What do you think the jitter rejection of a standard S/PDIF PLL is? I'll tell you: it's zero across the audible spectrum. Please read the datasheet for the Cirrus Logic 8416, which is by far the most popular digital audio receiver chip. The jitter attenuation is only -6dB at 20kHz. At 3kHz, it's +2dB! The PLLs in consumer electronics are a sad joke indeed.
Some random magnetic field will interfere with the cable and that will affect how the signal is received at the other end. This is basic EE stuff. How do you think an antenna works? Therefore a good digital cable would be one which is least susceptible to interference from stray magnetic fields, such as a well-shielded RG-6 of the type used for cable TV. Even more important is to have constant impedance along the transmission line, which is why BNC is much better than RCA connectors. Finally, a huge improvement would be to use a balanced cable, where the stray interference can be cancelled. Unfortunately, common consumer digital audio cables have none of these features. They are badly shielded, they don't have controlled impedance, and they aren't balanced.
I'm referring to the zero-crossing in the analog domain. That's the very worst case for S/PDIF because the encoding flips every bit in the word. Therefore crossing zero causes a maximal net current flow. Optical has the same problem because crossing zero causes the maximal net current draw from the power supply of the transmitter. But obviously you have no experience with any of these design issues, and your thoughts on the topic are based on pure speculation.
Yes, the way you have stated the problem sounds impossible. But the phase noise isn't at 700kHz, it's at low frequencies. Read DIGITAL-TO-ANALOG CONVERTER WITH LOW INTERSAMPLE TRANSITION DISTORTION AND LOW SENSITIVITY TO SAMPLE JITTER AND TRANSRESISTANCE AMPLIFIER SLEW RATE, Hawksford, M.O.J., JAES, vol. 42, no. 11, pp 901-917, November 1994. On page 904 you can clearly see that the jitter spectrum is well within the audible range, well below 2kHz even. Hawksford also clearly derives the relationship between jitter in the time domain and harmonic distortion in the frequency domain.
With optical, there is no such thing as higher or lower quality cable. In that sense, optical is superior to copper. But the optical receivers and transmitters which convert the signal from and to electrical signals are quite sensitive to the quality of their power supply, and because of that systems using optical interconnects normally have higher distortion on the analog outputs than systems using a good copper interconnect. And there's nothing about optical which ameliorates the fundamental design issues of S/PDIF.
I'm sorry, but your simplistic extrapolation of information theory simply does not model reality. The frequency of the phase noise is independent of the signal frequency. In fact, because of the sad design of S/PDIF, the phase noise is correlated with the signal, which is the worst possible case. In fact, the worst induced phase noise is on the zero crossing, which means that the lowest level signals have the highest distortion.
On the off chance that you are capable of educating yourself on this topic, I leave you with this link regarding phase noise: http://www.wenzel.com/documents/noise.html
Wow, there is no shortage of you ignorant dudes on Slashdot, is there? I'm not talking about bit-flipping errors. I'm talking about the fact that the broken-ass S/PDIF protocol relies heavily on the lack of phase noise in the inter-symbol arrival time on the wire. Therefore it also relies heavily on perfect transmission. Therefore you need a good cable. You say that the clock isn't generally carried on the cable, but you're wrong. Generally, the clock *is* carried on the cable. And most consumer electronics have clock recovery circuits that are so bad as to be jokes.
Ah yes, yet a third ignorant yokel, holding forth on topics he doesn't understand. S/PDIF is not ethernet. The data and the time of arrival of that data are both key to the protocol. Yes, the protocol is broken. But that means you need a good cable! Also I'm sorry to inform you that your understanding of "high frequency" is incorrect. The important part is not the symbol rate, but the rise and fall time. As transmitters have moved to newer and faster logic, the frequency spectrum on a digital audio cable (and any other digital cable) has actually become higher and higher over the years. You should consider an S/PDIF cable to be working in the domain of 100MHz or thereabouts, because of the very fast rise times we have in modern driver circuits.
Like the previous poster, you are woefully misinformed. Consumer audio equipment reads the clock straight off the wire. If the last bit arrives at the wrong time, then the analog signal is produced at the wrong time, which means you have distortion in the frequency domain. The quality of digital audio relies entirely on low phase noise of the sample clock. If you have phase noise on the sample clock, then you have harmonic distortion in the analog output. QED. This is all so basic, it's incredible that none of you understand it. Don't they teach anything in EE classes any more?
Your ignorance on this topic is staggering. Consumer audio equipment uses the inter-symbol arrival time to drive its internal clock. That's why interference on the cable is bad. Distortion in the time domain maps directly into distortion in the frequency domain due to the digital-to-analog convergence.
I hate to spark an argument, but I think the cable is actually far more important with digital than with analog. You can get a quite accurate representation of an analog audio signal at the far end of practically any conductor. Literally any wire will be a fairly good analog audio connector. But digital is different: with a digital signal the frequencies in transmission are much higher, and the signal is much more likely to be distorted or interfered with. The result is that the inter-edge arrival time will be distorted, and this, in turn, maps directly into harmonic distortion in the analog reproduction.
There are ways around this, including buffering and reclocking the digital signal at the receiver, but in general this step is not taken in consumer electronics. So you'll be well advised to spend a few dollars (not thousands) on a good piece of coax with BNC connectors, or another kind of good digital cable, if you plan to use a digital signal.
Optical isn't any better, because the optical receivers and transmitters are extremely sensitive to noise on their power supplies.
These fonts might be nice, but FreeType makes a complete hash of them. I tried Consolas at various sizes and FreeType makes the vertical lines variously too light and too dark. A sentence looks like a ransom note.
In the positive column for free software, Wine was able to run the PowerPoint installer perfectly.
You needn't worry about the kernel in this case, because the applications aren't paying any attention. Mozilla happily passes all open file descriptors (sockets, pipes, and files of any kind) to subprocesses like Adobe Reader. There's been a bug open on it for eons. Other applications have the same problems. It may be convenient to have O_CLOEXEC in open(2) calls, but it won't help of the application writers don't know what they are doing, or if they have "abstracted" their platform interaction to such a degree that they can no longer interact with any platform services (*cough* jvm *cough*).
The OLPC no longer has the crank. It can be charged by solar or by a pull-string charger but unless you have your own crank, you can't charge it that way.
eMusic is certainly not a mainstream music retailer. They don't sell you MP3s the way the grocer sells you a melon. You have to sign up for a month and you're allowed to download a song a day, roughly, although nobody does that. I can go to Amazon and spend 89c on a single song and never return. At eMusic, I have to pay $9.99 at least and then I have to remember to cancel it if I don't want it any more.
Re:Is this really different from the RIAA or MPAA?
on
GPL Lawsuit May Not Settle
·
· Score: 5, Insightful
If you think distributing GPL software is bad for your business, try distributing pirate copies of Microsoft Windows with your product. You'll get to see just how "dangerous" a license can be.
The fact is that infringing other people's copyrights is a bad business move. It does not matter at all what mechanism the rightsholders are using to protect their copyrights.
And incorrectly. Among foxit's many, numerous bugs is the inability to correctly number the pages, which makes it pretty fucking hard to discuss the document with someone who uses foxit, because eventually you will say "It's on page 8" and they will say "page 8 is blank!" and then you will find out that they use foxit and then you will just have to stop talking to them or possibly beat them to a bloody pulp for wasting so much of your time.
Not that I'm bitter. I'm just really tired of people saying "XYZ is faster than ABC" when the reason XYZ is faster is because it does half of everything incorrectly.
On my machine Beagle doesn't *ever* finish. It runs, for about a month, over my millions of files, until it finds one that makes it crash, and then it crashes in a loop. I thought the whole point of mono was that it can't crash and it's safe from overflows and other C problems. But in practice the implementation is appalling.
In constract Tracker can index my huge home directory in less than a day without crashing.
I'm sure you mean Gutsy. I just updated my Feisty machine and there's nothing new, and I wouldn't expect a new major GNOME release in an existing Ubuntu distribution.
I think you answered your own question. They insist on shipping Tomboy so they can have a reason to ship mono. Without Tomboy, and without Beagle, the search tool thousands of times more idiotic than Tracker, there would be no reason for anybody to install mono.
All previous GNOME releases had a problem where the volume control would wake up every 100ms to poll the mixer settings, which prevented the CPU from entering and maintaining deep sleep states. The new volume control does not do that, which may be good for a few watts at the outlet. Other applications have undergone a similar treatment.
Slashdot Mirror
Datasheet, which you have obviously never read. Please see page 55.
Some random magnetic field will interfere with the cable and that will affect how the signal is received at the other end. This is basic EE stuff. How do you think an antenna works? Therefore a good digital cable would be one which is least susceptible to interference from stray magnetic fields, such as a well-shielded RG-6 of the type used for cable TV. Even more important is to have constant impedance along the transmission line, which is why BNC is much better than RCA connectors. Finally, a huge improvement would be to use a balanced cable, where the stray interference can be cancelled. Unfortunately, common consumer digital audio cables have none of these features. They are badly shielded, they don't have controlled impedance, and they aren't balanced.
Actually what you said is "you're not going to get harmonic distortion" which is the opposite of reality. Here's the peer-reviewed watershed work on jitter and harmonic distortion, published in the Journal of the AES. Educate yourself. http://www.essex.ac.uk/ESE/research/audio_lab/malcolmspubdocs/C44%20DAC%20Transisiton%20distortion,%20jitter,%20slew%20rate.pdf
I'm referring to the zero-crossing in the analog domain. That's the very worst case for S/PDIF because the encoding flips every bit in the word. Therefore crossing zero causes a maximal net current flow. Optical has the same problem because crossing zero causes the maximal net current draw from the power supply of the transmitter. But obviously you have no experience with any of these design issues, and your thoughts on the topic are based on pure speculation.
Yes, the way you have stated the problem sounds impossible. But the phase noise isn't at 700kHz, it's at low frequencies. Read DIGITAL-TO-ANALOG CONVERTER WITH LOW INTERSAMPLE TRANSITION DISTORTION AND LOW SENSITIVITY TO SAMPLE JITTER AND TRANSRESISTANCE AMPLIFIER SLEW RATE, Hawksford, M.O.J., JAES, vol. 42, no. 11, pp 901-917, November 1994. On page 904 you can clearly see that the jitter spectrum is well within the audible range, well below 2kHz even. Hawksford also clearly derives the relationship between jitter in the time domain and harmonic distortion in the frequency domain.
With optical, there is no such thing as higher or lower quality cable. In that sense, optical is superior to copper. But the optical receivers and transmitters which convert the signal from and to electrical signals are quite sensitive to the quality of their power supply, and because of that systems using optical interconnects normally have higher distortion on the analog outputs than systems using a good copper interconnect. And there's nothing about optical which ameliorates the fundamental design issues of S/PDIF.
On the off chance that you are capable of educating yourself on this topic, I leave you with this link regarding phase noise: http://www.wenzel.com/documents/noise.html
That's just a typo. I meant conversion, not convergence.
Wow, there is no shortage of you ignorant dudes on Slashdot, is there? I'm not talking about bit-flipping errors. I'm talking about the fact that the broken-ass S/PDIF protocol relies heavily on the lack of phase noise in the inter-symbol arrival time on the wire. Therefore it also relies heavily on perfect transmission. Therefore you need a good cable. You say that the clock isn't generally carried on the cable, but you're wrong. Generally, the clock *is* carried on the cable. And most consumer electronics have clock recovery circuits that are so bad as to be jokes.
Ah yes, yet a third ignorant yokel, holding forth on topics he doesn't understand. S/PDIF is not ethernet. The data and the time of arrival of that data are both key to the protocol. Yes, the protocol is broken. But that means you need a good cable! Also I'm sorry to inform you that your understanding of "high frequency" is incorrect. The important part is not the symbol rate, but the rise and fall time. As transmitters have moved to newer and faster logic, the frequency spectrum on a digital audio cable (and any other digital cable) has actually become higher and higher over the years. You should consider an S/PDIF cable to be working in the domain of 100MHz or thereabouts, because of the very fast rise times we have in modern driver circuits.
Like the previous poster, you are woefully misinformed. Consumer audio equipment reads the clock straight off the wire. If the last bit arrives at the wrong time, then the analog signal is produced at the wrong time, which means you have distortion in the frequency domain. The quality of digital audio relies entirely on low phase noise of the sample clock. If you have phase noise on the sample clock, then you have harmonic distortion in the analog output. QED. This is all so basic, it's incredible that none of you understand it. Don't they teach anything in EE classes any more?
Your ignorance on this topic is staggering. Consumer audio equipment uses the inter-symbol arrival time to drive its internal clock. That's why interference on the cable is bad. Distortion in the time domain maps directly into distortion in the frequency domain due to the digital-to-analog convergence.
I hate to spark an argument, but I think the cable is actually far more important with digital than with analog. You can get a quite accurate representation of an analog audio signal at the far end of practically any conductor. Literally any wire will be a fairly good analog audio connector. But digital is different: with a digital signal the frequencies in transmission are much higher, and the signal is much more likely to be distorted or interfered with. The result is that the inter-edge arrival time will be distorted, and this, in turn, maps directly into harmonic distortion in the analog reproduction.
There are ways around this, including buffering and reclocking the digital signal at the receiver, but in general this step is not taken in consumer electronics. So you'll be well advised to spend a few dollars (not thousands) on a good piece of coax with BNC connectors, or another kind of good digital cable, if you plan to use a digital signal.
Optical isn't any better, because the optical receivers and transmitters are extremely sensitive to noise on their power supplies.
These jokers didn't even mention the most important part of a home theater: comfy chairs.
These fonts might be nice, but FreeType makes a complete hash of them. I tried Consolas at various sizes and FreeType makes the vertical lines variously too light and too dark. A sentence looks like a ransom note.
In the positive column for free software, Wine was able to run the PowerPoint installer perfectly.
You needn't worry about the kernel in this case, because the applications aren't paying any attention. Mozilla happily passes all open file descriptors (sockets, pipes, and files of any kind) to subprocesses like Adobe Reader. There's been a bug open on it for eons. Other applications have the same problems. It may be convenient to have O_CLOEXEC in open(2) calls, but it won't help of the application writers don't know what they are doing, or if they have "abstracted" their platform interaction to such a degree that they can no longer interact with any platform services (*cough* jvm *cough*).
The OLPC no longer has the crank. It can be charged by solar or by a pull-string charger but unless you have your own crank, you can't charge it that way.
eMusic is certainly not a mainstream music retailer. They don't sell you MP3s the way the grocer sells you a melon. You have to sign up for a month and you're allowed to download a song a day, roughly, although nobody does that. I can go to Amazon and spend 89c on a single song and never return. At eMusic, I have to pay $9.99 at least and then I have to remember to cancel it if I don't want it any more.
If you think distributing GPL software is bad for your business, try distributing pirate copies of Microsoft Windows with your product. You'll get to see just how "dangerous" a license can be.
The fact is that infringing other people's copyrights is a bad business move. It does not matter at all what mechanism the rightsholders are using to protect their copyrights.
And incorrectly. Among foxit's many, numerous bugs is the inability to correctly number the pages, which makes it pretty fucking hard to discuss the document with someone who uses foxit, because eventually you will say "It's on page 8" and they will say "page 8 is blank!" and then you will find out that they use foxit and then you will just have to stop talking to them or possibly beat them to a bloody pulp for wasting so much of your time.
Not that I'm bitter. I'm just really tired of people saying "XYZ is faster than ABC" when the reason XYZ is faster is because it does half of everything incorrectly.
I have only one thing to say about Rails and that is "waiting for twitter.com..."
On my machine Beagle doesn't *ever* finish. It runs, for about a month, over my millions of files, until it finds one that makes it crash, and then it crashes in a loop. I thought the whole point of mono was that it can't crash and it's safe from overflows and other C problems. But in practice the implementation is appalling.
In constract Tracker can index my huge home directory in less than a day without crashing.
I'm sure you mean Gutsy. I just updated my Feisty machine and there's nothing new, and I wouldn't expect a new major GNOME release in an existing Ubuntu distribution.
I think you answered your own question. They insist on shipping Tomboy so they can have a reason to ship mono. Without Tomboy, and without Beagle, the search tool thousands of times more idiotic than Tracker, there would be no reason for anybody to install mono.
All previous GNOME releases had a problem where the volume control would wake up every 100ms to poll the mixer settings, which prevented the CPU from entering and maintaining deep sleep states. The new volume control does not do that, which may be good for a few watts at the outlet. Other applications have undergone a similar treatment.