Does buffering time (latency) affact SQ in your system?

[Windows X]

I'm running 1ms buffering time (latency) in my system right now. It's making great different for me though it couldn't be possible to get better sound than my CD transport. What about your system? Do latency cause any concern on you?

[PeterSt]

Windows X, hi and welcome here.

For others : the question may look strange to you, but it really isn't. It's hopping over from somewhere else, and I think I'm not entitled to answer for objectivity.

A bit of clarification so you know what the question actually is about :

Windows X can't get it done to find a PC playback means that will superceede his CD Player (Esoteric P-05) and has the idea that this is because the latency of PC systems is too high to really squeeze out the details. Het got this by empirical finding (a sort of : the lower the latency the better it sounds).

To help you on the track with answering : 1ms is something like 200 24/192 samples.

The latency related phenomena in XXHighEnd are :

- Device Buffer Size;
- Q1 for KS Adaptive and KS Special Mode.
- In the end even SFS (split file size). <- I wouldn't be able to talk about that myself, haha.


Hopefully some of you want to spend a few lines on this subject for Windows X.
Thanks,
Peter

[manisandher]

Windows X, I think you will find that most (if not all) regular XXHighEnd users have found that buffer sizes affect the sound. On this forum, you will find many discussions related to this - just search for 'device buffer size' and 'split file size' (or 'SFS').

Furthermore, with the right hardware, you can get to super-low device buffer latencies with XX. There is a kernel-streaming mode known as 'Special' that allows you to get down to these super-low latencies. One sample is achievable!

My experience is that very low device buffer latencies give a 'sharper/tighter' sound and higher latencies tend to soften this. I've settled for a device buffer latency of 1024 samples which gives me the best balance of sharp/smooth sound with my RME AES-32 and Weiss AFI1 interfaces.

FWIW, last year I sold my Esoteric P70 transport because I felt that overall I was achieving a better sound from a PC and XX than from the CD transport. However, there is one area that I'm not sure I've matched - and that is in sheer dynamics. Of course, now that I don't have the CD transport, this is difficult to assess.

Mani.

[Windows X]

I used to think CD Transport can be inferior to music server until I get my hands of real hiend system. At first I don't think Esoteric P-05 can sound better than my project if it's done with hardware modifcation. But the more of time I spend to make my ears adapt to inner details and micro harmonics, the more CD Transport being placed beyond my grasp. And now with buffering and timing controlled for best possible 1ms playback, it still couldn't put up with Esoteric P-05 for all equipments I've tried excluding Mykerinos (That could be possible but it needs to challenge against P-03 or P-01 for its development cost).

I could not say lowest possible latency without dropout or spike always make better performance in every equipment. They need to be at stable state and not many people can archive stable 1-2ms without harshness sound. It involves proper optimization of hardware/software environment and application itself to archive that.

[PeterSt]

Quiet here ? ...

[Windows X]

Seems to be infamous topic for discussion  .

[PeterSt]

It looks like everybody in in church all day. But in the end ... if you do a search for "latency" it's quite easy to see it's everybody's subject.
But not for a Sunday I guess.

And otherwise ... since you are here I guess you believe me by now, which obviously is nothing different from believing in yourself. Now it's only a matter of trying (XXHighEnd) with the notice it is free, but with a kind of a downside that especially on Windows7 you might just "need" Core Appointment to get there, while this is about the only thing not in the trial version (which is coincidence, because Core Appointment has been outside of the trial version always, while Low Latency was introduced the beginning of 2010). But I guess you can get sufficiently low without glitches just to try, while I hope the rest of the "hi-end" stuff in here also will help. Whether you are ready to do it (try) I don't know, but *if* you do, please let us know your results.

Peter


PS: At browsing through some topics here, you will see peoples signatures and what they use for settings. You can learn from that too.

PPS: You will also see at reading through topics that it is a tad different here from elsewhere, and I dare say that without exception people here are like you. This will include for many the understanding "why", but once being here all is different, because you *are* here because of SQ. Not so anywhere else that I know of. So I'd like to challenge you the other way around : try it, and if you don't perceive differences by all the means you can find within XXHighEnd only (which are too many to ever combine while they all make a difference), say so, and know you will be the first. Haha. Good luck !

[Marcin_gps]

Hi Windows X!

My experience in this area is that I prefer lowest possible latency and I've tried everything, believe me The bottom line is: low latency gives you the most out of music files - best details, control and smoothness at the same time, but there are certain requirements from hardware and software side that need to be fulfilled in order to get 'proper' SQ. I promised Peter that I will not continue my topic (with all the tweaks that I find significant), but I'll be happy to send you a PM with my ultimate tweaks, which btw are quite different from those that I published some time ago. In few words to get the best out of XXHE playback you need:
- very fast PC
- slimmed OS
- ramdisk
- few OS tricks

If you choose wrong OS, you won't be able to get a good sound in the first place. Other things I mentioned will boost SQ, make your setup more transparent.

Cheers,
Marcin

[Josef]

It looks like everybody in in church all day. But in the end ... if you do a search for "latency" it's quite easy to see it's everybody's subject.
But not for a Sunday I guess.

LOL: Yes, but there is a reason why this is such a controversial topic - in short, I'm not sure anybody completely understands what exactly is going on as some of it is counter-intuitive (which is a nice way of saying it makes no sense at all  ) and experimental findings are not always positive i.e. sometimes low latency can become 'too much' and distort the sound or, should I say, impart our enjoyment of music. (which, by definition, is subjective, so it makes the topic even more difficult....)

But - that is why we have Marcin!

I promised Peter that I will not continue my topic (with all the tweaks that I find significant), but I'll be happy to send you a PM with my ultimate tweaks

Ah, forget about Peter  ( I'm sure he won't mind as everybody else went to church anyway )

We also exchanged some mails on the subject and I think you are really The guy who has tried more then anybody else out there! AND is willing to share what he had found 

I say, give us the goods and let the controversy begin!

[PeterSt]

Let me add a thing which is not all that obvious, and somewhat harder to find in this forum :

Where the ultra low latency really excels is in only a handful of electronic (ambient) albums, and *then* (for good comparison) only with an NOS DAC allowing for the resolution needed (which would be NOS 24/384 in my book). So, only with this handful of albums (because of their high transient synth music -> sustained by NOS) the ultra low latency gives that additional dimension of detail. These albums all sound GREAT throughout (and might you want to know : plainly distort on parts with an OS DAC because of the filtering in there really making a mess of it). However, use this ultra low latency (which really is near to the sample level) on more "normal" music and it is as Josef said : the joy has gone. Somehow ...

Only few people will disagree on this (the Man Down Under comes to mind), but then their system clearly outbetter mine on e.g. the amp or speaker department, so YMMV. The only thing I know that this preference is rare.

Also, maybe a bit contrary to what Josef said (for safety) ... not much is subjective here. At least not that *I* can recognize from this great group of people. Also, whenever I listen together with a random other person, it never happened that there was even a small sign of disagreement. So, these things are not subjective, if you're only in the realm of "close to reality" which makes the comparison to just that easy. We all at least have *that* reference you know. But you may need to jump some hurdles to get there.

Peter

[Windows X]

Marcin_gps, the hard thing is fast PC introduce more ripple from power and more interferences in the system degrading sound performance. Making lowest possible latency to lowest spec available today is quite a though ones. Especially making it stable is like impossible job though I managed to do it few days ago. Please send your tweaks to me. I'm open-minded enough to try out stuff before making conclusion about it. I found some from your post playing a great deal on it and some made it worse and I fixed something like "Clock Rate" though.

Josef, low latency at unstable state will make harsh sound. You need to make your system really stable and make sure your hardware is capable of handling it.

PeterSt, what interface do you use with your NOS DAC? Which hardware model do you use to connect to your DAC with among USB/Firewire/PCI/PCI Express?

[PeterSt]

Hey X,

Before we proceed, I'd really like you to do us a favour;

As I know from your posts by now (not only the few here), you are talking from some "distance" or generality perhaps. I mean, as how I observe it, now we are going to dive in the merits of PCs and OSes and all, while you are still at that "general" level, especially referring to you not using XXHighEnd for your references. This was quite ok for the initial question, but it is quite not if you are talking e.g. Foobar (or whatever) while we talk XX. Nothing is compareable, and while you seem to be dealing with 1ms of latency as the low side, here it is the high side. Still your low side seems to give you problems, while here, well, I think we have not at all. BUT, we can't be sure, and you may just as well be right on already our high side is subject to improvement.

But do you get what I mean ? you already are a man of far too many theories (from your own), and it is IMHO useless to keep on talking about theoretical situations which we both can not compare for their real life merits.
If you use XX for your reference, we both know what the reference is, and if *then* you post the same as you did, we will know what you are talking about. Next we will try to get it all together and improve if possible.

So : what you seem to suggest now is that "we" should improve while you have the problem. This is non-sense.
On the low side it is a different story I think, especially for those (like me) who think it doesn't sound the best. But *now* you must first judge this, and you can't do that with any Foobar.

Is that ok with you ?
Thanks,
Peter


PS: Please do note that whatever your theories are, your experience is of great value. This means that we will listen to your findings, while next we both and together try to improve again. It happens all the time, and as is known here : I can't do that alone. So, appriciate that you are appreciated.

PPS: Specs of the DAC will be up soon.

[Windows X]

No dude you got it all wrong. It's more than years before you start posting about XXHighEnd for me to get serious computer audiophile (may not as much as now but I do know lowering buffer size in ASIO improves SQ). I did try yours but not license ones. I did heard tons of licensed version of Amarra and XXHighEnd but none can beat real hiend CD transport so far.

Are you certain that you really manage to get 1ms at its stable level? Is your hardware/software environment really ready for it? I'll try to make it sound not offensive but choosing .NET Framework for player draws a lot of bottomneck for UI. I understand that your playback part isn't there but for many reasons behind it couldn't perform better than WASAPI when properly optimized.

When I'm to make judgement, I made it from actual result of experiments not just stood up theories of possibilities. Yes, those includes foobar, J River, Amarra, HQPlayer, Ayrewave and XXHighEnd (mine for trial and someone's machine from paid license version).

For paid license version, the owner of that machine said he couldn't feel that with Weiss INT202 being inferior to his $10k CD Player but tons among hiend listener of his comparison didn't agree with him though.

I'd like to hear your background about your system, Peter. Do you use real system at hiend level. They cost at $50k for minimum. Please don't think that I try to classify money for quality but that's the world for commercial products we live in. There're tons of garbage and real diamond in this field.

[Marcin_gps]

XXHE could play with latency of 1 sample not 1ms, which gives 1/44100 of second, right? But not many drivers allow for that.

[PeterSt]

Are you certain that you really manage to get 1ms at its stable level? Is your hardware/software environment really ready for it? I'll try to make it sound not offensive but choosing .NET Framework for player draws a lot of bottomneck for UI.

That is exactly what I meant. Theories theories. And a kind of deaf so it seems because this must be the tenth or so post addressed to you (including over at CA) without one single response back that it's great that there's actually a player out there which achieves the low latency what you asked for originally. The only thing "we" achieved is that you're here, which is a good start. But now it's time to stop theories and try. And read better maybe. Stop nagging about your old fashioned 1ms and let's see whether YOUR system is capable of 16/384000 which is 4/100 of your ms and is 25 times as low (hehe).

Look, there's this vertical bar. You are on it. To the right of that bar (somewhere) are the good people of CA and while they can't help it not to have your experience they are still good people. To the left of this bar is this community. As good people, but sadly for you, with more experience on this. Most don't understand what is hapening or why, as it is on the other side of the bar. On major difference : nobody (and I said *nobody*) doubts or questions. Why ? because they listen. That other side of the bar does not listen (that I know of or could notice in that thread). Here ? here it is you who doesn't listen. No matter your stories, because they don't fit reality, by itself because this is not about 1ms you keep on talking about.

I'd like to hear your background about your system, Peter. Do you use real system at hiend level. They cost at $50k for minimum. Please don't think that I try to classify money for quality but that's the world for commercial products we live in. There're tons of garbage and real diamond in this field.

I know what you mean, and to a certain degree I agree with that. Let me say, in between the lines, that it may not be all that important what system *I* have, and that it may be more important for both me and you what systems the others in here have. Here too, browse through the sigs of people and you will see many with speakers of that 50K+ alone. If they mention it at all, and e.g. I don't.
But as you already told, it maybe says something, but not all. I will say though that my system sounds better than yours. Just because I know what DAC you use and mine superceeds that by far. Theories here, but I know mine well. And forget about that Weiss you mentioned in the first place. I'd call that distortion machines. Nice if you don't know better. Haha.

but choosing .NET Framework for player draws a lot of bottomneck for UI

??

I understand that your playback part isn't there

Oh, uhm ... so you just like to make statements, or what ?

but for many reasons behind it couldn't perform better than WASAPI when properly optimized.

Now what ... I'm lost here on what you want to say. Maybe it's the language barrier. But I assume you know WASAPI is just in there, and also that hardly anybody uses it anymore because the KS is way better. But maybe you are telling me that the WASAPI engine is sh*t because of using dot-net in a UI which is not used somewhere.

Peter

PS: May you not have noticed, the tone of voice in this topic can't be found in any other here. It will be merely me, but you incur for it.
[reset ]