Isn't it blatantly obvious how this reflects the nature of all humans? Kindness directed towards strangers isn't exactly a common trait. There is even a theory to explain the its rare occurances; those with no prospect of reward: Humans are evolved in small social groups, where favours were more likely to pay off. These groups rarely encountered outsiders. Therefore, humans never developed the social instinct to differ between kin and stranger. The theory basically assumes that all humans are egoists. As a large company can be viewed as a collective conciousness of sorts, egoism is just a reflection of a basic human trait.
Actually, most music businesses seem to use WMA. Although OGG beats MP3, some tests show that WMA beats OGG. I've heard the opposite as well, though, in A/B listening tests published here on Slashdot a while ago. Many of my aquaintances are using WMA, because normal people don't know what OGG is. The über marketing skills of GNU geeks has also blesed us the ugliest codec name this side of the galaxy.
Okay. Our game consoles don't have SCART connectors, but there's usually a Phono-to-SCART-adapter included. Most old TV sets only take input via SCART if there's any input but the antenna one. Modern TVs usually have Phono and S-Video connectors on the front, and 1 to 3 SCART inputs on the rear. The SCART system has a couple of clever features: For instance, good TVs will auto-switch to the right SCART input when equipment is turned on, and back again when it's turned off. If you tell your DVD player you have a 16:9 TV, the TV will switch formats automatically. On a good, regular 4:3 set, this means that the set does automatic letterboxing. Of course, another nice thing about living in Europe is PAL's picture quality.;)
Funny. I've actually thought about a system like this. A digital replacement for the SCART connector. Do you guys over in America have SCART connectors?
There's one thing I've always wanted to try:
1. Record audio digitally at a high sample length (say 32 bits).
2. Segment it into short blocks of maybe 1024 samples.
3. Normalize each block and store the amount of amplification together with each block.
4. Quantize to lower sample length (say 12 to 16 bits).
5. Store reduced data heap to medium.
Wouldn't you be able to compress audio of quite high fidelity in this way, without actually having a noticable reduction in dynamic range nor quality? Introduce volume interpolation (smooth transitions) between each block and you'd constantly get the highest quality possible for each sample (the encoder and the decoder would interpolate in synch).
The HDCD format uses the least significant bit of the signal to add extra dynamic information. The same method could be used to encode expansion information. Imagine this being used by mastering plants to improve sound quality. With special plugins, compression information could extracted in the studio, stored in the least significant bit of each sample, and if the listener wanted, he could cancel out studio compression with a touch of a button, if he had the right equipment.
Of course it'll never happen. But it's fun to dream about.;)
Much more happens in a tape recorder than just the addition of noise. There is something called warping, that is inaccuracies in the speed the tape passes the playback/recording heads. There is crosstalk, that is mixing of audio across several channels. There is also the fact that the signal is being recorded with a bias to bring it into the most linear part of the tape. And then there's the most noticable effect: Compression. When you over-record a tape, you get compression, that is reduction in audio levels compared to the original levels. A digital recorder will just clip, sounding horrible. A tape recorder will do it more gently. Modern musicians and technicians are very fond of what's described as vintage sound. I need just mention the UREI LA-2A compressor, an opto-electric tube compressor, used on numerous recordings. I dare bet anyone who has ever listened to recorded music has heard the handywork of that machine, or it's digital emulations. I love the sound of warping, especially when it comes from record players. It's what I call a becoming untunedness or unstability in pitch. Creates a warm fuzzy feeling inside of me, at least. Tube equipment is popular for grunching things up a bit, the newest Korg synthesizer model, a purely digital machine in all other aspects, has a tube stage for adding an edge to the sound. Guitarists unanamously agree that tube amplifiers give the best sound... I can go on and on... But it's a part of musicians' culture, basically.
Well, I saw no other reason for mentioning it, so I wildly assumed creds were the reason.
Fine. He created all that, and for that I am grateful. But I don't like the way he does PR. I'm embarassed on his behalf, because I just know that so many disagree with aspects of his arguments. If the public views him as some sort of representative for the Linux community... Well, I just don't like the thought of that, because people will often interpret opinionatedness as narrowmindedness and stupidity. He's hostile to the whole model of closed source programming and he doesn't care if a zillion programmers lose their jobs. I just dislike that strongly. Although I am a fan of open source and free software, I think they can coexist with closed proprietary software. I don't want to abolish it. He's just so extreme.
Hey, I did Linux in the 90s too. My computer teacher saw that I wasn't like the other kids, so he let me play with old hardware instead of following the regular program. I installed Slackware on it and learned tons. I don't see how this has anything to do with your geek credentials, though.
I'm just saying RMS seems to be this very opinionated guy. He seems obsessed and overdramatisizes everything. I'm not the only one who thinks he's harming more than he's helping with his attitude.
Why does it have to be so that every time there is a disagreement, ever so civilized, on Slashdot, somebody has to shout "idiot"? Isn't it possible to be factual? This is the sort of thing that makes me read Slashdot comments more for the sake of contemptuous entertainment than true insight.
I don't see why this guy is such an icon for so many geeks. He seems unsympathetic and makes outrageous black & white statements. I think he gives Linux a bad name.
I'm using Firefox. I bet it would display correctly if I told it manually that the page uses ISO-8859-1 or what ever. But this is Slashdot's responsibility.
And don't come here and be nit-picky. You know as well as I do that I was speaking in the context of this message. In Norway, "æ", "ø" and "å" are called "Norwegian letters" even though they're used in Denmark too. I bet the Danes call them "Danish letters". Neither of us pick on the other for doing so. In this context, they are Nordic letters. So there!
You've noticed too? I swear, people, this is true. Outlook ignores viruses and blocks friendly files. My dad can testify about this. He got a bunch of e-mails, most of them spam, some had bad stuff in them, and Outlook didn't grey those out. However, when a friend sent him an MP3, that was greyed out for some stupid reason, and we had to disable the "protection". How is it possible to write such stupid software?
Hehe. The classic "LPs are better"-argument. I wish I had good LP equipment so I could actually go and compare things. Unfortunately I was born in 1983 and became part of the CD generation. I have never owned an LP in my life.
I also wish I had the money to buy or even build a tube amp, just to check how that sounds. I do have access to studio monitors. They tend to reveal a lot, because they're neutral. Most mixes sound cold and impersonal through them and I wonder if that is just my ears being used to listening to equipment that colors the audio to make it more interesting or if the sound picture would warm up if I connected a record player to them.
I can't say I'm 'blessed' with tin ears. I notice artifacts that many people don't hear even if I try to make them aware of them. My father accuses me of having fly eyes, because I can see the MPEG2 compression artifacts in the pictures from our DVB sattelite tuner.
If I ever get the chance to wire an oscilloscope up to a PC sound card's output, I'm definitely going to test various waves and see how accurately they come out.
As an experiment, have you tried digitizing some of your premium albums? Maybe you'll find that the digital recording will be able to capture the sound of the LP. It's popular belief these days that adding analog artifacts to a sound will make it sound 'warmer'. Maybe the manner in which LPs have to be mixed (the RCA stage) creates that particular sound that your ears like?
As for sawtooth waves being rounded off to sine waves, that would happen as the pitch gets close to the cutoff of the lowpass filter before the A/D, which is designed to cut off as close as possible to 22.05 Hz. Human hearing ends at about 20 kHz and most adults I've met (including myself) can barely hear any sine wave above 16 kHz.
The compression itself wouldn't cut off a sawtooth wave. It's just a DCT analysis (similar to FFT) that, if performed in inverse, should reproduce the original waveform. DCT converts from the time domain to the frequency domain, which in English means that it produces a frequency 'curve' of whatever waveform was processed.
A sawtooth wave will have many harmonics (sine waves with pitches that are multiples of the base frequency), both odd (N * 1, N * 3, N * 5,...) and even (N * 2, N * 4, N * 6), and these produce the characteristic sawtooth sound. DCT will not kill those harmonics. The psycho acoustic analysis following that might take away some of them, but at a good bitrate, it should only remove frequencies that are masked by other frequencies (and there you have a quick explanation about what psycho acoustics is about, heheh).
Because you can't compress a whole file in one huge DCT analysis, the audio is usually chopped up in blocks. Each block overlaps 50%. An inaccurate overlap might kill some frequencies (by inadvertantly creating a comb filter), but that would be a design error in the encoder/decoder, not a flaw of the MP3 format.
I'm not 100% sure about what follows, but: The remaining frequencies are then quantized (this produces artifacts) and the block is Huffman-compressed. During the whole process, there is a bitrate to adhere to, so there's something known as a bit pool, which is basically a "how many bits do we have left for encoding additional info about this audio" pool. Complex parts will get harder compression that simple parts. This is why VBR was invented. There's no bit pool, the needed number of bits is used all the time.
MP3 is more ingenious than one might think. One day, I'm going to figure out what it is that OGG does, apart from using VBR, that makes it better.
Slashdot Mirror
Isn't it blatantly obvious how this reflects the nature of all humans? Kindness directed towards strangers isn't exactly a common trait. There is even a theory to explain the its rare occurances; those with no prospect of reward: Humans are evolved in small social groups, where favours were more likely to pay off. These groups rarely encountered outsiders. Therefore, humans never developed the social instinct to differ between kin and stranger. The theory basically assumes that all humans are egoists. As a large company can be viewed as a collective conciousness of sorts, egoism is just a reflection of a basic human trait.
"Ignoranse" is how it's spelled in Norwegian... only we pronounce it "ee-gnoo-rahn-seh". XD
"The GIMP has a very good interface."
I *hope* that's a joke.
Ah! Thanks for clearing that up. I always wondered why the voltage drops when you pick up the phone. (Yes. I measured it.)
But is reduced to about a 10th of that or less whenever a call is put on.
Never said that, but it looks like WMA is winning out.
Actually, most music businesses seem to use WMA. Although OGG beats MP3, some tests show that WMA beats OGG. I've heard the opposite as well, though, in A/B listening tests published here on Slashdot a while ago. Many of my aquaintances are using WMA, because normal people don't know what OGG is. The über marketing skills of GNU geeks has also blesed us the ugliest codec name this side of the galaxy.
This is latin for "garbage in, garbage out", right?
Okay. Our game consoles don't have SCART connectors, but there's usually a Phono-to-SCART-adapter included. Most old TV sets only take input via SCART if there's any input but the antenna one. Modern TVs usually have Phono and S-Video connectors on the front, and 1 to 3 SCART inputs on the rear. The SCART system has a couple of clever features: For instance, good TVs will auto-switch to the right SCART input when equipment is turned on, and back again when it's turned off. If you tell your DVD player you have a 16:9 TV, the TV will switch formats automatically. On a good, regular 4:3 set, this means that the set does automatic letterboxing. Of course, another nice thing about living in Europe is PAL's picture quality. ;)
Funny. I've actually thought about a system like this. A digital replacement for the SCART connector. Do you guys over in America have SCART connectors?
Most of us here on Slashdot weren't around back then. I wasn't, but I think it's amazing still.
There's one thing I've always wanted to try: 1. Record audio digitally at a high sample length (say 32 bits). 2. Segment it into short blocks of maybe 1024 samples. 3. Normalize each block and store the amount of amplification together with each block. 4. Quantize to lower sample length (say 12 to 16 bits). 5. Store reduced data heap to medium. Wouldn't you be able to compress audio of quite high fidelity in this way, without actually having a noticable reduction in dynamic range nor quality? Introduce volume interpolation (smooth transitions) between each block and you'd constantly get the highest quality possible for each sample (the encoder and the decoder would interpolate in synch). The HDCD format uses the least significant bit of the signal to add extra dynamic information. The same method could be used to encode expansion information. Imagine this being used by mastering plants to improve sound quality. With special plugins, compression information could extracted in the studio, stored in the least significant bit of each sample, and if the listener wanted, he could cancel out studio compression with a touch of a button, if he had the right equipment. Of course it'll never happen. But it's fun to dream about. ;)
"Every day our father would kill us, and dance about on our graves."
Much more happens in a tape recorder than just the addition of noise. There is something called warping, that is inaccuracies in the speed the tape passes the playback/recording heads. There is crosstalk, that is mixing of audio across several channels. There is also the fact that the signal is being recorded with a bias to bring it into the most linear part of the tape. And then there's the most noticable effect: Compression. When you over-record a tape, you get compression, that is reduction in audio levels compared to the original levels. A digital recorder will just clip, sounding horrible. A tape recorder will do it more gently. Modern musicians and technicians are very fond of what's described as vintage sound. I need just mention the UREI LA-2A compressor, an opto-electric tube compressor, used on numerous recordings. I dare bet anyone who has ever listened to recorded music has heard the handywork of that machine, or it's digital emulations. I love the sound of warping, especially when it comes from record players. It's what I call a becoming untunedness or unstability in pitch. Creates a warm fuzzy feeling inside of me, at least. Tube equipment is popular for grunching things up a bit, the newest Korg synthesizer model, a purely digital machine in all other aspects, has a tube stage for adding an edge to the sound. Guitarists unanamously agree that tube amplifiers give the best sound... I can go on and on... But it's a part of musicians' culture, basically.
Well, I saw no other reason for mentioning it, so I wildly assumed creds were the reason. Fine. He created all that, and for that I am grateful. But I don't like the way he does PR. I'm embarassed on his behalf, because I just know that so many disagree with aspects of his arguments. If the public views him as some sort of representative for the Linux community... Well, I just don't like the thought of that, because people will often interpret opinionatedness as narrowmindedness and stupidity. He's hostile to the whole model of closed source programming and he doesn't care if a zillion programmers lose their jobs. I just dislike that strongly. Although I am a fan of open source and free software, I think they can coexist with closed proprietary software. I don't want to abolish it. He's just so extreme.
Hey, I did Linux in the 90s too. My computer teacher saw that I wasn't like the other kids, so he let me play with old hardware instead of following the regular program. I installed Slackware on it and learned tons. I don't see how this has anything to do with your geek credentials, though.
I'm just saying RMS seems to be this very opinionated guy. He seems obsessed and overdramatisizes everything. I'm not the only one who thinks he's harming more than he's helping with his attitude.
Why does it have to be so that every time there is a disagreement, ever so civilized, on Slashdot, somebody has to shout "idiot"? Isn't it possible to be factual? This is the sort of thing that makes me read Slashdot comments more for the sake of contemptuous entertainment than true insight.
I don't see why this guy is such an icon for so many geeks. He seems unsympathetic and makes outrageous black & white statements. I think he gives Linux a bad name.
*sings along* I'm a big Beatles fan! :D
Linus Torvalds is Finnish. The character is supposed to be an "o" with two dots on top of it. I don't know what you see.
Yes. It's called MPlayer.
I'm using Firefox. I bet it would display correctly if I told it manually that the page uses ISO-8859-1 or what ever. But this is Slashdot's responsibility. And don't come here and be nit-picky. You know as well as I do that I was speaking in the context of this message. In Norway, "æ", "ø" and "å" are called "Norwegian letters" even though they're used in Denmark too. I bet the Danes call them "Danish letters". Neither of us pick on the other for doing so. In this context, they are Nordic letters. So there!
Slashdot apparently can't handle nordic charsets. That's supposed to be 'glögg', as in 'juleglögg'.
You've noticed too? I swear, people, this is true. Outlook ignores viruses and blocks friendly files. My dad can testify about this. He got a bunch of e-mails, most of them spam, some had bad stuff in them, and Outlook didn't grey those out. However, when a friend sent him an MP3, that was greyed out for some stupid reason, and we had to disable the "protection". How is it possible to write such stupid software?
Um... Norway. The home country of DVD-Jon.
Hehe. The classic "LPs are better"-argument. I wish I had good LP equipment so I could actually go and compare things. Unfortunately I was born in 1983 and became part of the CD generation. I have never owned an LP in my life.
...) and even (N * 2, N * 4, N * 6), and these produce the characteristic sawtooth sound. DCT will not kill those harmonics. The psycho acoustic analysis following that might take away some of them, but at a good bitrate, it should only remove frequencies that are masked by other frequencies (and there you have a quick explanation about what psycho acoustics is about, heheh).
I also wish I had the money to buy or even build a tube amp, just to check how that sounds. I do have access to studio monitors. They tend to reveal a lot, because they're neutral. Most mixes sound cold and impersonal through them and I wonder if that is just my ears being used to listening to equipment that colors the audio to make it more interesting or if the sound picture would warm up if I connected a record player to them.
I can't say I'm 'blessed' with tin ears. I notice artifacts that many people don't hear even if I try to make them aware of them. My father accuses me of having fly eyes, because I can see the MPEG2 compression artifacts in the pictures from our DVB sattelite tuner.
If I ever get the chance to wire an oscilloscope up to a PC sound card's output, I'm definitely going to test various waves and see how accurately they come out.
As an experiment, have you tried digitizing some of your premium albums? Maybe you'll find that the digital recording will be able to capture the sound of the LP. It's popular belief these days that adding analog artifacts to a sound will make it sound 'warmer'. Maybe the manner in which LPs have to be mixed (the RCA stage) creates that particular sound that your ears like?
As for sawtooth waves being rounded off to sine waves, that would happen as the pitch gets close to the cutoff of the lowpass filter before the A/D, which is designed to cut off as close as possible to 22.05 Hz. Human hearing ends at about 20 kHz and most adults I've met (including myself) can barely hear any sine wave above 16 kHz.
The compression itself wouldn't cut off a sawtooth wave. It's just a DCT analysis (similar to FFT) that, if performed in inverse, should reproduce the original waveform. DCT converts from the time domain to the frequency domain, which in English means that it produces a frequency 'curve' of whatever waveform was processed.
A sawtooth wave will have many harmonics (sine waves with pitches that are multiples of the base frequency), both odd (N * 1, N * 3, N * 5,
Because you can't compress a whole file in one huge DCT analysis, the audio is usually chopped up in blocks. Each block overlaps 50%. An inaccurate overlap might kill some frequencies (by inadvertantly creating a comb filter), but that would be a design error in the encoder/decoder, not a flaw of the MP3 format.
I'm not 100% sure about what follows, but: The remaining frequencies are then quantized (this produces artifacts) and the block is Huffman-compressed. During the whole process, there is a bitrate to adhere to, so there's something known as a bit pool, which is basically a "how many bits do we have left for encoding additional info about this audio" pool. Complex parts will get harder compression that simple parts. This is why VBR was invented. There's no bit pool, the needed number of bits is used all the time.
MP3 is more ingenious than one might think. One day, I'm going to figure out what it is that OGG does, apart from using VBR, that makes it better.