Originally posted by KameZero
View Post
Announcement
Collapse
No announcement yet.
id Software: Linux Hasn't Produced Positive Results
Collapse
X
-
Originally posted by Scali View PostYea whatever... he laid the groundwork, and it's still an unofficial patch, hence 'RT-linux', rather than just the regular linux kernel
Originally posted by Scali View PostNo it doesn't, but you sound like such a zealot that you probably never even TRIED.
Originally posted by Scali View PostAgain, nope. OS X is good (much better than linux), but so is Windows. Besides, ASIO on Windows is perfect. CoreAudio is not bad, but it is not specifically made for low-latency professional audio as ASIO is. CoreAudio is more comparable to DirectSound (which you don't use on Windows with professional apps), sitting on top of audio drivers, where ASIO is a direct interface to the audio drivers, giving you direct access to hardware buffers. It doesn't get any better than that.
Besides, Windows can do < 1 ms latency out of the box (without extra CPU overhead obviously). Anything below 1 ms is purely theoretical anyway. It doesn't get better than that.
A recent DJ TechTools round up of soundcards sparked some queries about whether latency is a bigger problem on Mac or Windows machines, or if it’s a moot issue these days. Latency is a murky issue with many factors affecting it: the number of audio channels and effects being used, CPU power, and more. However, […]
There are also a lot of other articles similar floating around the web. ASIO isn't bad, it's actually pretty good, but like i said it really isn't necessarily the big issue ~ and your perception of CoreAudio is blatanly absurd.
Not to mention the only people i have witnessed at various gigs or jams that have problems with their computer are always Windows users. My Rackmount never crashes or does anything stupid - nor do any of the musicians i know using Macs. this may be ancedotal, but i know a lot of people with very similar experiences. This tends to be why Mac is the prefered platform for proaudio - Windows is less reliable and can be a big show stopper.
by the way, below 1ms isn't theoretical. there is hardware/software than gets below that now. So yes, maybe theoretical in Windows, but not universally true. And it is a matter of FACT that i can get lower latencies with my VSTi's in Linux than i can in Windows - your position that i can't is a crock of shit because i have compared this in Win7 this year - as at one point i was debating using Ableton Live for some slightly more electronic stuff, as well as trying out some other software. In the end it wasn't as reliable, smooth at extremely low-latencies, although not terrible - running the same VSTi's using Jack + other software was much better, i could use smaller buffers and have way much more flexibility ~ plus, my setup should be headless with on-off operation. i'm sure i could have figured out how to get that to work in Windows, but why bother when Linux is doing a better job...
Originally posted by Scali View PostUhhh, whatever. You just posted some quote about the development of Windows NT, as inthe OS, not the kernel.
The thing is, you are asking me to prove that there is NO name for the kernel, which you cannot prove. If the kernel has a name, then you *can* easily prove it. Because you can easily refer to various sources that define that name. So the burden of proof is on you: you need to prove that the NT kernel is the name, which you can't, because as I demonstrated, not even Microsoft themselves refer to the kernel as such.
You're full of shit, quite dishonest.
Originally posted by Scali View PostIf anything, 'Executive' would be the closest to the name of the Windows kernel. But since that only describes the microkernel itself, and not the Windows subsystem, it is not what people commonly see as 'the kernel' of Windows.
Originally posted by Scali View PostWhat a bunch of nonsense. Unlike linux, Windows and most other OSes don't develop the kernel separately from the rest of the OS.
Windows NT is an OS, the kernel is a part of that development. The NT kernel has ALWAYS had EVERYTHING to do with Windows NT, because it was designed as part of the Windows NT OS.
Originally posted by Scali View PostI hate to point it out, but Windows is not just one OS.
It is a variety of OSes that share some common APIs.
Originally posted by Scali View PostWindows NT was designed for 32-bit and newer processors (386, MIPS, PowerPC, Alpha etc). Before that, an earlier version of Windows existed for 16-bit x86 machines.
It's pretty obvious that Windows existed before the 32-bit kernel existed. But what does that have to do with anything?
Originally posted by Scali View PostAll 32-bit and newer versions of Windows have *always* used a kernel based on the NT branch. The first ever 32-bit version of Windows was Windows NT 3.1. And since 32-bit Windows was a completely new OS (just like OS X is completely different from MacOS 9 for example), the older 16-bit versions of Windows and the kernel they used are completely irrelevant.
So what exactly are you trying to say? It seems you are drowning in your own argument.
drowning...?! lol. No, you are going off the hinges on a very minute point (to do with OS/2) in a wonderfully stupid way. You 'think' i'm drowning in the argument that you are making up to smash...funny shit.
Originally posted by Scali View PostYes, we all know that. It is however painfully irrelevant.
The kernel never officially had a name, not 'NT kernel' or anything else. And you can't prove otherwise. Links directly from microsoft.com contradict you.
epic fail :\
Originally posted by Scali View PostHint: the kernel is not a process.
Hint: whatever you're referring to (could you be any more vague? I suppose you mean ntoskrnl.exe) is not a process on a modern Windows machine.
Originally posted by Scali View PostHint: just because the executable is named a certain way doesn't mean it's the official product name. Heck, most of my 64-bit DLLs are called *32.dll. Are you going to argue that they're actually 32-bit? No, they just kept the same names for better backward compatibility (there's that term again).
Originally posted by Scali View PostHint: You've lost the argument 3 or 4 posts ago, give up.
Originally posted by Scali View PostFor the rest, I've had just about enough of you. You just pile insult upon insult. If you actually had some proper arguments, you wouldn't need insults to get your point across.
the best you have done is provided one page from microsoft - yet counltess other pages do call it the NT kernel - and that shit isn't hard to find from the original page you linked to - no dice, dipshit. You also made claims about Wine that were nothing more than your opinions and then labeled them as being factual ~ which is dishonest, and your are full of shit. and the list goes on.
I don't 'need' to insult you, i choose to because you're a smug little asshole. if you don't like it - too bad. You've also insulted me quite a bit ~ so your argument holds no weight. Actually, really it just makes you look like a hypocritical dickhead.. Come up with better arguments that actually hold truth, back them up with citations and stop being a dishonest little fuckwad ~ then maybe you'll get a little respect.
later.Last edited by ninez; 14 August 2012, 05:26 PM.
Comment
-
Originally posted by ninez View Postit takes one single command to install an rt-kernel on my machines, no different than installing any other software.
Originally posted by ninez View PostSure i have, which is exactly why i am comfortable saying this shit. I've even tweaked friend's windows machines for better low-latency.
Simple fact. You can find tons of semi-professional audio cards that advertised < 1 ms latency with ASIO on Windows, and I've tried a few over the years, and I can verify that they work as advertised. Even USB devices can get ~3 ms latency, which is probably about as good as the USB bus gets.
Originally posted by ninez View PostCoreAudio is not comparable to DirectSound - where the fuck did you get that information?!
Aside from that, reading comprehension fail.
I quote: "but it is not specifically made for low-latency professional audio as ASIO is."
Which does *NOT* mean that it is *NOT* made for low-latency professional audio.
Sure, it can be used for low-latency professional audio. Same goes for DirectSound. But it's a more high-level approach, and also supports things like effect processing etc, where ASIO is just a direct interface, nothing more. No layers, no hardware abstraction, no processiong, no nothing.
If you weren't so blinded by zealotry, perhaps you could just read what I write, instead of your distorted interpretations because your common sense has been shut down by your raging zealotry.
Originally posted by ninez View Postyou are after all a big fan, long time user of Cubase
Originally posted by ninez View Postand think Windows is the best OS of all time forever...
Originally posted by ninez View Postby the way, below 1ms isn't theoretical. there is hardware/software than gets below that now.
I got < 1 ms YEARS ago... I'm just saying that once you've reached the 1 ms barrier, it no longer matters if you can get the latency down any further. So any further improvements in latency are merely theoretical. They no longer add value to the actual user experience. In fact, even my USB devices with 3 ms work just fine. I've used one as a guitar processor with a VST. Didn't notice any annoying latency while playing.
Originally posted by ninez View Postexcept for in the many links i have already provided to you. I am not asking you to prove there is NO name - i am asking you to prove it's not called the NT kernel - which you cannot do - since it IS called the NT kernel.
So yes obviously it is referred to as the NT kernel. This was especially true back in the days of Windows NT 3/4. With Windows 2000, most people still were very much aware that this was actually 'NT5', so the name was still quite common. However, since XP, Vista and 7, the NT name is no longer that familiar, so you see "Windows kernel" more and more, and "NT kernel" is mostly a relic of the past now.
Originally posted by ninez View PostI've met my burden of proof
More reading comprehension I suppose.
Originally posted by ninez View Postexecutive is part of the kernel, but is NOT the entire kernel, nor is it the 'microkernel' within the NT architecture - that is plain WRONG;
Originally posted by ninez View Postexcept for the fact that it wasn't called 'Windows NT'. It was actually developed as NT OS/2 for IBM..
However, due to the sudden popularity of Windows 3.0, Microsoft and IBM parted ways, since IBM wanted to use a special OS/2 interface, whereas Microsoft wanted to use the now-popular Windows 3.0 UI, and just sell it as Windows NT.
But it was still the same project.
Supporting multiple subsystems has been a core design decision for NT since day 1. Even when they were focusing on 'NT OS/2', it was always going to have a DOS and Win16 subsystem, and could run various other subsystems as well, even a POSIX one.
Win32 is just one of the many subsystems in the NT project, which eventually led to Windows NT (where Microsoft dropped the OS/2 subsystem).
So I'm not sure how you are claiming that the 'NT kernel' is not named after either NT OS/2 (early in the project) or Windows NT (later in the project).
Originally posted by ninez View Postmaybe you should learn how to read properly. dumbass.
Originally posted by ninez View PostWho said anything about Windows not existing before the 32bit kernel? (i sure as hell didn't). I pointed the 32bit thing, simply for the fact that oOS/2 was already a 32bit OS before windows 3.1 NT was ever released...
Originally posted by ninez View PostI've already proven otherwise, using microsoft.com DUMBASS - you jsut choose to ingore that - your just a little bitch, who ignores FACTS. I provided many links/citations, you provided one :\
*You* however are denying (despite links from microsoft.com) that the kernel is *also* commonly referred to as the 'Windows kernel' (which is more appropriate today, since the NT brand name has not been used since 1999).
Originally posted by ninez View Postthe best you have done is provided one page from microsoft
Comment
-
Originally posted by Scali View PostBut still you need this alternative kernel, you can't do it out-of-the-box.
Originally posted by Scali View PostWell, I've had < 1 ms latency with Cubase on Windows more than 10 years ago... so if you need to tweak a Windows machine for better low-latency with today's hardware, you're doing something quite wrong.
Originally posted by Scali View PostSimple fact. You can find tons of semi-professional audio cards that advertised < 1 ms latency with ASIO on Windows, and I've tried a few over the years, and I can verify that they work as advertised. Even USB devices can get ~3 ms latency, which is probably about as good as the USB bus gets.
Originally posted by Scali View PostIt is if you know what DirectSound is.
Aside from that, reading comprehension fail.
I quote: "but it is not specifically made for low-latency professional audio as ASIO is."
Which does *NOT* mean that it is *NOT* made for low-latency professional audio.
Sure, it can be used for low-latency professional audio. Same goes for DirectSound. But it's a more high-level approach, and also supports things like effect processing etc, where ASIO is just a direct interface, nothing more. No layers, no hardware abstraction, no processiong, no nothing.
Originally posted by Scali View PostIf you weren't so blinded by zealotry, perhaps you could just read what I write, instead of your distorted interpretations because your common sense has been shut down by your raging zealotry.
Originally posted by Scali View PostI am a user of Cubase, and various other software. Not necessarily a fan.
Erm, on the contrary. But if your sick mind interprets it like that, I'm not surprised. I guess you only see the world in black and white.
Originally posted by Scali View PostAgain, reading comprehension.
I got < 1 ms YEARS ago... I'm just saying that once you've reached the 1 ms barrier, it no longer matters if you can get the latency down any further. So any further improvements in latency are merely theoretical. They no longer add value to the actual user experience. In fact, even my USB devices with 3 ms work just fine. I've used one as a guitar processor with a VST. Didn't notice any annoying latency while playing.
My Mac has no problem doing less than 1 ms (and pretty much every mac before that, going back years and years) and the same is true of Linux... So your 'common sense' tells you to make up shit, does it?!!? As a guitar player why would you notice?! ...As a keyboard player, i like to keep it as low as possible... In fact, since you claim to be well above 30 - at least it would appear that way based on certain things you have said, the fact is you are probably used to guitar FX that historically have had much higher latency, i know i am from playing older 80s-90s guitar FX boards (i also play guitar, although prefer keyboards)... But i tend to notice it more with my keyboard playing (using digital synth modules) than guitar with effects. I can compensate (as you do with any old analog moog-style synth or piano), but when you are playing really fast leads (32nd notes), rolls, arpeggios, etc. Lower latency is better. And AFAIK USB (even 2.0) is of no use to me, Firewire or PCI cards only. I personally think USB soundcards are not a good way to go (in any OS). Especially, if you are using multi-in/out consistently (which i am).
Originally posted by Scali View PostAgain, reading comprehension: I have never claimed it is NOT called the NT kernel. What I have said is that there is no *OFFICIAL* name for the kernel.
So yes obviously it is referred to as the NT kernel. This was especially true back in the days of Windows NT 3/4. With Windows 2000, most people still were very much aware that this was actually 'NT5', so the name was still quite common. However, since XP, Vista and 7, the NT name is no longer that familiar, so you see "Windows kernel" more and more, and "NT kernel" is mostly a relic of the past now.
Originally posted by Scali View PostNot at all, you merely 'proved' that 'NT kernel' is one of the common unofficial names used. Something which I have never contradicted in the first place. I merely pointed out that this is not *THE* name of the kernel, since it does not have one.
More reading comprehension I suppose.
here is a page that tracks all NT kernels by version, right back to the first;
oddly enough, while you claimed those .exe files don't really have anything to do with the kernel, and then went on some rant about 'backwards compatibility' ~ it actually turns out that what you were saying was wrong.
Originally posted by Scali View PostAgain, reading comprehension. I *specifically* said that Executive would not be a correct name.
Originally posted by Scali View PostIf anything, 'Executive' would be the closest to the name of the Windows kernel. But since that only describes the microkernel itself, and not the Windows subsystem, it is not what people commonly see as 'the kernel' of Windows.
I think you're the one with the reading comprehension problem, and also some mild retardation to boot.
Originally posted by Scali View PostWhat is your point? The NT project started out as OS/2 2.0, everyone knows that.
However, due to the sudden popularity of Windows 3.0, Microsoft and IBM parted ways, since IBM wanted to use a special OS/2 interface, whereas Microsoft wanted to use the now-popular Windows 3.0 UI, and just sell it as Windows NT.
But it was still the same project.
Supporting multiple subsystems has been a core design decision for NT since day 1. Even when they were focusing on 'NT OS/2', it was always going to have a DOS and Win16 subsystem, and could run various other subsystems as well, even a POSIX one.
Win32 is just one of the many subsystems in the NT project, which eventually led to Windows NT (where Microsoft dropped the OS/2 subsystem).
Originally posted by Scali View PostSo I'm not sure how you are claiming that the 'NT kernel' is not named after either NT OS/2 (early in the project) or Windows NT (later in the project).
Originally posted by Scali View PostWhich was the same project... OS/2 was just done a few months sooner... what's your point?
Originally posted by Scali View PostI don't see me ignoring anything, because I never denied that it was commonly referred to as the 'NT kernel'. Heck, I used that name myself, back in the NT3/4 days.
*You* however are denying (despite links from microsoft.com) that the kernel is *also* commonly referred to as the 'Windows kernel' (which is more appropriate today, since the NT brand name has not been used since 1999).
For someone whom is claiming others having 'reading comprehension problems' - you clearly have not only reading problems, but severe memory problems, and retardation.
Originally posted by Scali View PostOne was all I needed. Besides, it was two.Last edited by ninez; 14 August 2012, 09:13 PM.
Comment
-
I dont think that audio latency is a huge problem on Linux. Ok, PA may not be the optimal case for everybody, but you dont need to use it on simple setups. Often sound is done via openal which is even available for Xbox (360). That abstraction layer doesn't hurt on win as well and lots of games use it. So whats problem with audio?
Comment
-
Originally posted by Kano View PostI dont think that audio latency is a huge problem on Linux. Ok, PA may not be the optimal case for everybody, but you dont need to use it on simple setups. Often sound is done via openal which is even available for Xbox (360). That abstraction layer doesn't hurt on win as well and lots of games use it. So whats problem with audio?
http://en.wikipedia.org/wiki/OpenAL#Games
Games generally work fine even with cheap onboard audio, which have quite poor latency.
The problem is with realtime audio processing. So you input the sound of an instrument/microphone/whatever, then add effects to that, and send the processed sound back out. Low latency is required here, not only to keep the sound in the rhythm of the rest of the music, but also because it can be very disorienting if you hear your own voice/playing back with a delay.
So it is important that the sound is sampled, processed, and sent back out to the audio device with a delay of only a few ms at most, so that you still perceive it as 'realtime'.
This requires very small buffers and very tight responsiveness from the system.
For example, with 44KHz CD quality, you have 44100 samples per second. A 1 ms buffer is only 44 samples. So every 44 samples you transfer back between the system and the audio device. That's quite intensive (regular task switching is generally done with 10-20 ms intervals).
Comment
-
Originally posted by ninez View Post...
Anyone can tell you that your own links don't even support your case (like that link you posted, where they literally say ASIO itself works fine, there is just a problem with some cheap hardware/poor drivers (hence WIFI issues and such). This doesn't happen on Macs because Apple doesn't ship cheap hardware and poor drivers in the first place. Can't blame Windows for cheap hardware with poor drivers).
Comment
Comment