2.10 sound quality

[PeterSt]

Peter hi,

I am back to this point about  q1 and q1x again,  where sound changes for different values of
q1 and q1x values which produce the same mathematical product.

I'v been playing using q1 30 and q1x 10.
Changing to q1 15 and q1x 20, "definitely" changes sound. The presentation is much clearer and noticeably more dynamic, emotional and musical with the latter setting. Since I last posted above on this, the resolution of my PC has majorly improved, and its now very easy to hear the difference between the setting. The change in sound is repeatable when switching between the settings, q1 15 and q1x 20 just sound so much more musical.

Hi there Nick,

Possibly you thought I was ignorantly silent, but this is not so. Instead I thought to need really a few spare hours to investigate this in the program, which I just now attempted ...

First off, I right away tried to hear what you hear, but it is too much apples and oranges for me because I'd need to change the "product of the factor" to be higher, which I already don't want; I see that others easily skip this phenomenon, or even skip your subject and find the higher product / resulting number better, but I can't go for that (still working on the 12/24 for me new processor and its best sound ever by now, for me).
And exchanging my 30x5 for 5x30 does not seem to do much.
One (very) funny thing : I was just looking up the numbers (like 30x5) to be very sure and see that I left it at 5x30. Is that fun or is it not ...

Investigating the real impact is too difficult. Already the Q1 is used throughout in the program which, mind you, is all about WASAPI and how "Engine3" ever started its life as that (WASAPI playback). So if Q1 occurs 1000 times in the program, it is 900 times in vain for today's subject because you and me are not using WASAPI. But finding that proper context in the program is too difficult because the program is too large (60K C++ program lines for this audio engine only).
So I gave up on it.
But also with a real reason ...

I see that Q1 is used to initialise timers for WASAPI. Btw, this is all (still) in today's secrecy because nobody uses (or can use I'd say) WASAPI the way I do it, which sheerly springs from me being 18 months ahead of everyody and the particular audio stack not being fully developed yet while I already started with it (when Windows Vista came about). Now, the fact that WASAPI can not be used for playback when its relevant services have been shut off in Minimized OS Mode (they are indeed with WASAPI services set to shut off in XXHighEnd's Settings), does (apparently) not prevent the initialisation to go ahead with everything, that possibly really setting timers which ARE there to let respond the OS differently when Audio Playback is in order (read : this dedicates to Audio). So Yes Nick, Q1 alone can very well influence SQ. I actually set myself to an "it will" because I could like it.

Like it ?
Well, we thought to know the SQ parameters by now a little, but it appears someone just found a new dimension within it.
Q1 on its own could "apply" a little of what is applied to WASAPI when used, but now it is separated from WASAPI itself. Read : buffer sizes which are influenced by Q1 explicitly for WASAPI, are now NOT in order when xQ1 is involved (only the product of the both (together with Device Buffer Size !!) is relevant) BUT the timing applied to the OS still happens. Not all the timing, but some part of it. This could be seen as "response resolution" which is within our Clock Resolution which has a life on its own (and was invented later than when we started to use WASAPI) and is explicitly not about that Clock Resolution itself, which is *also* controlled by WASAPI.



So as you see, complex stuff.

If anything, the lower Q1 would be the better one. This is a bit my own theory, but is also a remainder of the WASAPI era, although thus back then mixed with other buffer settings (plus a few miles less knowledge of it all).

Now why was my own sound so outrageously good yesterday. Maybe coincidence. But trouble is : I forgot when I set these settings, which also is related to rebooting and how they could have gone back with e reboot without saving first (of course I started with checking this out when your first posts about this emerged).
One thing I have in mind for very sure : the best sound I have now emerged 3 days ago after a reboot, but this was a special reboot which applied something special ... new to 2.11 ...


Thank you Nick !
Peter

PS: I notice that a year after, we are still not finished with exploring 2.10, right ? haha

[Nick]

Hey Peter,

I had guessed that you might have been looking at the possible reasons for the SQ differences. 60k lines of C that is some program.  It's a facinating explanation of what could be happening, thank you.

So I think you are saying that buffer size is the simple product of q1 an xq1, but q1 is also used elsewhere in the engine by itself to influance timing. Now you have explained this it may make sense in terms of what I hear. The SQ difference I get when selecting a lower q1 value (of course whilst ensuring the q1, xq1 product remains the same) is similar in character to changing the timer resolution setting.

It's good to know my hearing is not changing slightly each time q1 and xq1 "pairs" are changed 

One thing I have in mind for very sure : the best sound I have now emerged 3 days ago after a reboot, but this was a special reboot which applied something special ... new to 2.11 ...

I am really looking forwards to this, I cannot wait.....

Kr,

Nick.

[Robert]

I find this setting Q1 at 5 and xQ1 30 is even better than the opposite. I'm currently finding Xxhighend is very responsive to Alt commands and almost responds immediately.

Still running custom filters Low 705600. Differences between filters easily noticable.

Robert

[Rmalits]

Hi there,

since a few weeks I am using the new settings for Q1 and Q1x:
15 x20. With the small SFS = 0,69 the sound is lot better now, more precise and dynamic. The sound stage improved too.
Before that parameter change the sound was a bit „smearing“ and I wasn‘t sure about my new speakers...

Thanks a lot to all of you who found these new Q1/Q1x settings!

During the last months I also tried different values for balanced load between 8 and 16. I ended up with 12.

Kind regards
Richard

[Nick]

Peter hi,

I think I may have come across another music initialization condition that could be of use for sound quality, similar to the q1 q1x setting further up in this thread.

Over the past couple of months I have noticed that my system always has a particually live sound (eg extra dymanic, strong presence and very nice highs) when I first boot everything up and play the first few tracks each day. I could not pinpoint when this nice initial sound stopped but later in a listening session, although the sound is still very very good, the super nice initial sound is gone.

Whilst doing some equlisation testing I stumbled on a way to swich this nice initial sound on and off. I wounder, if this also works for you and if might be able to pinpoint / exploit what is going on. It seems like playing a 192k with no upsampling / arc prediction could be resetting something important to sound quality that makes the sound quality of 44.1k 705.6k upsampled tracks beter when played straight afterwards.


First getting the not so nice sound.

Play any 44.1k sample track at 705.6k with arc prediction and whilst playing change volume down then up a 1.5 db click whilst listening. Now stop playback and play the next track at you normal volume.  This gives the not so nice sound for me.

With the sound now not so nice play any 44.1k track you know the sound of especially well to see how it sounds (this will be your track B in the next stage that you should compare how it sounds now with nice sound I hope you get from the next step).

(If the above process does not result in the not so nice sound for you, maybe just listen to you test track B after a long playback session of a few hours and then try the next step.)

To get the nice sound.

First play any 192k track (track A), importanly set arc prediction OFF and with NO upsampling ie 192k samples per sec going to the dac. Now stop track A playing and select track B in your playlist (the track you selected that you know very well). Play this track with arc prediction back ON and back at 705.6k samples per second, and listen to the changed sound. Here there is a quite a significant difference with track B now being much more present and musical, no decernable down sides. If there are more tracks following track B in your playlist these will also playback with the nice sound.

My XX settings are pretty much identical to your signature and are up to date in my signature

Could this be useful ?

Best regards,

Nick.

[Robert]

Hi Nick interesting. Some of us went back to custom settings preferred to Arc. I think this was at the time you made some changes back in March this year. I still prefer Custom settings low 705600. Peter's settings online are out of date. Probably best to wait for next version 2.11 which will have even different settings I'm guessing. But you are right the sound does go off somewhat after fresh bootup, although I find it ok for a day at least.

Robert

[PeterSt]

Hey Nick - Not sure whether I comprehend all of that, but I feel the gist and may be able to find out myself blindly
… by now.

You will discover a new SQ feature in 2.11 which looks mighty much similar to what you describe, with the difference that I know what it is doing internally in the program. I still would not be able to reason any culprit or explicit "feature" as such, but it is controllable and looks to have the same behavior as your finding as of now.
In my case the sound changes so drastically that it factually requires different settings elsewhere (like a new PC or new DAC - to name something - LOL).

You will see …

Meanwhile I will try to absorb what you laid out. Maybe it can be automated. And let's agree that it can't become more mad if these matters help consistently. 

PS: You are mad as well; how else could you find out.

Regards,
Peter

[Nick]

Peter hi,

You will discover a new SQ feature in 2.11 which looks mighty much similar to what you describe, with the difference that I know what it is doing internally in the program. I still would not be able to reason any culprit or explicit "feature" as such, but it is controllable and looks to have the same behavior as your finding as of now.
In my case the sound changes so drastically that it factually requires different settings elsewhere (like a new PC or new DAC - to name something - LOL).

I'm really looking forwards to 2.11, it sounds like there is going to be some real fun using it.

Transparency and live qualities have progressed so far in the last 2 years with your new cables, latest XX versions and 621 based PCs. I don't think it would have been so easy to be sure of the changes in sound quality in this post before but now it just stands out so clearly. Iv been trying it out some more this morning, once you know its there its hard to listen without the nice sound set. I hope its possibly for it to somehow be built into the way XX initiates playback. Its already becoming a bind playing the 192k file and changing settings backwards and forwards.

PS: You are mad as well; how else could you find out. 

Haha, yes I think I might be. Its been an obsession lately wringing sound quality out of my system. Mainly PC and Orelo setup work but certainly worth the effort. This SQ thing showed up I was when repeatedly playing 192k sweep files to EQ my bass and listening for SQ changes 

A slightly different topic but I think another factor that impacts the SQ 621 boards when they used from cold is the oscillator. Its located under the chipset heatsink and that whole area goes from ambient to circa 60-70 degrees as the board warms up. I am pretty sure this changes the oscillator speed as the temp goes up proberbly by something of the order of -100 to 400hz. In my case this reduces amplitude of highs and increase bass slightly. I prefer sound cold but in other systems the speed change could just as likley to improve sound depending on the clock speeds of other parts of the system. Im running with a fan on the chipset's heat sink just now but a more perminant fix is coming 

I hope this 192k nice sound thing can be reproduced and proves useful.

Kind regards,

Nick.

[PeterSt]

Hey Nick,

A slightly different topic but I think another factor that impacts the SQ 621 boards when they used from cold is the oscillator.

I prefer sound cold but in other systems the speed change could just as likley to improve sound depending on the clock speeds of other parts of the system.

Right. This is exactly what I was referring to (or tried to) - that it requires different settings elsewhere, which such a thing is in order (or seems to be in order). Of course I could try to reverse the clock speed into some negative cycle, but then I better work on the perpetuum mobile first.
Anyway, envision :

This heating up is done on purpose and it creates a stable environment because eventually saturated. Now adjust the environmental parts to it … (which can just as well be a USB cable configuration - anything).

And oh, to be 100% clear : I don't think I like the sound such a tweak creates, myself. But therefore it now is about finding that other tweak. So our approach is different, the solution is therefore different as well, but the thinking is the same. Well, as usual.

Best regards,
Peter

[manisandher]

I think I may have come across another music initialization condition that could be of use for sound quality...

Thanks for sharing Nick. But I hope you don't mind if I give this one a miss - I think it'd be the straw that breaks this particular camel's back 

Mani.

[Nick]

I think I may have come across another music initialization condition that could be of use for sound quality...

Thanks for sharing Nick. But I hope you don't mind if I give this one a miss - I think it'd be the straw that breaks this particular camel's back 

Mani.

Mani,

Its a nice sound, problem is when I have been listening today I keep thinking wouldn't this be great with the nice sound, then its Click click click click.... click click click ..... click click click ..... ahhhhhhh ! Definitely too much setting up so don't blame you.

I hope that it works at Peter's (dont even know if it will be repeatable in other systems) and that Peter feels its worth / can be incorporated into XX. Although it sounds like Peter has way way more in store for SQ in 2.11. Can't wait 

Nick.

[Nick]

Hi Nick interesting. Some of us went back to custom settings preferred to Arc. I think this was at the time you made some changes back in March this year. I still prefer Custom settings low 705600. Peter's settings online are out of date. Probably best to wait for next version 2.11 which will have even different settings I'm guessing. But you are right the sound does go off somewhat after fresh bootup, although I find it ok for a day at least.

Robert

Hi Robert,

Im back to straight arc prediction again at the moment. I managed to get the highs very detailed but smooth here so unfiltered is working nicely at the moment.

I am certainly looking forwards to Peter releasing 2.11. That is a strong statement he made above about being like a new PC or Dac. Its going to be fun  .

Kind regards,
Nick

[PeterSt]

That is a strong statement he made above about being like a new PC or Dac.

Am I not always full of strong statements ?

Last night went a bit different than what I expected of it.

The day before I had been playing fairly daft electronic music. Now mind you, electronic music as such is about the farthest from what I like. On a side note, you must know what it is first. This is not Vangelis or Klaus Schulze, no, it is (in my view) the most stupid computer fabriqued repetitive (not)music, most often from German origin (think Kraftwerk and you know the direction). Some times such an album ends up in my playlist because I tend to try a lot which looks to be OK from a distance.

Track titles speak (of nothing) and the Cover … could attract me. Track lengths look dangerous(ly way too long to remain of interest throughout).
And so I found my XXHighEnd at track 03, yesterday. Not completely ignorantly I pressed Play on it again, because I was prepared for my tweak (which I, btw, only know for a week or so, and which requires preparation which I forget half of the time - forgot it last night as well).

I didn't know what I was hearing.
From a stupid nothing it had turned into an enormously interesting sounds and soundscapes all over from left to right and in to out and what not. And it occurred to me that even that 14 minute track seems to have passed by in 5 minutes. It went straight into my "Nice Stuff" Gallery for any next time - being sure that this next time I will wonder what happened to me putting it in there. Btw, I recall that the one but last track was nothing much after all; the first two tracks I never played last night, so skip those for safety. Anyway, this album, so you can try it for yourself :

https://listen.tidal.com/album/39109162

I bet you all that you'll regard this the biggest nonsense ever. Or close to it. But not last night over here ...

[Nick]

Peter,

I tried Atmos track 3. It presents as quite bland here with a too heavy drum beat at 55hz. I definatly can see where the performers are going with it and get a sense of what it could be but the sparkle and shimmer in the mid / top end electronic voice sounds never quite happens to make the track pop.

Now if 2.11 can make this spell binding that would certainly be something very special...

Nick.

[PeterSt]

Hi Nick,

Nice that you tried. But yes, if I interpret you well, you really respond to it like I did the first time listening to it. The spellbound thing would indeed be what I dedicate it at listening that second time.

It is as if the resolution doubles without being annoying. A gland "stream" of highs (showing a mere hiss sound) turns into the individual fast on/off sounds and because this is different in the left channel from the right, the sound gets very spacious and holographic. Already the fact that left is different from right in this aspect, adds a dimension.

Something else, which not often is a subject, is how tracks like these can survive their length. So for this type of music, always the same (what about Autobahn, right ?) it is do or die for the nuances. I have something like : we MUST know at all times that no artist makes something really A-B repeating throughout 15 minutes of track. There has to be a build up or else each measure should be slightly different from the previous. Well, if that can't be perceived, you miss out.
This is such an example. It is in there and only theoretically you know there should be more in it. But you surely are not going to try to squeeze out more if you don't know for 100% sure there has to be more in it. But for me it *is* a kind of measure whether all is OK with the system.

This latter surely plays a role when I play something I put to my Demo Gallery ever back, and today I perceive it as way too long and always the same from the start. I definitely know myself of not putting such a track to my Demo Gallery. Still it is in there. So what changed since …
A cable ? the Lush^2 configuration ? Something in the software ? a software setting which improved something when I applied it, but was detrimental to other matters ? (and with the uncomfortable knowing that it went for the worse)

It also goes the other way around, but this is more difficult to catch : An album like this does not go to Nice Stuff and no track ends up in Demo (it does not even survive track 03). Will it get an other chance ? No, probably not. Unless an other album of the same artist makes it to Nice Stuff etc. so I might revisit others too, later.
Btw in this case, might you be interested, I obtained it from Tidal because of this one with 5 out of 9 tracks in my Demo Gallery (which really is many, for percentage) :
https://listen.tidal.com/album/38757515
https://listen.tidal.com/album/38756661 (bonus EP)
If all is right you like this much better from the start.

Peter