JPlay Part Deux: A new controversy?...

[Marcin_gps]

Wow - thanks guys!

Well, let me add my bit. It sounds really, really nice. The low-level detail is quite astonishing. Well done!

Mani

PS. Latest listening done with 64-bit version.
PPS. It still doesn't like being installed and being run from my RAMDisk...

I have the same problem on W2008, not on W2008 R2 though...
I kind of agree with boleary regarding 'the weight' or 'body', but I like it better than XXHE anyway at the moment and maybe it's supposed to be this way, I mean - I feel that this is more lifelike and instruments, vocals sound very natural

[manisandher]

Does anyone else find that japlay, though very crisp, clear, and beatiful sounding lacks the lower end "fullness" of XXkernel streaming.

These were my thoughts when I first heard it:

XX sounds fuller and more musical, but JPlay sounds cleaner.

Mani.

[Nick]

Josef, Boleary,

Thanks for the advice, I have work away for a few days so I won't get chance to try to get play going till later this week. Good to know the hiface can sort though. It's a very interesting experiment with memory I am very keen to hear for myself. Will post thoughts on sound when I get it playing.

Nick

[Josef]

Will give it a try, but I thought that the hiface was limited to a buffer size of 2048?.

Ah, well, you should be able to increase buffer to 2048 in same way - Don't believe there will be much difference though...

BTW: Do you have system priority set to Programs or Background? (Mani, same question for you?)

[boleary]

Josef: My system priority has programs selected, I see your pic has backround services selected. Which is preferable for jplay and/or xx?

Marcin: For me, on my system, the fuller sound of XX is more representative of what one hears live, especially with vocalists. For what it worth (not much I'm afraid) my daughters career has required a few hours of listening in clubs around the US east coast (where I'm often "helping" the sound guy in the room for "best" presentation of her voice or guitar, though I have no training in this area....just what sounds good to me  ). I am always surprised at how similar the sound of XXHighend is to the sound I get at her shows.

Just my 2 cents.

Can't wait to hear Z4 cause I'm sure "things" will be incorporated as a result of the collaboration that continues to amaze  .

I know I've done this before but what the heck: check her out here: carsieblanton.com.

[Calibrator]

Again, please share your thoughts & have fun!

hmmmmm .... well now isn't this interesting.

The aussies are getting pummelled by the poms in the cricket so I took a break from viewing that and had a play with JPlay64 on my music server ( running Vista 64 ).

I set the buffer size in Jplay64 to 64 and, playing from the RAMDRIVE, proceeded to flip back and force between it and XXHE playing my usual set of test tracks:

"Samba Pa Ti" from Ottmar Liebert & Luna Negra's album "Solo Para Ti"
"Canto de Oxum" from David Byrne's album "Girls on My Mind"
"Amanda" from Sheffield Lab's compilation album "Creme de la Creme"
"Spanish Harlem (Rebecca Pidgeon)" from Chesky Record's compilation album "The Ultimate Demonstration Disk"

My overall impressions were that XXHE still had an advantage in transient response and "cleanliness", and this was assessed by careful listening to the metallic percussion instruments within the tracks. I usually find that if I get a good sense of metal objects sounding metallic, then the rest of the audible spectrum generally follows along in it's "believability".

I then started to lower the buffer parameter in JPlay64 until I reached the lower limit of 5 buffers. 4 or less and it refused to play. At this setting, 0.1134ms, it was getting rather difficult to discern differences between it and XXHE. I think a lengthy period is needed before a firm preference could be made, but a conclusion could be reached that however Jplay is using memory allocation and resources, should be attempted to be incorporated into XXHE.

Regards,

Russ

PS. Get well Josef

[Marcin_gps]

Calibrator, but you're on Special Mode-  that's another tale and very much different from adaptive. And Q2,Q3 at 30, you are aware that it totally changes the presentation towards 'fullness' and makes sound more on 'your face'...

[PeterSt]

Marcin, what does that matter ? Are Russ's observations not valid for that ? Ok, it looks like he writes it a bit from down under , but his conclusion (also) is that Josef's means work out for the better (I read it the other way around at a first read).

By itself Russ's observations *are* interesting, because we can do something with it. Set out lines/paths. Well, I think I can.

Peter

[Marcin_gps]

I didn't write that they're not valid, but Q2 and Q3 at 30 changes sound completely Besides special mode is available for limited users (ESI drivers preferable). I haven't been able to play stable with neither RME, Cantatis, M-Audio or Ploytec drivers.

[PeterSt]

OK, I guess I don't get your point then ?

So, I hear what you say, and I understand what you say. What I don't see is the relevance ...
(and you maybe better make it clear before *I* mess up ! )

Maybe you are saying something like : someone who has Q2/3 dialed to 30 is not allowed to judge ?
I could go along with that to some point. But since Russ his observations are not contrary to other's I don't see a problem in it.

Anyway, it is important enough to understand eachother well.

Besides special mode is available for limited users (ESI drivers preferable).

I guess this is merely your point (and not "besides"). Well, I like to put this the other way around : I can use Special Mode (as many can) and I wish I would be able to keep on preferring it. So, it clearly has way more detail, but somehow isn't good. What if "this" (memory approach) would solve that ? It would mean a huge leap forward, with some downside of some people needing another soundcard. So what ? if it is for the (vastly) better ... Of course, if you use the soundcard not to passthrough (but use the analogue outs) it's another matter - and I think you do the latter. Well, *that* would be no sot "valid" I think ... (you will be one out of 100, roughly estimated).

Ok, let me put it all differently. I think I am here to judge all your findings ("all" is "all you"). I also think I can do that, or at least managed rather well so far. Whether I can make something from it for the better is always to awaiten of course. IOW, don't be too afraid I'll triple over one "wrong" judgement, be it a random person or be it you (and you know that !!). In other words : it is far far more useful to have the input from as many people as possible, than none because he now thinks "I am not allowed to speak because I have Q2 set to 30".

... As it is allowed to observe something like that, like you just did. But "observe" is different from turning it into some negative, no matter how well meant (of course). So, I only want to protect a bit against that. That's all.
Well, you know ...

Peter

[Marcin_gps]

Peter, but it's like comparing apples and oranges, you know - to many variables IMHO, but sure everyone can have his own opinion, I didn't mean to criticize or sth. I just wanted to rather point few, in my opinion, important differences:

- Kernel Streaming vs. WASAPI
- double arc prediction vs. native 16/44
- ultra low buffer vs. higher
- and finally Q2, Q3 at 30

I think that a 'fair' comparison would be with XXHE set to Engine#3, the same buffer size and playing native 16/44, but that's my opinion
If Calibrator prefers what he does then good for him! But I can reply to his post, can I?

And you know that I also have dCS Elgar Plus with RME-AES32 in dual AES configuration. The dCS can work in 24/192 in dual AES only, other inputs are limited to 24/96, so I don't think I'm gonna switch to JULI@, which lacks BNC for word clock BTW. I see that you're a bit frustrated that someone could prefer other player? I get it, but you don't have to search for weak points in my system to rationalize my statement, because that what this meant, right - 'of course, if you use the soundcard not to passthrough (but use the analogue outs) it's another matter - and I think you do the latter. Well, *that* would be no sot "valid" I think ... (you will be one out of 100, roughly estimated).'

Cheers,
Marcin

[CoenP]

Hi Marcin,

Peter, but it's like comparing apples and oranges, you know - to many variables IMHO, but sure everyone can have his own opinion, I didn't mean to criticize or sth. I just wanted to rather point few, in my opinion, important differences:

- Kernel Streaming vs. WASAPI
- double arc prediction vs. native 16/44
- ultra low buffer vs. higher
- and finally Q2, Q3 at 30

I think that a 'fair' comparison would be with XXHE set to Engine#3, the same buffer size and playing native 16/44, but that's my opinion

Personally I think there is more to comparing. Just like CPU benchmarks, there are more tests that might work out completely diffferent for completely different configurations. Just like with all audio testing one would like to gain insight how a product compares to its peers right out of the box ánd in a highly tweaked situation. Just like the occasional dac that apparently does not like to be fed from a powerconditioner or brings extasy only connected by cable brand X, can be a winner or looser just the same by 'all things equal' comparison. I would opt for comparing products it their most favourable configuration.
I realise this is not a sustainable principle since allmost everything is variable and makes a difference. Still you could experiment within your own context and still get a meaningfull result.

Since the Jplay is a demonstration app it cannot be so extensively tweaked as XX. I think is is therefore justified for everybody who has gone through the effort of optimising XX compare it with (naturally optimised as possible) Jplay and publish his or her findings.

Unfortunaltely no comparing for me since I have no WASAPI enabled OS  !

regards, Coen

just an afterthought: are you shure the OS and PC XX optimisations are also optimal for Jplay....?

[PeterSt]

I see that you're a bit frustrated that someone could prefer other player?

I really don't get where you got that from ! Unless you mean "settings". And then I still don't see it really.

I had a lot more, but scratched it. But I'll summarize : you have a hard time in knowing what apples and apples are. One example that hopefully opens your eyes (and it counts for everything I tell you) : what "you" are actually asking for is that I insert a preamp again, just because you use one. And Josef. And everybody trying his version of the memory approach. And then what ? Man, I won't even try.
You also want us to listen to native 44.1 while we know whatever Arc Prediction sounds better. Now I don't care who will tell it, but if that sounds better than what we already judged to be worse from another theory, it is completely fine with me, but I won't listen to it because THAT is apples and oranges.

So ... if you prefer to overlook that, or maybe think that you can "hear through" that all, good. But please don't dedicate others to be able to. I can't ! (and even won't attempt as said, which not only counts for the one (two) examples I gave, but for 20 other things you don't think about).

One more thing, and just for fun : maybe you want to hear my ever native WASAPI; you won't believe what you will be hearing; Maybe you were there when the "Less dynamics" checkboxes were there. They were there for a reason. Sadly, these things can not be sustained when things get bigger, can more, have more branches, anything. Already *this* is one reason I won't listen to Josef's attempts if it were for the absolute merits (the relative amongst them is another story). It will say few for the end result. Did you think about that ? no, you 100% sure did not. But then again, the relative difference between the two versions (now four) *will* say something to me. But only that.

Anyone's judgement is valid for that. And yes, I *can* look through that.

Shall we now stop please ?
(not with listening to Josef's attempts !)

Thanks,
Peter

[Windows X]

Im onside with native boat. Arc prediction is good for not really highly resolving system. When use in high resolution system, I feel little annoyed with artifacts (not as much as amarra, though). Though peter said there won't have anything like that but I don't feel it's natural sound to my ears.

My 0.02

[Josef]

Peter wrote:

Shall we now stop please ?
(not with listening to Josef's attempts !)

It is very, _very_ easy to misinterpret someone thoughts when there is no eye contact and only letters on the screen - I'm afraid that just happened between you two guys but please let's carry on as Peter suggests:
Please keep your impressions coming: No matter what settings or amps or whatever!

In the end: This is only about increasing our enjoyment of music!

I bet you all it's not only Peter who is following changes in jplayer and looking for new ideas The more there are - the better! Maybe somebody will get a completely new angle and make a player that blows both XX & JPlayer out of the water - Great!
I personally don't care 'which player is better' as long as I enjoy how it sounds!


And let's keep in mind that, after all, some people will always like transistors and some will never switch from tubes.
Some are 100% digital and yet others won't touch anything but Vinyl.
Some like BMW, others Audi, and yet others Mercedes (poor chaps )
Some like blondes, some brunettes, and others reds... (and some like them all   )

Is it then crazy to suggest that some may prefer WASAPI & others KS?

I honestly don't believe there is such a thing as 'absolute best' in anything, much less audio.

My hope with this little 'JPlayer controversy' is to push the limits to the point where it will be up to personal tastes!

And then we can make our own Arsenal vs Manchester United teams and really fight it out!  

Keep enjoying the music! (and also: guys, apart from Marcin who clearly prefers 64-bit (and I tend to agree) there still are no other opinions on 32 vs 64bits....)