XXHighEnd Application

[The Computer Audiophile]

Hey guys - Just wondering if anyone has used the XXHighEnd program. I've seen it compared to many programs like foobar. It only runs on PCs and supposedly there has been a recent development that increased sound quality substantially. I'm really interested and would love some opinions about this program. Heck, if anyone want to review it that would be fabulous. Or, if anyone from XXHighEnd wants to chime in here we would all appreciate it.

 

Here's the link - XXHighEnd.

 

[glt]

I've been reading about it also, but it seems it "works best" with Vista and I only have XP running iTunes (through Airport Express to get bit perfect output -or so the theory goes). But the fundamental question I have is: how many ways there is to get "bit perfect". Conventional wisdom tells me that getting bit perfect information should be a simple task. Imagine if your software programs were not bit perfect everytime you read it to memory!. Are the claims with different tweaking and different software about "bit perfect" or is it all about jitter. Conventional wisdom tells me that once you can get bit perfect transfer, then it is perfect and there is nothing more to do. Thus it appears that the different tweaks on software and systems are all about jitter. So how would software affect jitter? since the entire system is isolated from the inner workings of the hardware though the operating system, a software system would basically make some system calls from hard drive to memory and then to the USB bus (or though some Direct memory access mechanism). This whole thing really got me confused.

 

[The Computer Audiophile]

Hi glt - While it may be "easy" to provide bit perfect output, many applications do not do this. I am curious to know where you heard that iTunes on a PC has bit perfect output? It is my understanding that iTunes on a Mac is bit perfect, but not on a PC. I am certainly not ruling out the possibility that my industry sources are wrong, but I would be surprised if they were.

 

[glt]

Hi Chris. I don't know where I read it, but I believe that if you use airtunes with an airport express, you get bit perfect, but if you know that this is not true, then I better start looking at getting a Mac...

 

[The Computer Audiophile]

I am double checking with some sources right now.

 

[The Computer Audiophile]

Hold Up!

 

I have to correct myself. I mis-spoke about this one. I wasn't thinking straight about it and I have been corrected by one of the finest DAC manufacturers in the business. When iTunes uses the express it output through the WiFi interface and skips the KMIXER. I should have thought about it a little more before speaking up!

 

Sorry guys!

 

[glt]

Chris, thanks for checking. That would save me at least $600 for a Mac mini :-). I'm using an old IBM T30 laptop with XP which so far has been rock solid for iTunes. I use both the laptop and the airport express with wired ethernet because itunes can connect with airport very, very fast. When I was using them wirelessly, it would take a long while to reconnect to airport after a pause, and sometimes it would not reconnect at all (perhaps due to my generic router).

 

[PeterSt]

Hi Chris,

 

I bumped into your forum by accident, and thought to respond. I am not sure what you want to know though ...

 

Regards,

Peter

(XXHE dev)

 

[The Computer Audiophile]

Hey Peter thanks for jumping in on the Computer Audiophile forums. We are all curious about XXHE. I heard there was some sort of breakthrough that enabled much better sound in Windows Vista. Do you have any information about this or anything you'd like to share with us about the application?

 

[PeterSt]

Hi Chris,

 

Allow me to describe it in brief like this :

 

As an audiophool myself I had been thinking for a very long time (say, 10 years) that playback over computers should just be possible. Say, 5 years ago it was out of the question to propose this to any full hearted audiophile ... that just couldn't work.

In the beginning of 2005 I actually started trying, after -by then- the workout of the theories it just should be better than anything else. A networked player already sounded better than my Teac P1 (both playing over the same Audio Note Sig 3-whatever DAC).

 

All attempts for over a year or so miserably failed, and I got crazy already because of the inconsistency over players, them all being "bit perfect", but still.

 

Since some skills are combined in me, like being the audiophool, a programmer focused on system programming, having a high grade (horn) system so I can "listen" to things, I thought I'd better write a player myself. Only in that case I could know what's going on in there, and then maybe I could find the cause of tinny sound, as I so far appreciated all software players.

 

Strangely enough already the first version outperformed anything I heard so far (this must have been Sept. 2006), and since I got the grasp of it a bit, combined with other people wanting to use it, I set myself to Vista near the end of 2006. Well, *that* did tricks.

 

The mode of programming I use in Vista allows for nearly everything I can think of, with the most important property (for me) of the latency being so low (XXHighEnd can use 48 samples of latency, or 1 ms). It is a huge struggle though, because actually nobody is using that programming mode (which would be WASAPI), so for all things Google brings my own messages only hehe.

 

What it all comes down to, is that -while the software is as bit perfect as the others- sound can be influenced significantly, and what it kind of takes is : well, not much else than someone like me who believes in it (compare Hydrogen), and has the before mentioned "skills" coincidently, and of course the time to do it (I estimate some 4000 hrs so far).

 

To be really honest, it takes *you*. If you'd follow things a bit on phasure.com where XXHighEnd is hosted, you'll see that there is no way I can do this with my ears only, already for the by now numerous variations on DACs and playback sources (like hirez). So what in fact happens, is that I apply what "you" think can be better, or changed for the worse in a before version. Of course I judge the real merits of things, but it really doesn't take much for believing people, once they are part of the community of the (due) best playback ever. There's no theoretical shouting (like all so often), and if people say something, they are just plain to be believed.

 

If you only think you'd benefit, please try it. Trying is just free. Then, if you ever might have the idea of having such crazy (good) ears that you might contribute in judging, I'd say you MUST. It's for all of us, but why not let it be for yourself ?

 

Thanks Chris,

Peter

 

 

[The Computer Audiophile]

Wow, thanks so much for the very detailed post. I think many CA readers will love using your application and helping to make it even better.

 

Thanks again Peter

 

[Wavelength]

Peter,

 

I and many of my user's have tried earlier versions of your product and found the sound somewhat flat. I know you are on a pretty fast track of updates at this point can you say what the latest changes that you have made to make the sound fuller.

 

Also I think latency is useless for a player. I am also a Windows and Mac programmer for years. I was Chief Engineer of the 6th largest hardware/software company for 15 years.

 

What I think a player should do is actually cache and entire song in ram. Actually in thinking this through the other day I was thinking of something like this...

 

malloc for the song, read the entire song, stream the song....

 

also then malloc for second song, read the entire song in... wait for first song to complete.

 

free the first song, repeat.

 

Having the entire song in memory allows and assures the low level drivers of have a continous source of info.

 

kinda like what the memory player unit is doing and they proved that it sounds better than partial buffering.

 

Just some ideas...

 

Will download the latest and have my team try it out again. We have found the USB Audio drivers written by DJ to be allot better than the XP drivers. He did an excellent job and the sound is better for it.

 

Thanks

Gordon

 

[PeterSt]

Hi Gordon,

 

I really can't say whether the sound on your DAC (and yours only !) is fuller now than before. In fact I think it can't matter much because of the, say, high degree of anynchronous working. On the, "normal" DAC the sound has become some "way out" more direct. If all is right this is easy to describe : as your DAC should do already by itself. Not that I heard yours, but I've been working on a similar type, and that one came to me as "in your face" (this is a positive).

 

Assuming that XX can't influence your DAC (which is from the technical standpoint only, but hey, what did we think about these things a couple of years back) I can't say why XX sounds flat, or relatively more flat than others. Again with the DAC I worked on, I experienced similar, but wasn't (for the time I had it in my hands) able to tweak or really try.

It is my assumption (and nothing more than that) that the "influence" which is applied normally (and I mean by each and other software player) may just be needed to let it sound right. So, whether you are experienced on XX or not, whether you may (want to) believe in such things or not, aparantly (!!) there is influence over players the least. And no, you don't hear me say that I have the answers ready for the asynchronous situation ... Now, if we then as a kind of emergency-jump tend to think that XX by means of settings afterall *is* capable of influencing your DAC, it's only a matter of finding what/how. Ehh ...

 

Continuing this little brainstorm (sorry for that) I can tell you that XX, say, natively is very "short attacked". Think in terms of dry, no additional (unreal) enveloping attacks, or (post) ringing. You might just as well say, less warm. However, it is my experience that the more "dry" it becomes, the better it becomes, right until pieces starting to fall together. Or, until your brain starts to think they fall the right way.

When a player expresses more warmth, it is not so difficult to think that it is more fuzzy, right ? And in the end this (by my own experience) can be proved rather easily by means of exploring more detail or just not (the fuzzy ones). On the bass side this would imply the tout bass, or the fumbling around bass. Now on the latter try to imagine that fumbling bass can be cold because so wrong, half fumbling bass warmer and actually good sounding,because we didn't hear better so far. But bass complety free of fumbling suddenly shows a. the individual vibes of the string etc. and b. therewith the instrument. *This* is the time when suddenly things are super warm, no matter how dry it actually is. On that matter, I think you can just as well compare an undecorated concrete room (will sound full of bass but super cold) with the room full of bass traps (bringing just warm bass).

 

With this in mind we should be able to proceed, because you now actually know what to head for.

Also, as I told in the before post (I think) : the really only thing I do is the better 1:1 representation of the digital wave (careful here, I deliberately say digital), and each thingy I apply to get that better, works out for the better indeed. Same should count for your DAC without a player involved, and for theories on eliminating jitter, to name one (important) aspect. But keep in mind this too :

 

It would not be the first time that things got over the hill, looking at an "improvement" which just was not and of which I encountered a few. On that matter, one of the first versions of the Vista engine showed such enormous dynamics, that each smash on a drum made you fall from your chair. It didn't matter whether you knew it was coming or not. The most strange thing was : it couldn't be measured by an SPL meter, but in the mean time all music could only be played 18dB or so lower. It looked like the best there ever could be, but in fact you could hear everybody spitting in the microphone, it appeared that *every* guitar player was rambling over the wound strings with his thumb nail, and Mike Oldfield was just not able to play any instrument on Tubular Bells I. It was all completely wrong, and it came down to instruments falling apart. Not only the individual position of each string out of 6 on a guitar could be pointed out, but the sound emerging from the strings was separated from that emerging from the cabinet. Over the hill.

 

All concluded, it can well be that a player which is too accurate for whatever not understandable reason, doesn't coorporate well with a DAC rather free of jitter by itself. I don't know. I'm learning every day myself here.

 

On to your cache etc, suggestions, what made you think it isn't operating like that from the beginning ? IOW, it just is. You could also say that you just gave yourself the legitimate reasons for a software player to be better than average. Or, maybe now you can understand better why people indeed claim XXHighEnd to be beter than anything else. Anyway it's my explicit objective like you do the same with your DACs, only the means are different.

 

So FYI, exactly what you all propose, I'm doing, and it also happens with DSound in a special way :-) That's probably why people already perceive the DSound parts (that would be Engine #1 and Engine #2) better than ASIO which really has higher latency opposed to "my" DSound already. On Vista (Engine #3) I work with 48 samples latency, and so far the only reason it's not less is because of instable timers under 1ms (and I didn't spend the time of my life yet on finding better timers). Those samples go right into the DMA buffer which is shared with the DAC, which for USB is under the influence of OS settings you know better than me. The influence on USB sure is different, but as sure is there just as well. Again, for your DAC I have my doubts, but the best one to try is you yourself.

Higher the latancy, and it is really one second work to hear the difference. BUT, you'd have to learn what to listen for. A most golden tip is to get yourself some subwoofers, and just feel the difference.

 

This all sure is not about not being able to keep the buffer full, although many will think that, and of course situations/players exist which can't properly. If you'd only know that I could do it even with 1 sample latency (if I had only the proper timers), you'd know it is not about this. Or, run a 176400/24, and count the cpu useage; subtract the driver's useage and you'd see a virtual zero.

That I do exactly as you suggest is for quite different reasons, and as you can guess this is the last about cpu useage (which is a still a, say, 1000 times more with this low latency opposed to the other way around : the large buffer). That I wouldn't be able to meet a latency of 48 samples when streaming from disk is quite another matter.

 

Well, I hope I have been confusing enough to leave some questions (and I don't say I'll answer them all :-).

Thanks,

Peter

 

[geraldedison]

I've been a foobar 2K user for couple of years now and it has been the best performer on my setup even though it is rather feature less compare to Winamp and others. However, I like yourself are interested in obtaining optimal playback...at least further improvements.

 

I'm interested in what you have develop but want to test it in a fair setting. I'm current using my dell laptop with windows XP to do the playback going to my benchmark USB Dac1. Does your software perform the same in XP or should I upgrade first?

 

Gerald

 

 

[PeterSt]

Hi Gerald,

 

Although many people say that even with XP XXHighEnd sounds better than anything else, I would not do that. It's a kind of waste IMHO (and I myself by now think it sounds "ugly"). So you'll need Vista ...

 

There too, in your case, there is a problem, because the Benchmark isn't enough "Vista compatible" as far as I can tell. IOW, it just won't work. http://www.phasure.com/index.php?topic=431.msg3161#msg3161

 

Sorry Gerald.

Peter

 

[markr]

.... WHEW! Hard stuff. Keep it up Peter! I'm adjusting my 401k so that I have enough money available to buy the hardware required to run VISTA 'properly' in the future! JUST so that I can run your application.

 

markr

until then, mac (unix) rulz!

Win XP Pro / AMD Athalon 64X2 5000 4 Gb mem / Macbook Pro C2D 2.4 Mhz / 4 Gb mem / Fireface 400 A/D/A sound / QSC (or Marantz - American made) power / Klipsch LSI-A transducers / lots more undocumented......

 

sorry for 'showing off', I've been reading the threads on your forum. You CERTAINLY work hard. I admire that!

 

[geraldedison]

Peter,

 

I appreciate your insight on this and I will investigate some more.

 

Computer audio is highly flexible but getting it right is certainly not easy...but then again, nothing is easy to get right in this hobby

 

Gerald

 

 

[Wavelength]

Peter,

 

I actually work very closely with the Windows Sound Group and know the USB driver very well. This is one of the reason's why I tell PC user's to use Vista over XP because the USB drivers are much better written (thanks DJ!).

 

I am still at a complete loss why you even talk about latency. I could see if you were developing a mixer application and the requirement of the input to output would have to be low. But really with output only on a 1ms timing I would think making sure the pipe is full or better yet to the fullest would be the best avenue.

 

One of the things that is frustrating the Sound Group is this. Why is it that with all the applications that the same song sounds different? I mean I can see the ripping effecting the sound. But really why is it that these all sound different.

 

I also realize that Direct Sound is not bit perfect and loses the LSB except when playing into a 24 bit dac with 16 bit data and the option for 24 is selected in Control Panel. This is too be fixed in SP1 from what I hear.

 

Also why is it that different ASIO's sound different?

 

My flat comment actually came from my set of beta testers which actually have PCI and other USB dac's other than my own.

 

I will download an evaluate your latest when I have a chance.

 

Thanks

Gordon

 

[PeterSt]

Hi Gordon,

 

I think I never was clear on this, because indeed it is so obvious TO ME that latency by itself does nothing here. So, I *use* it, but not for its normal functional implications.

 

But really with output only on a 1ms timing I would think making sure the pipe is full or better yet to the fullest would be the best avenue.

 

Or just empty ...

Or half full

Quarter

Random ...

 

The point is, with the 1ms (which could be less, but out of decent control then) I am "in control" at 48 sample boundaries. Actually, things start to be highly dependent on the time slices the playing thread gets, and it is acting as a clock ...

 

Don't think about this too much, because it is still a derival from things. It is much easier to think of super micro power surges traversing through all that is connected, and those micro power surges thus electrically influencing.

 

If you ever *want* to understand (read : I myself so much seek for answers !), it is a good idea to believe the story from two (separated) persons who burned CD's during playback with XXHighEnd, therewith burning the signature of XXHighEnd with it and could sort out (ABX) in 100% of cases the correct CD.

Keep in mind : these two persons suffer from a high degree of hearing jitter as a pain (so, they really suffer if there's too much of it), and these persons can discern these things.

 

So what happened, was that a random CD was burned, while (without sound) XX was playing another CD. One of these persons used 3 versions of XX (all with different "jitter signature"), and burned three CD's this way. After this, he could sort out the XX version which played during burning, and he listened with his car radio.

 

There is no way I don't believe this (so, I 100% do), if it were only that with one of these persons (the other than last mentioned above) I started XX and he teached me what he hears when (think in terms of 100ds of emails).

 

The only logical explanation would be that all the "influences" as they emerge upon hardware - no matter how small in terms of Volts - go everywhere, en with the example of the burned CD up to the power of the laser. It would have created jitter in the burned pits, that jitter by itself impeeded by voltage variances riding on the same variances XX incurs for.

 

If you can't believe this (but I guess you do), all that leaves for believing is that this changes sound :

 

i = i + 1;

vs

i+=1;

 

So better believe in something like the burning CD story, which by itself would explain the latter.

I know it may sound the most stupid, but I'm sure you are with me when I say that both commands above require a different "load" from the processor. That this changes sound is something else (but it just does), and that one can manage this is ... well ... a skill ?

I'm starting to learn better and better which does what, and of course it is not as simple as the above example, because it always is about the whole of commands that do something.

At least now you know why it can't be explained. Anyway, by means of indeed something like the latter example, I yesterday improved SQ again by (virtually) a factor of two (that would be 0.9u-8 which is not up yet).

 

I hope this answers your questions a bit.

Peter

 

 

[thermalbat]

Anyone who is a fan of xxhighend player or anyone who's never heard it should check out the latest

version- http://www.phasure.com/index.php?topic=934.0

 

This version is a huge improvement over previous versions(not that they sounded bad!).

Really...check it out!

 

Rich W

 

[PeterSt]

Hi,

 

Please allow me to add that the (release notes) page the link from Rich refers to, talks about "no good if you don't have an NOS DAC" or similar. Or at least at the time this was written, this was a big question;

 

By now it is fully clear that even OS DAC users hugely benefit, except for one report thus far.

I must honestly say that I don't quite understand this myself, but it should be the "reconstruction" of the analogue waves the filtering from OS DACs normally apply, which just don't get (or may not get) the "room" to operate. This, because it has already been done to begin with.

 

An even larger puzzle (to at least myself) is why this -which is about 16/44.1 material- even outclasses native hires material by far. As I said on my own forum "it won't be that hires material is produced by the same normal (and "wrong") filtering as is comonly used in DACs ?" ... but right now I have no other explanation.

 

While thanking Rich for giving me the kind of opportunity to respond from here, I would like to encourage everyone to NOT find this good sounding. And if you do (not) please take some time to explain as exact as possible what you think is wrong with it, over at the Phasure forum.

This is about a new means of filtering (meaning : never applied before) and where I myself think it is the future of digital, it really is not if "you" think it is not. But then I need to know.

 

Thank you all,

Peter

 

[GrayFractal]

I tried it out about a year and a half ago and got frustrated with it.

Sometimes it wouldn't work at first, other times it locked up. Mind you

I didn't use any stripped down OS at the time. It's a simple "looking" piece of software (which isn't necessarily a big deal) with, if I recall correctly, a specific way to access your music files. It just seemed awkward, but all of that I would trade for significant sound improvement which apparently some people wax eloquent about. If they have a newer version with even better sound perhaps it is worth it for me to take a second look. BTW it's not my aim to give any low key disparaging remarks about the software; I never gave it a thorough testing with my best efforts at the time. Also, it seems to me there was a greater difference with NOS dacs, maybe the software author could comment on that.

 

Regards,

GF

 

[XP9433]

It would be more illuminating to have current experienced users of XXHighEnd give their impressions.

 

Fron what I have read on the site, there is exceptionally positive feedback from experienced users to the latest incarnations of XXHighEnd.

 

Frank

 

[manisandher]

Not sure if I count as being one of the experienced users of XXHE (have used it for a couple of years now)... but I'll give you my impressions nevertheless.

 

Yes, it has a simple interface... and isn't the most intuitive player, until you're used to it.

 

I rarely have any problems with it, though on the website (phasure.com), there are posts by others who do. However, these tend to get sorted pretty quickly, due to input from a cadre of dedicated enthusiasts... and of course, Peter... who is simply superb.

 

But the real reason for using XXHE is its sound quality, which just seems to get better and better with time. Actually, this isn't too surprising considering Peter's creative yet absolutist approach, his measurement techniques, and the continual honest and constructive feedback he receives from users.

 

On a final note, I'm in the fortunate position to be able to afford pretty much whatever I want in this hobby (within reason) and XXHE is my choice for playback software... though I'll be keeping a close eye on what Merging come up with.

 

Mani.

 

[givemegut]

How do you guys get an account on XXHighEnd's site. They use e-mail verification to validate a new account, but it doesn't work. I've tried to sign-up or requested the validation e-mail three times now and have had no response from the folks at XXHighEnd. Anyone have any suggestions.