thanks ...
Peter

There's an other effect becoming more and more clear to me as an exhibit of the LUsh^3 in general : Stereo phase works out for the better.

By now I am 100% certain that the more I listen in the middle opposed to the speakers (compared with more to the side(s)), the stronger the bass is. And not for a little bit.
At the same time, the same applies to the highs, or more spooky : the distortion on them.

In the middle all falls into place, say "infinitely" more than prior to the Lush^3; it's almost like the speed (which is clear audible if you'd ask me) takes care of phase angles of waves (from left vs right) now are less vague (thus more "straight" or robust or accurate), so a chance now exists that they meet how they should (and therewith emphasizing each other when in (good / exact) phase).
This is a totally new phenomenon for me. So ...

True, being/listening in the middle always has been a good thing for listening to music to (stereo) loudspeakers. However, for me this has never been about pinpointing of instruments (the imaging Ramesh also refers to) because this always has been all right to begin with, for me. But it is not about this at all ... it is about the waves meeting correctly at one and one place (position) only, and how that gives a sense of 6dB more bass. Or 6dB less distortion on the highs, if you want.

More strange you can't have this, for a USB cable which transports music via a protocol.
Maybe later we'll find out what's really going on ...

Peter

Maybe not unimportant : I changed my settings for the Mach III (see my signature and the ** at the changed settings).
The crucial one is the highering of the SFS from 10.19 to 20.69.

I deemed this necessary because the bass is now so "impressive" that with the lower SFS settings (and in my room) it starts to higher frequency buzz a little bit. I feel this is caused by too much energy which in itself is some kind of higher resolution workout of the PC (settings), them now let through by ... well ... a USB cable ?
Anyway, something like that.

Mani's lowering of the SFS clearly only emphasizes this mere negative behaviour. At least over here it does. But good that I tried Mani's suggestion, because otherwise I may have never tried to explicitly higher the SFS.
There could be a slight pinch of that energy left, but I only notice this when I am fairly much off-axis of the speakers.
Also see next post.

Peter

PS: But Ramesh, at least you should try this too, I think.

Haha Ramesh - Nice, very nice.

Yes, I don't have the feeling that you are mimicking my words while you (surely) describe exactly what I perceive of it myself.

By now (actually since day before yesterday) I have the very clear perception of that this is the Lush for its "analogue" behavior times 10 or so. In the beginning I was a bit afraid it could be a more digital sounding cable again (I told about it in an earlier post), but the contrary is true by far. Nothing sounds more analogue than what I perceive of it now WHILE (and this is the contradiction) there's so an enormous additional pile of detail all around. And not less highs either. But different. Eh "better" ?

Thanks a lot, Ramesh.
I hope others will take the effort of writing a bit about the Lush^3 - especially on the findings on the configurations so far.
Thanks !

Peter

Hi Vincenzo,

If you can select an other output than one prefixed with KS:, it will be WASAPI. At least, when XXHighEnd produces audible sound.

XXHighEnd does not have any other means for playback than "Bit Perfect" (and yes, KS and WASAPI are equally that).


So if I understand correctly, this should be WASAPI.
Please notice that WASAPI is an additional layer on top of Kernel Streaming. Now :

This is possible (sound sweeter) for very many reasons, all a little hard to explain in the context of your current knowledge of XXHighEnd (that will grow automatically).

One reason would be that you (most probably) did not optimize anything yet. This makes it a bit of a wild guess what will "be" "better" as such ("better" is not really subjective here - it is a measure of accuracy I am referring to).

An other reason will be that Kernel Streaming has many different (and more !) optimization parameters towards the OS itself, which btw are many more for e.g. Windows 10 vs Windows 7. Also, the Q1 parameter works very differently for WASAPI than it does for KS.

So Yes, both are/play Bit Perfect, but both will sound very different with in the end KS as a clear winner (you will see later). How this can happen while both are Bit Perfect is relatively easy to explain : XXHighEnd sounds way better than  e.g. Foobar, because, well, it is XXHighEnd. Foobar too can play in Bit Perfect Mode, but it sounds like ...
So there is a LOT more that can cause better or worse sound. But you wouldn't be able to measure it easily.

Kind regards,
Peter

Hi Vincenzo,

The best answer is No. Under the hood they still play a role, but it is not user selectable any more.

Otherwise this relates to WASAPI vs Kernel Streaming.
For best music quality you'd want Kernel Streaming. And this eventually (in the end, when you feel comfortable) combined with
1. Going to Minimized OS;
2. Use Unattented Playback (= no GUI).
That should be your base for further fine tuning.

Kind regards,
Peter

I now added it to the original post about this config.
Peter

Vincenzo, very well done !!
For the next time you will know it by heart !

Kind regards,
Peter

With some time and rest now ...  YES, that is correct !

Regards,
Peter

Haha. Why don't you show a photo yourself ?
(teasing a little, but I don't make photos at 6 am)

The jumpers are not related / in order for this latest config. They are only for protection (hence leave them off to avoid confusion).

Ramesh, One. And with MQA less I suppose.
But I only listen to albums when they are new. And then one is "logic" I think. Otherwise I only play from Galleries and I take care that no more than 60 minutes or so is loaded. With Hires/MQA less (say half of normal maximum).

Peter

Yes ...