0.9y-1d ... wow!

[Calibrator]

Earlier this week I had some exposure to version 'y' but that was more at an arms length, ie. listening out of the sweet spot to one side, and I was more interested in troubleshooting the niggling bugs that had been introduced.

Well, today I had a chance to sit down and spend many an hour in my comfy recliner and critique 0.9y-1d in earnest, listening to a variety of albums I am very familiar with. I'm still using the same parameters I have for months now, and feel no need to change.

I now have to ask one question though:

Peter, what the hell have you done to the program to make it sound like it does now?

The level of control and articulation I am hearing now is remarkable! I thought SQ was pretty darned good with the last of version 'x', but this is on a whole new level. Everything has been tightened up, with the bass registers exhibiting the most startling improvement. Stanley Clarke is one of my favourite bass players, and to hear "Illegal" from his "East River Drive" album is a joy to behold. I'm hearing the nuances of his finger work on those strings in a way I have never experienced before. Spine tingling stuff!

Throughout the entire spectral range, I am hearing better definition of instruments, with more 'space' around them.

Whatever you have changed, and however you have altered it, deserved a big round of applause. It's no going back, regardless of the minor bugs still lurking.

You've introduced a dilemma though. Which of my many favourite albums do I queue up first to re-listen to in a new light!

Well done Peter !

Cheers,

Russ ( the Mad "Librarian" ) 

[SeVeReD]

Yep, I agree with all of the above.  For my system/music I find it even more easy to hear phase changes... seem to be playing with alt-I a lot.

[arvind]

Hi Peter,

I agree the bass response is quite amazing; crisp & very tight, just the way it should be. However I feel that the high's could be turned down a couple of dB's to make the mids more pronounced.

Anybody else feel the same way?

Arvind

[Calibrator]

Hi Peter,

I agree the bass response is quite amazing; crisp & very tight, just the way it should be. However I feel that the high's could be turned down a couple of dB's to make the mids more pronounced.

Anybody else feel the same way?

Arvind

Hi Arvind,

IMHO the overall response is very linear and detailed, with no apparent weakness or emphasis anywhere. The mid's have always been balanced on my system and several friends I have created music servers for, including the guys who designed and built my fabulous Octagons.

If you are perceiving some inbalance in the midrange, try some alterations to the QC parameters to see if that helps.

Have you ever performed an in-room frequency response of your system at the listening position? The results can be VERY enlightening and an ear-opener to what is happening with regard to playback characteristics. Room modes are easily discerned, and it's not uncommon to see peaks and troughs of 15+ dB showing up in suboptimal setups.

I use a program called ETF from http://www.etfacoustic.com/index.html and my last check earlier in the year is shown below.

I should point out that the approx. 3dB elevation from 35Hz to 100Hz is deliberate, and it's the way I chose to integrate the subs with my mains. The mains by themselves have a meaningful response down to the very lower 20's .. the subs just add a little more emphasis without being overpowering. It is seamless blending at it's best.

Cheers,

Russ

[summers_tim]

It's great that XXHE now works with my Empirical Audio Off-Ramp.  The sound is a BIG step up from 0.9x-7, Engine 1 or 2.  Clarity (vocals), attack (guitar or violin transients), and space and air around the instruments.  I've never heard such good source.  Thank you!

[PeterSt]

Glad it works for you too Tim. And thanks for you help on this behind the scenes !
Peter

[arvind]

Hi Russ,

I was doing a A to B comparison with the original CD & the ripped one; with the same set up. I did notice that the high's were a bit over powering, which made the mids feel a bit subdued.

Maybe this is with my setup. Also will check the Freq response in my room again.

Arvind

[ivo]

Hi there,

Well, after several days listening to 0.9y I can agree that bass is better than with 0.9x.
Regarding high's: yes, to me they sound a bit to harsh or agressive if compared with 0.9x.
So, cannot still decide which sound is better for me?
I use Q1=14, others set to 15 (I chose this randomly and also found one person on this forum using the same and can agree with him that Q1=14 and others 15 make overall sound fuller and more detailed, etc)

Ivo

[arvind]

I agree with ivo the high's sound harsh & aggressive & the mids could do with a bit more warmth. Bass is spectacular. My Q settings are 5/0/0/0/0 on 0.9y-2.
I compared with the original CD using the same set up & the result is the same. Peter you may need to look into this. I believe audiophile's using very transparent amps & very high resolution speakers, would be facing the same situation.

Arvind

[PeterSt]

Ok, I listen.

Arvind, by now you may (please) put your preferred settings and all in your signature. At least I feel very comfortable sometimes at judging things, and it may avoid a few rounds of questions. Thanks !

Allright. Now what.

As happens more often lately, it is very hard to judge myself because of the ever changing DAC parameters. So, right now I'm playing with the intended Firewire interface and there is no way I can compare the sound to how it was a few days back.
But I found a potential problem, although it may be a long shot;

Somewhat longer ago (before December 18, 2008) the memory was used in a certain way, and besides the OS being busy with that a main characteristic was that *more* memory was used. Now, the 0.9-y versions (but not the first I think) do that similar, and it is wrong.
This has already been worked out and solved, and the only thing I can say is that when this "new memory management" was introduced, that brought a nice boost to SQ. I only want to say, it can well be so that we went back to that old situation, and it really needs the solution as how I have it here for two days. So, 0.9y-3 will be back to normal to that respect BUT it will not be equal to 0.9x since an essential (and explicit) change was made, which should be for the better (and which has been working ok in the first 0.9y version only).

So, I am quite sure you will notice a difference on 0.9y-3, and whether that is better than 0.9x I don't know, but in theory it should.
If we can agree upon it is not, the essential change is reversed in 30 seconds really and all is back to 0.9x.

Ivo, Arvind, are your judgements about Attended or about Unattended ?

Thank you,
Peter

[ivo]

Peter,

I judge and compare on Attended mode.
Actually, before 0.9y I was using 0.9x6a and I was happy with the way all frequencies are brought to me as a listener! I was thinking: Oh GOD! how my mp3s and Flacs sound. So, now with 0.9y I have partly lost this feeling, however bass is great.

Ivo

[PeterSt]

But Ivo, as you know (or at least I think you do), judging SQ at Attended is "not allowed". It would even be a waste of the SQ you can perceive.

The sound quality for Attended Mode is completely out of my control, and this is exactly why Unattended Mode emerged.
In other words : I can't help it if SQ got worse for you, nor can I reverse things or anything.

So, sorry ...

But did you try Unattended ? (yes you have) and what is your judgement from that compared to Attended ?

[arvind]

Hi Peter,

I will put in my settings in my signature, probably today. The SQ I mentioned about was in Attended Mode; pls forgive my ignorance but what is the reason for difference in SQ between Attended & Unattended?

Of course I will try Unattended today & give you a feedback.

Arvind

[PeterSt]

Hi Arvind,

No, this isn't your ignorance, just the lack of a manual (but remember, all is still beta, so that's a good reason (honestly)).

Unattended - meaning uncontrolled, running by itself - eliminates everything which can influence. "Everything" is not 100% true (yet, but there is more to come), but at least it eliminates the influence from the GUI. And this influence really is there (and to a larger extend you'd wish for). So, what happens is that the GUI program closes (it will really quit and nothing is left of it) and the audio engine (XXEngine3) just runs on its own. Now, no other external influences assumed (which starts with ticking that "Stop Services" checkbox), the SQ *is* under may control. Thus, if then something happens to SQ, I know what should have caused it, but more importantly, within that program I indeed *can* control everything. So, for instance, if you would be asking for something like turning the Coverart upside down, and I would make that, it needs an additional checkbox, and that new checkbox influences sound (additionally). Believe it or not, it just is so.
In the GUI much more is going on like timers running, processes doing things in the background, etc., and they all "consume" SQ. In XXEngine3 there's nothing of that, and what *is* in there is there for a SQ reason, or has been left out for the same reason.

You really should try it, and I promise you sometyhing you again never heard before ...
Peter

[ivo]

Peter,

Thanks for input.
I come from engineering minds closely related to IT and computers so therefore probably will not stop thus far
So, regarding unattended vs attended mode:
I sat down in my coach and played my well known tracks via XX version 0.9x6a (Attended) and was happy.
Then I  switched off 0.9x6a and switched on 0.9y-1d and again sat down in my coach and played the same tracks (Attended) again.
So the latter playback is better in bass, but more harsh&aggressive in highs, so I just cannot believe that if nothing else is changed in my setup except XX version, then I blame new XX version for these harsh highs. You said, that you do not control anything in Attended mode then if nothing else is changed except switching versions, then there is something that makes newer version sound worse than older in Attended mode, which you have forgotten or missed?
I just cannot agree with you that there is no your control in attended mode if you are the master of the code.
Is it possible that 0.9y may require different Q parameter settings than 0.9x to sound the same? Anyway, if you have managed to improve bass response, I may think you have improved jitter handling, so this may mean that the Q stuff may be turned a bit down?

Ivo

[PeterSt]

Hey Ivo,

I could respond to each of your questions or statements and it even might be useful or interesting. What is less useful I guess, is that you don't believe me in what I tell, advising to use Unattended in the first place. Next, I see no reason in "defense" so to say, on things I said in the first place (if it were the other way around, maymaymaybe ).

That you can't get what this is all about, I can't help. But this is why I do this and you do that.

Yes I sure understand that you feel I destroyed your sound (that assumed -> destroyed), but if you buy a car, and the sales man tells you to never shift from 1 to 3 hence don't skip 2, I guess because you know better you skip 2 afterall, and claim some money when the thing falls apart ? This sure looks like your post !

Never mind what I think about it, I SURE (which means SURE) understand how you feel. But then again this is exactly why Unattended was created, just to AVOID such a situation. I guess you were already there when it was created for the first time ?

You may not see it, but in XXHighEnd some 15000 lines of code were added in between 0.9x and 0.9y. I would guarantee for a million sound would change, whithout a single listen.

Question : can it be so that you perhaps never understood what this is all about ? I mean, if I tell you once again that both 0.9x and 0.9y are as bit perfect as can be ... maybe you didn't expect that ? well, it is. So it is external influences changing the sound. When one line of code is inserted at a crucial point, it already does.

Let me make a topic for this, I planned that anyway (coming up -> The Sound of your PC).
Peter


PS: But ask ahead whatever you want, as long as it does't show toooo much "i.....", ok ? hahaha

[ivo]

OK, Peter, I see you are a serious guy
I will then step back from this thread leaving myself in this state:

Addition of 15000 lines of code to 0.9x has improved bass response which is great AND has degraded for some amount the highs response.
The next I will do is try to find ways of getting back the great highs with version 0.9y by experimenting with settings and playback modes. If I manage then I will know what other (1500-bass improvement lines) have done to the code.

Ivo

[PeterSt]

Wrong solution ! Right after you managed (which will most probably NOT be the case) you can start all over because of an upgrade ...

You should use Unattended, and complain on that one as much as necessary !!
I'm serious.

[AUDIODIDAKT]

I honestly find the difference between Unattended and Attended has become bigger in the new version of xx.

Unattended has become the standard for me.

Have to say going in and out unattended has become much better (more understandeable, seeing xx loading files to C: is a good thing)

And when in Unattended you are still in control, because xx has become faster, so you in and out, just like that

Hope we will loose those other sevices too, because I think I can "hear" them.

[AUDIODIDAKT]

And,

When swapping the Spinning disc for a SSD, I have reported that I have lost some bass.

But have that back now, with this new version (and much better)

For me an the other hand, thought my hights has become much better.
Because my tweeters and X-over, are finaly getting good data in that area.

So I don't have to put extra resistor in there. (for months I have that on my to do list, but not anymore)

My calculations were right, the problem was the source. (not really talking about xx, but all cr*ppy cd-players outhere)

Did some testing with a friend of mine last friday,
- full load pc and HDD, and listening to Foobar
- just Foobar
*- XX with full load PC and HDD (differences are hugh now)*
- and xx Unattended (at thist point I get a WOW from him, before he felt off his chair)


*CORRECTED*

[Telstar]

- and xx Unattended (at thist point a get a WOW from him, before he felt off his chair)

OK, I now gotta try unattended again

[PeterSt]

Before people get confused, let me add this :

Assuming Engine#3 of course ...
Running Attended vs. Unattended really uses the exact same Audio Engine (program, XXEngine3.exe). That behaves 100% the same (which is since 0.9y and before it played quite the same, but it was not 100% the same because of the way it grabbed the files to play).

So ... might you hear a difference (before falling of the chair) ... the only difference is that the GUI has gone. The GUI (program) really does NOTHING to the sound explicitly. But it influences anyway ... (again, out of my control, sorry).
The differences (all indirect) become more when those Services are shut down, which won't (or shouldn't) happen with Attended (because all starts to look real ugly).

If it's not clear, ask away.

[arvind]

Hi Peter,

Unattended is definitely better than attended; no doubt about that. However with my set up I still find that the high's are a bit harsh & bold. Tried various Q settings but none improved this. Wonder what is causing this since it appears that I am probably the only one hearing this?

Arvind

[AUDIODIDAKT]

Hi Peter,

Unattended is definitely better than attended; no doubt about that. However with my set up I still find that the high's are a bit harsh & bold. Tried various Q settings but none improved this. Wonder what is causing this since it appears that I am probably the only one hearing this?

Arvind

Hey Arvind,

Not emplying anything here, just out of interest, can I see your speaker specs and x-over layout.

Like I said just out of interest.


Grtz Audiodidakt (Roy)

[PeterSt]

Arvind,

You might try a higher Q1 to start with. Use 14 which is a nice Joe-average. If anything, a too low Q1 (though it depends on the system) may make the sound more "digital". It will be more detailed as well, but it can be over the top.

Next, try setting Invert (On). It may be a difference of night and day.

Generally try to imagine that things will be more "sharp". If all is right more accurate, but this means "sharper". So, it is not said that this may workout for the better in your system, although you would be the first one where "better" is perceived as "less".

Do you have "zooming" noises in your room (IOW standing waves, which may occur at any frequency)? if so, you might - no, will have an impedance problem somewhere. If you feel this may be the case, for fun try without preamp. That will be way MORE "sharp". Does it get better ? then your preamp can't cope.

All rather wild guesses.
Take some time at analysing it. And ... keep in mind what I said earlier, the current version you have out there may not be the best (wrong memory useage). So that may be a reason too, and you may be hearing things the others can't. But this is related to your system of course (and I think it was you who started "high resolution system" ... I think I have just that ... ).

Note : I believe you. Until otherwise proven. Ok ?

[LydMekk]

Hi Peter,

Unattended is definitely better than attended; no doubt about that. However with my set up I still find that the high's are a bit harsh & bold. Tried various Q settings but none improved this. Wonder what is causing this since it appears that I am probably the only one hearing this?

Arvind

I have the same here. Well, not harsh really, but too high in level. Pointed that out earlier regd. the new versions, sounds like theres too little level of bass compared to the rest of the spectrum, and then the top will stand out.

Don't know if this is something Peter has designed to go better with his own equipment?

Something like Windows Media Player e.g. has a much more balanced recital of the frequency range, but of course all the nice staging etc. from XX goes down the drain.

[LydMekk]

To the above, yes, have played with the Q-controls. Doesn't help. The only thing these does is softening/hardening the playback somewhat.

I have suggested it earlier, give us a possibility to implement convolver filters or just add a simple 8-10 band graphical EQ into the player so we can pull down the highs etc. in a system where that is overpowering.

[arvind]

Hi Peter/Roy/LydMekk,

Ok now with all the suggestions the best of the lot is Q=14/0/0/0/0 & Invert. This gives the least harshness with the high's; though I still feel its not good enough. Sharpness is good to the point it does not hurt or fatigue the ear.
I dont have standing wave problem in my room, have installed bass traps as per the freq/dB chart. My preamp & amps are of the same make so impedance matching problem would be highly remote.
Roy, my speakers are Ascendo System M (ascendo.de) the Xovers are 80Hz & 3200 Hz. Tweeters are Ribbon's.

LydMekk, I dont a problem with the Bass as a matter of fact it is spectacular but the mid's could be warmer & the High's are harsh & bold; probably that is why the Mid's are not coming out so warm.

I am using the 0.9y-2 version of XX.

Arvind

[PeterSt]

Well, as expected, Arvind's speakers look better than Roy's. Hahaha, I said "look".
If they are piano black of course.

Let's wait for 0.9y-3 and judge again. I think it will be different.


Arvind, don't underestimate the "phase resonse" in general. It has come forward more than once in here : different versions seem to express a difference to that respect. That is why I suggested the inversion of the absolute phase. You may also see that at lowering Q1 at some stage you may need to invert again ...
Especially when you're into accurate time alignment, it well may be that it needs adjustment; this is about the same subject. And no, you can't do that with random other software (calibrate) because that won't express like XXHE does it.
But wait for the new version first (couple of days).

Peter

[Telstar]

Roy, my speakers are Ascendo System M (ascendo.de) the Xovers are 80Hz & 3200 Hz. Tweeters are Ribbon's.

Well, well, some ribbons do indeed sound harsh. they have a bump in the very HFs. And that's not the source fault.
I'll wait for y3, but i suspect that y version compared to x is more transparent and truthful.

I also would like the chance to add VST plugins, but that's not happening because it woudl compromise the SQ (Peter's words and I believe him).
There will be FIR filters and one can add the equalization in those (by means of secondary programs and not with realtime controls).

[PeterSt]

Hi Telstar,

Although I will agree with you, maybe it is not a nice thing to blame things on that, "just like that". So, on behalf of you too, let me try to correct this (I know you will agree).

In the before post I started blabbering about my own Infinity R90's ribbon tweeter and mid-high, but scratched that just because of the same reason : it should not be the culprit per definition.
At first I extended my (now scratched) post by mentioning the tweaks I applied in the opposite direction : mounting lamp hoods to make the sound more directional, and avoid room modes (in the high frequencies). It really worked.
I again went further in that post, and referred to my own boosted highs (on my current horn speakers, 16dB towards 20KHz via the XOvers) and tried to explain that if enything, I would hear "wrong highs".

But look what I described earlier today in the DAC topic :

If I had to describe the sound as how it is now, I would say

- Aggressive;
- Very good separation of individual elements like a plunk on a nylon string;
- Something like the latter being very forward and straight (no vibrato no flanger) indicating few jitter;
- The sense of wanting to jump along. It is "performing".

The "aggressive" is a dangerous one, because it is quite the opposite of "dead bird" and comes along with "not laid back". You could almost say "less suitable for background music". Dynamics are fairly high (but not too much this time).

... and I should at least recognize that there's some "aggresiveness" in the sound which may not have been there before.
But how can I know if something as drastic has changed to the DAC (direct Firewire) while at the same time I'm changing the software ?

In all honesty this only tells me not to trust my own ears, which implies I must trust yours out there. This means I trust yours, but also Arvind's, and LydMekk for that matter, although he is hopelessly lost in never giving in to Unattended. Only kidding of course, but please read the truth in between the lines.

All summarized, I try to be in the leage of not blaming someone's system as long as it can be the software. As said earlier, until proven otherwise, and not the other way around.
Besides that, with all the different systems everybody has, it NEVER has been a subject. Instead, it ALWAYS has been the software. No matter how it's done and no matter how impossible it seems to be.

I can only hope this is a good attitude, and when it comes down to me having messed up ... so be it.

Peter

[arvind]

Hi Peter,

Waiting for the 0.9y-3 version.

Arvind

[LydMekk]

lol...thx, Peter! 

Nah, not lost in attended, doing un-attended for maximum involved listening. But some listenening are happening together with other activites on the computer and in real life and then the un-attended session creates dropouts, noise etc. the longer it has played. The possibility to stop services etc. further aggravates Vista and the possibilities to run other software simultaneously and creates dropouts with error reports from Vista etc.

Selecting prosessor schemes and thread prios further aggraves these error conditions and therefore have not been enabled in my system.

So yeah, sometimes im doing playback in attended - shoot me.