Yes, Brasov is great. The scenery around there is breathtaking. And you can go skiing in Poiana just up the road. I got married to a beautiful Romanian woman last year, in Brasov, her home town. I don't know Bucharest at all, my father in law picked me up from the airport and we drove straight to Brasov. I'm really looking forward to going back to Romania soon.
Well, yes and no. "Digital", in the context of sound recording, means that:
* The signal is sampled, so is no longer continuous in time;
* The signal is quantized, so it is no longer continuous in value;
* The resulting set of values is stored as a bunch of numbers. Now "digital" only really means the last one, but we can stretch it a bit if we know the context. In reverse, a digital to analogue converter converts numbers to "values", usually in the form of a voltage. That's all a DA converter has to do, and it's all it does (ideally, but it will have a non-flat frequency response, so there's some inherent, but unwanted, "smoothing" built in). The output from the DA converter is not especially "smooth", it only changes when you give it a new number. In the context of digital sound reproduction, we would also like to undo the effects of the other two steps - the sampling and the quantization. Well, bad news, there's nothing we can do directly, because information was thrown away. The only thing we can do is minimize the nasty side effects of aliasing in time and value (amplitude if you like). That's not really the job of the DA converter, it's that of the numbers->waveform reproduction system, of which the DA converter is an important part. Now you talk about "interpolating between samples". Well there are a number of ways to interpolate between discrete values. Some of the resulting curves are "smooth", some actually pass through the values, and some are better approximations of the original signal, for a given metric. What we want with audio is to interpolate in such a way as to minimize the extra frequencies created by temporal aliasing. So basically we bung a filter in. Having said that, I think vinyl boys are nutters. I'm just nitpicking a bit because I used to be really into signal processing.
I love those meatballs. I have to trudge around for what seems like an eternity, carrying whatever crap the missus has taken a fancy to, but it's worth it for those meatballs. Oh, and there's the squeezy cheese and ham.
You're right, I wouldn't trust GPU's to do IEEE 754 at all well. I actually posted something along those lines in another thread, but in this one I got the wrong end of the stick.
I'm sorry, I don't normally do this. I like to leave it up to moderators to do what they feel is right. I will never say "mod this up" or "mod this down" because I expect moderators to do the right thing. But the idiot who moderated this as "Troll" has just sent my fuckwitometer right off the scale. Way to miss the point, you sub-moron...
Well actually, to be honest, I know Daniele and Carol. Or should I say, Carol and Daniele. Eheh. Anyway I used to be one of Daniele's "Awkward Moments".
If I could borrow your opening, "Yes, but" GPUs are designed for rendering realtime games. They are not designed for general purpose vector work. You could design a processor using modern silicon processes that would outperform GPUs by several orders of magnitude for general purpose vector operations. There has to be some sort of tradeoff, some point where it would be cheaper to use purpose built hardware. I would suggest that that point lies somewhere between l33t fraggers and serious vector processing mofos.
The Rendition Verite was known for having a RISC processor onboard. I don't associate it with ARM, although it might have been. Anyway it wouldn't have been that quick.
The tagline, "... because she's cleverer than you", obviously hit a nerve didn't it? Here you are, geek, clearly the supreme intelligence in the universe, and someone is purported to be smarter than you. Inconceivable! And she's a woman! You have been well and truly trolled.
Actually his name was Al-Khwarizmi. But the title of the book was "Al-kitab al muhtasar fi hisab al-jabr wa-l-muqabala". "al-jabr" means "the reunion of broken parts".
If I only read one newsgroup, and had no other source of interesting stuff, then you might have a point. As it is, you don't. My Iraq example was clear enough. If I wanted to hear arguments about the war I would go to the appropriate newsgroup. Or go outside. So, what part of "off topic, against the charter, don't post" do you not understand? It isn't tolerated on moderated newsgroups, or mailing lists, and those fora have much higher SNR.
But regulars are often worse than the spammers. Once people get to be regulars, they have no problem with writing endless off-topic, mind-numbing crap. They often enjoy putting "OT:" in the subject line as if to say, "hi, this has nothing to do with the subject of the newsgroup, and hence is against the charter, but I'm a regular, so that shit doesn't apply to me, so anyway..." and on with some boring stuff that happened to them that morning. Or the war in Iraq. I don't think I have seen a single unmoderated newsgroup that hasn't been full of pro/anti war flamefests over the last few months. Er, hello, is it "off" or "topic" that you are having difficulty with? And what the fuck is it with people that reply to spammers and trolls? Spammers aren't listening, trolls just feed off it, and you just reduced the signal to noise ratio. Well done pal. <slaps head> I'd understand if usenet was invented last week, and people were just getting used to it. But it's twenty years old and people really should know better. Does anyone remember Bertrand Meyer's Self-Discipline for usenet? Putting '[++]' (etc) in front of your subjects? It was a nice idea but it never caught on, nor did I ever expect it to. Basically, sadly, ultimately, undeniably, a large proportion of usenet posters are idiots. And yet I can't understand why. Most people you meet in person, when pressed, can put up a reasonable argument. You could have a reasonably entertaining evening debating with them. But on the internet, everyone knows you're an asshole. I used to think that the internet, in bringing us unprecedented global communication, would lead to a more peaceful world. How naive was I? Now we don't just hate people from other nations, we hate just about anybody with a typing finger. And why do people rant so much? Oh, wait...
In general, and IANAL, and I am oversimplifying, and I am from the UK so our legal systems only share a common ancestor, etc,
slander:
- transient, for example:
- speech
- holding up a sign
- pulling stupid faces
- internet chat rooms (maybe, and distributing logs would be libel)
- damage must be proven to have occurred
- except in certain cases, for example:
- accusing someone of a crime
- saying someone's got the lurgy
- calling a woman a slag (only works for women)
- saying someone is shit at their job
libel:
- permanent, for example:
- newspapers
- magazines
- books
- comics
- films
- web pages
- usenet
- damage is presumed to have occurred, so the claimant need not prove it
A while back I was a bit busy so I stopped checking my home email regularly. As I was getting over a hundred spams a day, it was quickly mounting up. A couple of times I sat there and deleted 500 or so, looking carefully for anything from friends and family. Just over a month ago I was overwhelmed. I got 127 today, so now I have 5521 messages in my inbox. That email account is basically fucked. I have no idea if there are any legitimate emails in there. Fucking spammers. Hanging, drawing and quartering is too good for them.
Re:offtopic
on
Nuke-Lobbing
·
· Score: 2, Informative
The United Kingdom of Great Britain and Northern Ireland. Not England.
There's no need to be afraid! I am not mistaken. The whole point of WFMO is to block your thread until something happens. You call it, then it returns when one (or all, iff bWaitAll) of the objects is signalled. Then you do something with that object. Then you call WFMO again. If more than one object was signalled, then WFMO will return immediately with the next index. You don't have to test the other objects, because WFMO does it for you. You are not using it correctly. What you are doing is called 'polling'. Polling is bad, mmkay, that is why WF[S|M]O exists. Yes there is a reason why a pipe created with CreatePipe can't be overlapped, because it just doesn't take the required parameters. Win9x sucks, you will get no argument from me there.
CreatePipe creates an anonymous pipe, and you can't ask for it to be overlapped. But it just does a CreateNamedPipe with a unique name (and you can do overlapped on named pipes). You could do that yourself, although I admit it's a bit strange the way the API is laid out. I quote from the documentation for WaitForMultipleObjects: "The function modifies the state of some types of synchronization objects. Modification occurs only for the object or objects whose signaled state caused the function to return. For example, the count of a semaphore object is decreased by one. When bWaitAll is FALSE, and multiple objects are in the signaled state, the function chooses one of the objects to satisfy the wait; the states of the objects not selected are unaffected." [emphasis mine]. You're not supposed to test objects individually, you just peel one off at a time and go back for another WFMO. I'm off to read about the UNIX stuff.
The display drivers aren't integrated into the kernal, they just run at the kernal's protection ring. This eliminates the overhead of crossing between user and kernal spaces. In short, you don't understand the issues involved. NT *is* a microkernal based OS, j00 n00b... Smack!
Not really. Peace breaks out not when B has no shit left, but when A stops kicking. Oh, and if the Nazis* had kicked the shit out of us [Brits] in WWII (which wasn't far off) do you suppose we should have rethought our position? No, because might does not equal right. Only the superpower of the day is arrogant enough to think so.
For the avoidance of doubt, then, will you now agree that what you are doing is not legal? Oh well, but your conscience is clear because you made "an effort". *tap tap tap* Hello? Deep down (well, not so deep really, I can almost hear you sniggering) you know full well what you are up to. You are kidding no-one, except perhaps for yourself...
Slashdot Mirror
Yes, Brasov is great. The scenery around there is breathtaking. And you can go skiing in Poiana just up the road. I got married to a beautiful Romanian woman last year, in Brasov, her home town.
I don't know Bucharest at all, my father in law picked me up from the airport and we drove straight to Brasov.
I'm really looking forward to going back to Romania soon.
Well, yes and no. "Digital", in the context of sound recording, means that:
* The signal is sampled, so is no longer continuous in time;
* The signal is quantized, so it is no longer continuous in value;
* The resulting set of values is stored as a bunch of numbers.
Now "digital" only really means the last one, but we can stretch it a bit if we know the context.
In reverse, a digital to analogue converter converts numbers to "values", usually in the form of a voltage. That's all a DA converter has to do, and it's all it does (ideally, but it will have a non-flat frequency response, so there's some inherent, but unwanted, "smoothing" built in). The output from the DA converter is not especially "smooth", it only changes when you give it a new number.
In the context of digital sound reproduction, we would also like to undo the effects of the other two steps - the sampling and the quantization. Well, bad news, there's nothing we can do directly, because information was thrown away. The only thing we can do is minimize the nasty side effects of aliasing in time and value (amplitude if you like).
That's not really the job of the DA converter, it's that of the numbers->waveform reproduction system, of which the DA converter is an important part.
Now you talk about "interpolating between samples". Well there are a number of ways to interpolate between discrete values. Some of the resulting curves are "smooth", some actually pass through the values, and some are better approximations of the original signal, for a given metric.
What we want with audio is to interpolate in such a way as to minimize the extra frequencies created by temporal aliasing. So basically we bung a filter in.
Having said that, I think vinyl boys are nutters. I'm just nitpicking a bit because I used to be really into signal processing.
you would literally be blown away
I do not think that word means what you think it means.
- Inigo Montoya
I love those meatballs. I have to trudge around for what seems like an eternity, carrying whatever crap the missus has taken a fancy to, but it's worth it for those meatballs. Oh, and there's the squeezy cheese and ham.
True... sorry, I've been on the Stella all night.
You're right, I wouldn't trust GPU's to do IEEE 754 at all well. I actually posted something along those lines in another thread, but in this one I got the wrong end of the stick.
I'm sorry, I don't normally do this. I like to leave it up to moderators to do what they feel is right. I will never say "mod this up" or "mod this down" because I expect moderators to do the right thing.
But the idiot who moderated this as "Troll" has just sent my fuckwitometer right off the scale. Way to miss the point, you sub-moron...
Well actually, to be honest, I know Daniele and Carol. Or should I say, Carol and Daniele. Eheh. Anyway I used to be one of Daniele's "Awkward Moments".
What do you mean, "probably"? You need 8 bits gamma corrected or 11 bits linear. 32 bit floating point is 23 bits... More than enough.
If I could borrow your opening, "Yes, but" GPUs are designed for rendering realtime games. They are not designed for general purpose vector work. You could design a processor using modern silicon processes that would outperform GPUs by several orders of magnitude for general purpose vector operations. There has to be some sort of tradeoff, some point where it would be cheaper to use purpose built hardware. I would suggest that that point lies somewhere between l33t fraggers and serious vector processing mofos.
The Rendition Verite was known for having a RISC processor onboard. I don't associate it with ARM, although it might have been. Anyway it wouldn't have been that quick.
The tagline, "... because she's cleverer than you", obviously hit a nerve didn't it? Here you are, geek, clearly the supreme intelligence in the universe, and someone is purported to be smarter than you. Inconceivable! And she's a woman!
You have been well and truly trolled.
Actually his name was Al-Khwarizmi. But the title of the book was "Al-kitab al muhtasar fi hisab al-jabr wa-l-muqabala". "al-jabr" means "the reunion of broken parts".
If I only read one newsgroup, and had no other source of interesting stuff, then you might have a point. As it is, you don't. My Iraq example was clear enough. If I wanted to hear arguments about the war I would go to the appropriate newsgroup. Or go outside.
So, what part of "off topic, against the charter, don't post" do you not understand?
It isn't tolerated on moderated newsgroups, or mailing lists, and those fora have much higher SNR.
But regulars are often worse than the spammers. Once people get to be regulars, they have no problem with writing endless off-topic, mind-numbing crap. They often enjoy putting "OT:" in the subject line as if to say, "hi, this has nothing to do with the subject of the newsgroup, and hence is against the charter, but I'm a regular, so that shit doesn't apply to me, so anyway..." and on with some boring stuff that happened to them that morning.
Or the war in Iraq. I don't think I have seen a single unmoderated newsgroup that hasn't been full of pro/anti war flamefests over the last few months. Er, hello, is it "off" or "topic" that you are having difficulty with?
And what the fuck is it with people that reply to spammers and trolls? Spammers aren't listening, trolls just feed off it, and you just reduced the signal to noise ratio. Well done pal. <slaps head>
I'd understand if usenet was invented last week, and people were just getting used to it. But it's twenty years old and people really should know better.
Does anyone remember Bertrand Meyer's Self-Discipline for usenet? Putting '[++]' (etc) in front of your subjects? It was a nice idea but it never caught on, nor did I ever expect it to. Basically, sadly, ultimately, undeniably, a large proportion of usenet posters are idiots.
And yet I can't understand why. Most people you meet in person, when pressed, can put up a reasonable argument. You could have a reasonably entertaining evening debating with them. But on the internet, everyone knows you're an asshole.
I used to think that the internet, in bringing us unprecedented global communication, would lead to a more peaceful world. How naive was I? Now we don't just hate people from other nations, we hate just about anybody with a typing finger.
And why do people rant so much? Oh, wait...
No, it's libel.
In general, and IANAL, and I am oversimplifying, and I am from the UK so our legal systems only share a common ancestor, etc,
slander:
- transient, for example:
- speech
- holding up a sign
- pulling stupid faces
- internet chat rooms (maybe, and distributing logs would be libel)
- damage must be proven to have occurred
- except in certain cases, for example:
- accusing someone of a crime
- saying someone's got the lurgy
- calling a woman a slag (only works for women)
- saying someone is shit at their job
libel:
- permanent, for example:
- newspapers
- magazines
- books
- comics
- films
- web pages
- usenet
- damage is presumed to have occurred, so the claimant need not prove it
A while back I was a bit busy so I stopped checking my home email regularly. As I was getting over a hundred spams a day, it was quickly mounting up. A couple of times I sat there and deleted 500 or so, looking carefully for anything from friends and family. Just over a month ago I was overwhelmed. I got 127 today, so now I have 5521 messages in my inbox. That email account is basically fucked. I have no idea if there are any legitimate emails in there. Fucking spammers. Hanging, drawing and quartering is too good for them.
The United Kingdom of Great Britain and Northern Ireland. Not England.
There's no need to be afraid! I am not mistaken.
The whole point of WFMO is to block your thread until something happens. You call it, then it returns when one (or all, iff bWaitAll) of the objects is signalled. Then you do something with that object. Then you call WFMO again. If more than one object was signalled, then WFMO will return immediately with the next index. You don't have to test the other objects, because WFMO does it for you. You are not using it correctly. What you are doing is called 'polling'. Polling is bad, mmkay, that is why WF[S|M]O exists.
Yes there is a reason why a pipe created with CreatePipe can't be overlapped, because it just doesn't take the required parameters. Win9x sucks, you will get no argument from me there.
CreatePipe creates an anonymous pipe, and you can't ask for it to be overlapped. But it just does a CreateNamedPipe with a unique name (and you can do overlapped on named pipes). You could do that yourself, although I admit it's a bit strange the way the API is laid out.
I quote from the documentation for WaitForMultipleObjects:
"The function modifies the state of some types of synchronization objects. Modification occurs only for the object or objects whose signaled state caused the function to return. For example, the count of a semaphore object is decreased by one. When bWaitAll is FALSE, and multiple objects are in the signaled state, the function chooses one of the objects to satisfy the wait; the states of the objects not selected are unaffected." [emphasis mine].
You're not supposed to test objects individually, you just peel one off at a time and go back for another WFMO.
I'm off to read about the UNIX stuff.
Why don't you tell us all *why* the mere fact that it runs at a different protection ring makes it "less portable"?
The display drivers aren't integrated into the kernal, they just run at the kernal's protection ring. This eliminates the overhead of crossing between user and kernal spaces. In short, you don't understand the issues involved.
NT *is* a microkernal based OS, j00 n00b... Smack!
Yeah but hardly anyone actually shares that amount.. They are all running hacked clients that report their favourite made-up figure.
Not really. Peace breaks out not when B has no shit left, but when A stops kicking.
Oh, and if the Nazis* had kicked the shit out of us [Brits] in WWII (which wasn't far off) do you suppose we should have rethought our position?
No, because might does not equal right. Only the superpower of the day is arrogant enough to think so.
* Invoke Godwin's Law if you must...
For the avoidance of doubt, then, will you now agree that what you are doing is not legal?
Oh well, but your conscience is clear because you made "an effort".
*tap tap tap* Hello? Deep down (well, not so deep really, I can almost hear you sniggering) you know full well what you are up to. You are kidding no-one, except perhaps for yourself...
recieve -> receive
negavive -> negative
punctiation -> punctuation
pot -> kettle -> black