That requires 1000+ PCs and the associated software licenses! For such a large project, I think you want to use a proprietary hardware system. A sort of hook it up and it works type thing. All of the those PCs will have to be maintained too!
Is 30 fps really necessary? Is 640 x 480 necessary. At 300kb/s, that is broadcast quality. That is the compression range that expensive, proprietary NLEs like Avid and Media 100 use. Really I think you can get away with 10 fps and 320x240.
A lot of people in the discussion are talking about compression. Can you imagine the size of the encoding farm that would be able to process that many video streams and turn it into a Divx, MPEG, etc. You get good compression from a product such as Sorenson, but I wouldn't call it a fast process. So, I think you want to avoid compression. If you use compression, make sure it is hardware based - embedded on the camera. Remeber to that by reducing the picture size to 320x240, you have eliminated 75% of the data already. Reduce it to 10 fps and you have eliminated over 90% of the original data.
Also, if you are not getting a digital signal from these cameras, then you will need 1000+ capture cards. If you are getting something like a DV signal over Firewire then you still have a conversion process that must take place. DV runs at about 3.5 MB/s. That's still a high data rate for the number of cameras that you have.
Really, this doesn't seem feasible without a huge budget. If you have the money though, I'd love to consult on this project!
Oh, and one more thing. Look at your compressed file compared to the uncompressed file in a good audio editor where you can zoom in very tight on the waveform. There is some visible difference in the way the waveform is reconstructed following the compression process.
You couldn't be more correct. Recently, the quality of different audio codecs became a topic in a Digital Audio class I was taking. The feeling of the instructor and the class was that most compression schemes, such as MP3, degrade the quality of the audio. Of course, perhaps they're are all audio snobs, particularly because we were working with digital microphones and complete ProTools systems.
So for my final project in the class, I decided to do a blind taste test of sound. Working with a Macintosh (anything else would be uncivilized), I ripped a variety of tunes from many eras in many styles. I then converted them to AIFF files and burned them to a CD. I played the CD in "Control Room A' at school. Listening to the same sample with different compression schemes, I asked them to identify preference.
Basically, 160 kbs joint stereo with a decent compressor is where most people's tolerance ends. Below that, most people are able to detect a difference. Some are not. Individuals demostrated varying degrees of tolerance. When asked to rank a series of tracks, some answered "correctly" every time. Some people were way off.
I used an Incubus sample. I played a 96 joint stereo sample followed by an uncompressed sample. Some people said that they thought the first one was better. I concluded that tolerance for audio quality is widely varied, but that there is a point where people are able to detect a difference in audio quality that they object to. Some of the subjects had very astute ears!
In other words, if you are serious about your audio collection, as many recovering Napsterholics are, then you should test your tolerance level for audio fidelity.
To quote from benedict: "So basically, MOSX/Darwin aren't using Mach as a microkernel, they're using it as the basis for a nicely modular but "monolithic" kernel."
That seems like a good way of putting it. Apple's own documentation (Inside Mac OS X: System Overview) states that Mach handles "processor resources such as CPU usage and memory, handles scheduling, enforces memory protection," and handle interprocess communication. "Integrated with Mach is a customized version of BSD... BSD serves as the basis for file systems and networking facilities of Mac OS X." Also provides the process model, basic security (file system), POSIX threads, BSD sockets, and kernal APIs.
"Darwin integrates a number of technologies, most importantly Mach 3.0, operating system services based on 4.4BSD...... Darwin is highly modular, you can dynamically add such things as device drivers, networking extensions, and new file systems."
"The Mac OS X kernal environment is a subset of Darwin. The kernal environment contains everything in Darwin except the BSD libraries and commands that are essential to the BSD Commands environment."
This is taken in random snippets from the May 2000 revision of this book. Some info may have changed.
I think the confusion is that Mac OS X actually runs Darwin (which I believe Carnegie Mellon University worked on). I might be wrong as well, but Darwin is a merger of unadulterated BSD with an integrated Mach microkernal. This was chosen, I assume, for its extensibility over time as hardware changes, and because Apple needed something that talked to Mac hardware. BSD doesn't. Overall though, we shouldn't be so concerned about the minor inefficiencies of a microkernal. As hardware continues to get faster, there won't be a problem running fatter code. Or allowing different layers of an OS the ability to communicate with each other, even if it might be slightly inefficient. It is stable.
Software developers should be more concerned with extensibility, rapid development, reusable code, etc. That's the idea behind OpenStep and now Cocoa. Get the software to the people and let the chip engineers build faster hardware.
Re:It's official: Apple has become Microsoft...
on
OS X
·
· Score: 1
First, what is so confusing about 10 (X). The last big number release was 9; 10 comes after 9. System 7.5 was called that because it was bigger than a.1 upgrade but not big enough for a full number upgrade. Makes good sense to me. Apple has kept the same relative naming scheme for 17 years and it seems that they are continuing. Besides, no one at Apple is calling it Mac OS X Ten 10.0. It is version 10 of the OS, 10.0 to be exact. X is a good marketing strategy, especially because of its roots in uniX.
Apple did get a lot of flak for charging for the beta. The purpose: to keep everyone from trying it. They only wanted feedback from a sample of their market. The public beta was really research in disguise. The price was to make what they were doing feasible and not overly time-consuming. The price was an intentional deterrent. The target market, until July, for OS X is Mac aficianados and Unix geeks. They don't want iMac owners to get it yet. They're trying to change everything about their user experience and do it smoothly without alienating the entire market.
And are you using OS X? Because my system is anything but sluggish and buggy. I love OS X. The only thing better might be a smoke after sex.
You've got to be kidding me. Are you trying to compare the cost of a new Macintosh to 5+ year old legacy Pentium hardware. Yes, it is cheaper, now. We could even compare prices over time, I mean a TRS-80 was much cheaper then, than a new Mac is today. (even if we take inflation into account) That must make it a better machine. Twisted logic like this is ripping apart the world.
And besides, what are the people in third world countries going to do with their $200 Pentium box, run Linux? These people don't know how to use a faucet, or a TV remote. I sure they'll understand the world's endlessly customizable, completely confusing, and easily mangled OS... Linux.
Apple has delivered. They have provided a powerful, stable OS for geeks. (At the terminal, it really is just BSD.) And a simple, intuitive GUI with balanced performance for consumers.
Their is a reason stop signs are large, red, simple, and the same everywhere. Most people are dumb. Most people don't sit up at 2:42 AM clarifying your twisted logic on slashdot. Most people just want their computer to work.
If the Linux community wants their OS to proliferate, they will have to follow Apple's lead and deliver an OS that anyone can use. Meanwhile, back at the ranch, the vine-ripened tomatoes always cost more than the unripe yellow/brown/red crap next to them. "Thank you waiter, I'll have a consistent hardware platform and I would like my operating system well done."
Slashdot Mirror
That requires 1000+ PCs and the associated software licenses! For such a large project, I think you want to use a proprietary hardware system. A sort of hook it up and it works type thing. All of the those PCs will have to be maintained too!
Is 30 fps really necessary? Is 640 x 480 necessary. At 300kb/s, that is broadcast quality. That is the compression range that expensive, proprietary NLEs like Avid and Media 100 use. Really I think you can get away with 10 fps and 320x240.
A lot of people in the discussion are talking about compression. Can you imagine the size of the encoding farm that would be able to process that many video streams and turn it into a Divx, MPEG, etc. You get good compression from a product such as Sorenson, but I wouldn't call it a fast process. So, I think you want to avoid compression. If you use compression, make sure it is hardware based - embedded on the camera. Remeber to that by reducing the picture size to 320x240, you have eliminated 75% of the data already. Reduce it to 10 fps and you have eliminated over 90% of the original data.
Also, if you are not getting a digital signal from these cameras, then you will need 1000+ capture cards. If you are getting something like a DV signal over Firewire then you still have a conversion process that must take place. DV runs at about 3.5 MB/s. That's still a high data rate for the number of cameras that you have.
Really, this doesn't seem feasible without a huge budget. If you have the money though, I'd love to consult on this project!
Oh, and one more thing. Look at your compressed file compared to the uncompressed file in a good audio editor where you can zoom in very tight on the waveform. There is some visible difference in the way the waveform is reconstructed following the compression process.
You couldn't be more correct. Recently, the quality of different audio codecs became a topic in a Digital Audio class I was taking. The feeling of the instructor and the class was that most compression schemes, such as MP3, degrade the quality of the audio. Of course, perhaps they're are all audio snobs, particularly because we were working with digital microphones and complete ProTools systems.
So for my final project in the class, I decided to do a blind taste test of sound. Working with a Macintosh (anything else would be uncivilized), I ripped a variety of tunes from many eras in many styles. I then converted them to AIFF files and burned them to a CD. I played the CD in "Control Room A' at school. Listening to the same sample with different compression schemes, I asked them to identify preference.
Basically, 160 kbs joint stereo with a decent compressor is where most people's tolerance ends. Below that, most people are able to detect a difference. Some are not. Individuals demostrated varying degrees of tolerance. When asked to rank a series of tracks, some answered "correctly" every time. Some people were way off.
I used an Incubus sample. I played a 96 joint stereo sample followed by an uncompressed sample. Some people said that they thought the first one was better. I concluded that tolerance for audio quality is widely varied, but that there is a point where people are able to detect a difference in audio quality that they object to. Some of the subjects had very astute ears!
In other words, if you are serious about your audio collection, as many recovering Napsterholics are, then you should test your tolerance level for audio fidelity.
To quote from benedict: "So basically, MOSX/Darwin aren't using Mach as a microkernel, they're using it as the basis for a nicely modular but "monolithic" kernel." That seems like a good way of putting it. Apple's own documentation (Inside Mac OS X: System Overview) states that Mach handles "processor resources such as CPU usage and memory, handles scheduling, enforces memory protection," and handle interprocess communication. "Integrated with Mach is a customized version of BSD... BSD serves as the basis for file systems and networking facilities of Mac OS X." Also provides the process model, basic security (file system), POSIX threads, BSD sockets, and kernal APIs. "Darwin integrates a number of technologies, most importantly Mach 3.0, operating system services based on 4.4BSD... ... Darwin is highly modular, you can dynamically add such things as device drivers, networking extensions, and new file systems."
"The Mac OS X kernal environment is a subset of Darwin. The kernal environment contains everything in Darwin except the BSD libraries and commands that are essential to the BSD Commands environment."
This is taken in random snippets from the May 2000 revision of this book. Some info may have changed.
I think the confusion is that Mac OS X actually runs Darwin (which I believe Carnegie Mellon University worked on). I might be wrong as well, but Darwin is a merger of unadulterated BSD with an integrated Mach microkernal. This was chosen, I assume, for its extensibility over time as hardware changes, and because Apple needed something that talked to Mac hardware. BSD doesn't. Overall though, we shouldn't be so concerned about the minor inefficiencies of a microkernal. As hardware continues to get faster, there won't be a problem running fatter code. Or allowing different layers of an OS the ability to communicate with each other, even if it might be slightly inefficient. It is stable. Software developers should be more concerned with extensibility, rapid development, reusable code, etc. That's the idea behind OpenStep and now Cocoa. Get the software to the people and let the chip engineers build faster hardware.
First, what is so confusing about 10 (X). The last big number release was 9; 10 comes after 9. System 7.5 was called that because it was bigger than a .1 upgrade but not big enough for a full number upgrade. Makes good sense to me. Apple has kept the same relative naming scheme for 17 years and it seems that they are continuing. Besides, no one at Apple is calling it Mac OS X Ten 10.0. It is version 10 of the OS, 10.0 to be exact. X is a good marketing strategy, especially because of its roots in uniX.
Apple did get a lot of flak for charging for the beta. The purpose: to keep everyone from trying it. They only wanted feedback from a sample of their market. The public beta was really research in disguise. The price was to make what they were doing feasible and not overly time-consuming. The price was an intentional deterrent. The target market, until July, for OS X is Mac aficianados and Unix geeks. They don't want iMac owners to get it yet. They're trying to change everything about their user experience and do it smoothly without alienating the entire market.
And are you using OS X? Because my system is anything but sluggish and buggy. I love OS X. The only thing better might be a smoke after sex.
You've got to be kidding me. Are you trying to compare the cost of a new Macintosh to 5+ year old legacy Pentium hardware. Yes, it is cheaper, now. We could even compare prices over time, I mean a TRS-80 was much cheaper then, than a new Mac is today. (even if we take inflation into account) That must make it a better machine. Twisted logic like this is ripping apart the world. And besides, what are the people in third world countries going to do with their $200 Pentium box, run Linux? These people don't know how to use a faucet, or a TV remote. I sure they'll understand the world's endlessly customizable, completely confusing, and easily mangled OS... Linux. Apple has delivered. They have provided a powerful, stable OS for geeks. (At the terminal, it really is just BSD.) And a simple, intuitive GUI with balanced performance for consumers. Their is a reason stop signs are large, red, simple, and the same everywhere. Most people are dumb. Most people don't sit up at 2:42 AM clarifying your twisted logic on slashdot. Most people just want their computer to work. If the Linux community wants their OS to proliferate, they will have to follow Apple's lead and deliver an OS that anyone can use. Meanwhile, back at the ranch, the vine-ripened tomatoes always cost more than the unripe yellow/brown/red crap next to them. "Thank you waiter, I'll have a consistent hardware platform and I would like my operating system well done."