Windows vs Linux: round 2 (Off Topic)

by Eddie @, Sunday, August 12, 2007, 22:38 (6073 days ago)

Hi all,

Five months ago I was at Bert's place to compare the audio performance of a laptop using GC under Windows XP and another one using XMMS under Linux. At that time GC was the winner on points, but XMMS was not made for high-end audio and was not optimally tuned.

Since that time a lot of progress has been made on both fronts. Windows has changed to Vista and PeterSt developed engine 3 for it, supposedly much better than the previous ones. Meanwhile the Linux camp changed into real-time operating systems and a professional audio player called brutus.

On saturday, august 11, the second meeting took place, again at Bert's studio. Present were Bert, Klaus and his wife, Peter and myself. We listened more than four hours to different settings and combinations. Far too much to give an accurate description of all details, I caNot even remember them all.

So just some overall impressions here. Both systems made big progress in a relatively short time. The general impression was that brutus definitely had a more detailed and deeper bass, more high frequency and a wider soundstage. On the other hand GC-3 was more detailed in the midrange, e.g. on voices. The question about which soundstage is better could not be answered.

A few other remarkable points: brutus sounded better without a hardware volume control (the volume can then be set on the laptop). We were all amazed by the large and unexplainable differences between different settings. Klaus wrote a message to diyAudio ( http://www.diyaudio.com/forums/showthread.php?s=&postid=1276574#post1276574 ). You can also find a lot of information there about Linux/brutus, if you are interested.

My conclusions:
1. computer playback is by now miles ahead of what can be achieved with very expensive, commercial CD-players.
2. the progress is to a large extend achieved by trial-and-error. There appears to be virtually no information on what should be done on hardware and software in order to get the maximum performance out of a CD.
3. the approaches under Windows and Linux are very different. In case of Windows you buy a computer and the software, you can play your music and that is it. If you go for Linux/brutus you must install the realtime kernel and run from a command-line screen, you may want to convert all your CD's to 48 Khz, etc. So the second option will cost less money and much more time, but it will give also an enormous amount of opportunities to tweak your system to your demands.
4. This time, to my taste, the winner, again on points, was brutus.

I finish with an interesting question that a collegue posed to me once: "Eddie, what is more important for you: beauty or truth?". My first thought was, of course, thruth. But if you think longer you realize that it is often impossible to realize thruth, so what is left then is to go for beauty. It seems that this is the situation that we are facing in high-end music playback at the moment.


Kind regards,
Eddie

Tags:
0

Windows vs Linux: round 2

by Bert @, Monday, August 13, 2007, 01:19 (6073 days ago) @ Eddie

Thanks Eddie for the "review". It was fun to have it all happening at my place but my personal conclusions are not all the same as yours...

The first initial tests showed that the Manger CD did sound better with Brutus (Linux setup) but for the rest of the CD's I have heard I found GC more in balance, more realistic.

To my ears Brutus has a kind of loudness effect making the Manger track more fun to listen too but giving more bass and more high frequencies makes real instruments less recognisable (Garbarek - Officium) where an instrument screamed like h*ll on Brutus but sounded like a saxophone on GC and this was much more bearable to listen to.

I didn't like the results of the second test where Brutus was used without the Django as attenuator, despite the more "open" sound (overshoot?) there was lack of integration and the natural tonal balance was gone. We should have done that with the TwinDAC and GC too, just to hear what that would have brought for an objective comparing test...

For me GC is still the better player for my system and ears but, to my knowledge, it was not the intention of this afternoon to compete, but reading at diyaudio it seems that it was a kind of contest after all?

As Klaus is writing at diyaudio, it will be needed to prepare things better. Especially if it is going to be some kind of contest and not just checking each others player and hear the progress.

GC is still under development and so is Brutus.

BTW Klaus, the dog was barking all the time (as usual) and does not care about music at all...

Bert

--
BD-Design - Only the Best!

Tags:
0

Windows vs Linux: round 2

by MikeH @, Monday, August 13, 2007, 02:31 (6073 days ago) @ Eddie
edited by unknown, Monday, August 13, 2007, 02:51

A few other remarkable points: brutus sounded better without a hardware
volume control (the volume can then be set on the laptop). We were all
amazed by the large and unexplainable differences between different
settings. Klaus wrote a message to diyAudio (
http://www.diyaudio.com/forums/showthread.php?s=&postid=1276574#post1276574
). You can also find a lot of information there about Linux/brutus, if you
are interested.

I doubt the hardware volume would work well. Reducing volume in the digital domain without drastic increase in bit depth first would lose resolution wouldnt it?

Tags:
0

Windows vs Linux: round 2

by PeterSt. ⌂ @, Netherlands, Monday, August 13, 2007, 03:13 (6073 days ago) @ MikeH

Just an attempt for explanation. Let it even be a first attempt.

Mike, you talk like the digital volume control sounded worse. But with "merits" it did not. This is waaay difficult though, and is related to the merits;

One IMO very important remark to make in advance : Klaus stated that he has tuned his Brutus to the digital voume control, including loosing 2 bits in this particular case (not so much particular, because he applies that in his "home" situation just the same). Anyway :

With this he avoids a Django which indeed did everything to *his* sound.
What it did to XXHighEnd ? ... we don't know, because we did not try it (we could not).
I too "tune" GC to an environment, albeit I try to take care of keeping a distance of it. The funny thing is, that (mind you) with Django the Brutus player performed so, say, strangely, that it led me to some conclusions which indirectly can be found here : http://www.phasure.com/index.php?topic=161.0

Btw, apart from the apples and oranges thing it is like Bert already said : without the Django in between all still did not sound right to my ears (at all). But it would be true that with the Django the Brutus sounded way bad (this is subjective, and we hardly discussed it).

An interesting puzzle so far, is that without the Django, Brutus sounded rather normalized, while with the Django GC already sounded normalized. With this I mean : balanced. So, GC without Django wasn't tested, and Django + Brutus was a kind of misery (overexpressed highs and ditto lows).
Again, Klaus said he tuned all without preamp. This is just a legit explanation (to me).

Maybe more later.

Peter

Tags:
0

Windows vs Linux: round 2 1/2

by GC, Monday, August 13, 2007, 06:38 (6072 days ago) @ PeterSt.

Musceteers...:friends:

It is to my criterias of fruitfull trancending fact finding religion a summit meeting you all had. Not in the league of how to solve climatic changings on Mother Earth or to feed starving stomachs, preventing floodings and draught problems of this planet....but anyhow an event which should regularly happen.

It is as I see it, a laboratory of fact findings, which means much more to me than any Big Sony Blue Ray matrix data storage or earlier Phillips findings of the CD media possibilities....or for that sake 1000's of attemps to "sample" digital sound in religious ways and claim that to be a breakthrough for the next millenium to come. The last just to my standards a pile of "puke".

Eddie, Klaus, Bert, Peter, Wifes and barking dogs: Do that regularely. Make it a tradition to meet and exchange. There is only only one outcome of it:

BETTER(much better)PLAYBACK OF RECORDED MUSIC.

It is a very good idea to have these meetings in one and the same place, such as Berts studio. There you find a pair of speakers, some amps and cables etc. that stay unchanged and all would get familiar with as a reference. (I did not say "the" reference. Just a reference among many).
Then I suddently see that each one of you conclude advantages and set-backs by the source givers in play, that it can only lead to improvements.

This is what I call progressive development. Thanks guys.

Have you ever thought about how important roles you are playing to break throughs to playback? Yes, of course you have, but maybe not in the sence that I preceive it?
Call it MS or Linux, tweaked or non tweaked PC platforms: It is the biggest break through I experienced for a decade at minimum for audio playback you actually are dealing with. (If you should have forgotten it).

Not a moment after the session Peter went straight to his PC environment and applied new settings offering more options to acheive what we all beleive will result in ultimates. That's really dedicated entuisiasm. Klaus worked hard on tweaking his Linux based playback since session 1. There you go.

Now this will likely show to be a (hopefully) never ending proces to the benefit of all and what I particular like about it is that this is an example of that no one is a God and true cutting edge overcommings only comes along when several people contributes. :heart:

Peter kindly mention me as a mostly "off liner" e-mail contributer on his Forum: http://www.phasure.com/index.php?topic=161.msg891;topicseen#msg891

Thanks Peter...and to understand it in a context it's caused by his and my exchange of ideas about the whole chain including likely un-understandable way of taking each our brains into the matter of music perception. CaNot belong even in a "way out of topic" category. Don't think so at least.

Conclusion: None...just go on pals as it is of most importance what you all deal with. :good:


GC

Tags:
0

Windows vs Linux: round 2

by PeterSt. ⌂ @, Netherlands, Monday, August 13, 2007, 12:10 (6072 days ago) @ Eddie
edited by unknown, Monday, August 13, 2007, 12:18

Hi Eddie,

As GC so beautifully expressed it, we had a most fruitful meeting; no matter points for A or B, I could do *much* with what I heard.

Since we were all there in the atmosphere of learning, and what I always provoce(d), I see that you handed out some points anyway, and since you did it with arguments, please allow me to disagree with the arguments here and there. So, only that, nothing about points.


I caNot be sure how (un)prepared this all was, and as you know I was there by kind of accident. Also, I didn't capture the first 1.5 or so hour.
What *I* noticed, is that you were comparing apples and oranges so much (hence explicit as I perceived it) that I really had to ask the question : What are you comparing here ? The answer, luckily, was the only answer that could have been logically given in this situation : Nothing.
So, what happened further, was just interesting. To me, very interesting.

A first thing, that you could not know, was that the things I heard, were already on the phasure forum, from just a few hours before ( http://www.phasure.com/index.php?topic=159.0 ). Now, although we discussed that in between the lines, it wasn't a subject really, because "you" as the majority put your attentions to somewhere else. But let us not forget that what you call more deep bass and more highs (on Brutus), I called a far too saturated sound with over expressed highs. Lacking of mid all over.
You sure remember the (to my ears) really unbearable women singers and saxophone (of which we could not even guess that it was a sax). This was with Brutus, and not with XXHighEnd (the sax remained a problem, but far far less). You, and maybe others too, went for low lows and high highs. I go for less hurting ears ... :yes:

To some extend it wasn't even a subject, because Klaus could prove a 200% different world by leaving out the Django. And I say it in advance ... which we did not try (nor could we) with GC. And still ... it wasn't about comparing, it was about learning. Because of that, again we didn't even discuss what we were hearing, but if *I* am allowed to judge afterwards :
Without Django (to my subjective ears) nothing was right again. You may scream about better transients, and similar to more deep bass and higher highs ... I don't know what to do with that if the balance has (completely !) gone. Without the Django, Brutus showed *no* bass, except for the subwoofer hunking at a lower end. An in fact impossible situation of which I wouldn't even know where to begin thinking what had happened.

Both situations were very wrong to my ears, and somehow with reversed results : with Django too much saturation and too much treble, and without Django no bass but sub low, and a disappeared livelyness because of ? I don't know.

With GC the story was different;

Before I came in you were listening to the TwinDAC+ over the USB connection.
True, the comparison would be more honest (the Brutus laptop couldn't do otherwise) but any judging as such is out of order then. The USB connection is not the best for the TwinDAC+, and under way it converts to SPDIF. A matter of preparation, but I don't think it all was about that. But now stop judging ! grab a beer instead and have an interesting day.

Because of convenience on the Brutus side you started to use SPDIF with GC anyway (a switch now could switch players), and this time we were comparing apples and oranges explicitly.
One has more punch such, the other distorts the voice more so, and personally I really don't know what to do with this. Personally ? nobody should even want to judge, if things are to be taken seriously.

Anyway, again referring to my findings a few hours earlier on phasure.com, I could exactly hear what was going on / wrong with GC, with the nice conveniency of Brutus not performing at all in the same area. Apparently we had found a nice piece of music to test "these" things.
Bert got hold of the 0.9d version of GC which was supposedly to be better (my idea about it), and it was.
What to do with that ? nothing much for the day, but what I did with it in the end is already beautifully expressed by GC.

*If* there would be a conclusion that should be outside of all contexts I mentioned above, it is about the screaming voices and the sax and which player could deal with it in the end.
Mind you, this is about very detailed aspects, which REALLY needs treatment accordingly. This is not about what you say Eddie : trial and error. Not. Oh, I saw it happening, but this was at the Brutus side. The most interesting (I mean that), but trial and error indeed.
I predicted what would happen with the 0.9d version, and I also predicted I could attack the phenomenon. And I did with the 0.9i version. Hahaha, PeterSt is good at predicting ... nope. It is not about that at all. It *is* about paying very very good attention to these things, and I love you for being a witness of the day. Without that, there was nothing to predict nor solve, with an explicit means of testing.

To conclude this for now, it might not have occurred to you Eddie, but all the time I was there, I did not hear GC fail except for the sax. What *you* heard however, was more bass from another player. More highs too.
But what are "we" going to do with that ?
The next time I will personally tweak the xover from the Orphean. I will show you detail you never heard in your life. It's not a promise, but a guarantee. You won't even hear too much highs. Just detail.
But it doesn't work. Never mind the detail apparently is in there, it doesn't work like that. It's over expressed. Like so much spitting Lisa that listening to her becomes unbearable. Or the Mike Oldfield TB-I I wrote about in these forums. The man can't play any instrument because of our beautiful detail.

The moral is balance. To my ears Brutus showed no balance. Not with the setup of Bert, where Brutus had to play over the very same "rotten" USB connection I mentioned. Or via an all over the place colouring Django ... to Brutus. The most interesting. A Django colouring a software player like hell. I can hardly imagine that GC is coloured by it too so much, but possibly it does. But I can tell you in advance : when it does, it does it differently. What to do with that ? I don't know yet. I hope to find it out some day. Just out of interest.
What I do know, however, is that there is no NO *NO* way points can be given to any situation.

If, and only if there is something to judge by me, it can only be the digitally controlled Brutus operation, the wrongish USB connection still being in my mind : this sounded wrong to my ears, and the only thing I can think of afterwards is that possibly we listened to 44K1 -> 48K upsampled material. That would be a satisfactory explanation for me, knowing that these unnatural-like anomalies can come from there (including the sublow honk !).
If so, still no points or prizes to hand out. Just leave out the upsampling and re-listen. Is it worse ? then the digital volume is doing it to you. It was and is just not allowed. No matter it's 2 bits only you lost.
If you're stuck to a preamp because the digital volume destroys otherwise ... then you're apparantly stuck to an over saturated sound with screaming highs.
Then another preamp for a solution ... may be ... but it's not the way to go really.


Well guys, these were my objective votes.
I really enjoyed the day. :clapping:
Thanks,
Peter

Tags:
0

Windows vs Linux: round 2

by PeterSt. ⌂ @, Netherlands, Monday, August 13, 2007, 13:10 (6072 days ago) @ PeterSt.

Just some more comments towards Klaus :

When indeed you applied to 48K upsampled material the poorer the SRC program, the more aliasing you'll have. It is difficult to explain what happens without showing it on a screen, but this is "capable" of creating frequencies at volume levels of the exact same volume of the original (the base where these "harmonics" come from). So, a frequency of, say, 18K at digital level 20000, will incur for a 20Hz frequency with level of 20000. The 20Hz frequency is completely fake, and is audible at the same level obviously.
This would explain the sub woofer honk with no further normal bass audible !!
Btw, it would workout the other way just the same : an original frequency at 20Hz may incur for an 18KHz frequency of the same level.

Since these aliases come from squarish sound, and are squares themselves, the energy of them is already uplifted because of them being squares (the math upon sines versus squares).

The aliases will be all over the place, and individually they harm, but possibly inaudible. Where it goes wrong is when aliases meet. So, there are a few special (or prone to this) frequencies where all aliases (which create aliases by themselves !!) meet. This typically happens at the beginning and at the end of the "official" spectrum, with 22050 and 0 being mirrors. Remember, these are INSIDE mirrors, and not outside as the aliasing we normally talk about. Outside mirroring happens with an 18K frequency that occurs at 26100.

If I must guess this gives the lacking mid with the Django (which really would be over expressed bass and high aliases and which is how it sounded), and without the Django all shifted more to the outsides of the audible spectrum, that possibly being incurred by loosing the 2 bits of resolution in that setup (I just don't know what would be the result to aliasing and less resolution).

If you didn't apply upsampling, this all doesn't count as a relative difference (!!). It would be there anyway because of the nos-DAC principle but less, and probably more randomly organized (less chance that aliases meet).

Just trying to find explanations ...

Peter

Tags:
0

Windows vs Linux: round 2

by Eddie @, Friday, August 17, 2007, 22:54 (6068 days ago) @ PeterSt.

Hello Peter and all others,

I enjoyed the visit at Bert's place, it was very interesting not only to listen to the equipment but alos to meet the people involved.

It does leave me with a lot of questions. From a pure technical point of view I really have no idea how all the differences are caused. There are still no measurements that pinpoint the sound quality aspects that we are discussing, at least I have never seen anything of it. That is why I call this work trial-and-error.

You bring up the matter of aliasing when upsampling to 48 kHz. If this is the case than the upsampler does its work very poor. I did not check it by measurements. This could be done relatively straightforward but I do not have the equipment for it.

Unfortunately, we could not compare one and the same CD at 44 and 48 kHZ with the same player at Bert's. At home I compared a lot of them and all upsampled CD's sound better and I know that aliasing sounds very nasty if it occurs. Three months ago I was strongly hesitating to try this because, again from technical reasons, I caNot understand why it should be a substantial difference.

I want to finish with saying that I am very happy with the equipment that I have at the moment. It is much better that I could dream of when I started this hobby more than thirty years ago. Further improvement is still possible so I remain curious and I hope to spend a lot of time listening to music.


Kind regards,
Eddie

Tags:
0

Windows vs Linux: round 2

by Flecko @, Saturday, August 25, 2007, 21:42 (6060 days ago) @ Eddie

Hi Eddi,
I searched the web but didn't find the page where I can download brutus and the linux distribution you talking about. Can you pleas give me a link?
Is brutus capable to play 24bit96khz files?
Thanks and greetings Adrian

Tags:
0

RSS Feed of thread