If everyone listened to his advice, they'd at least have a good picture of the real requirements of the project they were undertaking, which would leave them in a much better position to learn while doing than just plunging in without a clear definition of problem.
And besides that, a smart person could learn alot from doing the requirements gathering and then working closely with an SI to define and implement a solution.
You do realize that every food crop/animal that man has raised since the dawn of time has been slectively bred to produce higher yields, disease resistance, and/or other physical traits (Dog and Cat breeds look nothing like their ancestors, neither does your baked potato). GE/GM crops are just allowing us to add factors that would normally take millenia to add.
And you do realize that a millenia gives a lot of time to identify and weed out undesirable side-effects?
With a decent blogging tool I can post a link to a page with a text excerpt and some brief commentary with minimal effort:
1. Select text 2. Click blogging bookmarklet. 3. Add any comments 4. Click the post button
I can't do that with notepad. Of course, I can do it with free software.
Other things that take more work with notepad. 1. Cycling stuff off my front page. 2. Creating archive pages. 3. Creating navigation by topic 4. Keeping a consistent template for all my blog pages 5. Syndicating blog content in RSS and/or other formats.
I tend to agree that Slashdot isn't the best place for this type of question, when there have to be thousands of real-world and online support groups out there. On the other hand, the person asking the question obviously has an affinity with the Slashdot community, and so experiences of Slashdotters may have more resonance for them than say that of a bunch of random AOL users.
It is really refreshing to see someone so willing to demonstrate their wrongheaded ignorance. Saves us all a lot of trouble.
I've found most of the people on Hydrogenaudio to be incredibly pragmatic. Perfection isn't the only parameter of importance. If it were, they'd not be wasting time testing codecs at 128kbps, except to demonstrate their unsuitability compaired to losless formats. They'd not be wasting time letting phillistines with their waxy untrimmed ears particpate in listening tests with their $20 sony earbuds.
As for the vendors lauding useless gear, um, what vendors lauding useless gear?
But hell, why let any of that get in the place of a perfectly good piece of ranting rhetoric. Still, it would be better if you'd unloaded at a deserving target. There are certainly enough of them out there.
While true that at 128kbps the differences will be more obvious to untrained ears, I wouldn't want people to think that the results of this this test will says anything about the relative performance of a given codec at a higher or lower bitrate.
A given codecs performance relative to other codecs at a given bitrate doesn't necessarily predict the performance of the same codec at another bitrate. For example, the codec implementation may be optimized for 128 kbps (something people suspect of the Apple AAC codec, for example).
The guy arranging the listening test at Hydrogen audio is well aware of this fact, so I think its rather likely that 128kbps wasn't chosen simply because its easier for untrained ears to hear the difference.
I think it was chosen because it is "disputed territory" where audio quality is acceptable to most users under causual listening conditions at a reasonably low bitrate.
I'd read the thread when they were discussing which version of Apple's ACC codec to use for the test, and concluded based on a few samples that the new version was subpar.
If they'd included both versions of iTunes/QuickTime in this test, perhaps they could have helped shame Apple into fixing what they broke.
And how do you know what you are asserting? Have you done properly controlled listening tests with 128kbps encoding using a variety of codecs?
The fact is that for a lot of people, knowing the best codec at 128kbps is worth knowing because:
1) They are using portable devices where they are space constrained 2) They are using portable devices that may not have the perfect fidelity of a high-end sound system, but can go anywhere with them. 3) They are using their portable device in a somewhat noisy environment that overshadows any sound quality issues caused by a lower bitrate.
And, the bitrate isn't a moot point. Download speed is only one consideration. Bitrate also directly impacts filesize so if you are storage constrained, like say a solid state digital media player, lower bitrate performance is important.
Different codecs and implementations of those codecs may be optimized for different bitrates, so its important to test codecs at various target bitrates.
SGIs early version of the Reality Engine used multiple units of the intel i860, a RISC chip and one of intels many attempts to move beyond the x86, to do the serious crunching. If one reads the paper in IEEE Micro when the i860 was released, one can see that it was positioned as a general purpose CPU for engineering workstations. I don't think it was used in said capacity in anything other than a reference design.
This way we wouldn't have to keep getting a new video card every time we want to upgrade our systems 3-d performance.
How much do you think you'd save that way? $20, maybe? GPUs are usually paired closely with the graphics memory subsystem. I'd guess that GPU + memory amounts to most of the cost of goods in higher end vid card.
At this point there isn't much of a market for massive parallel floating point (or integer) performance for anything but 3D graphics, much to the chagrin of CPU makers, so its not clear to me what incentive GPU makers have to produce a more general purpose and more commodified product.
I'm trying to imagine how the market might evolve from here.
Clearly there are some computationally intensive tasks that could benefit from a more general purpose math unit, but not as general purpose as todays CPUs.
Some of those tasks are of interest to a relatively small number of people, but those people will buy as much as they can afford. If they could add four or six big vector units per PC in a commodity cluster, they'd do it. So, while it may be a relatively small market in terms of # of customers, they are likely to buy a disproportionate number of units.
Some of the those tasks are of interest to a much larger number of people, but those people may be served, and satisfied, with just being able to tap the lone GPU on their lone vid card for things like video or audio encoding.
Taken together, the combination of the two markets may be big enough to be interesting to one or more GPU vendors, and both markets would benefit from a standardized general programming model.
What is less clear is whether both markets really demand an architecture that supports multiple GPUs per system for non real-time 3d tasks. Of course, the successors to PCI and AGP may make this question moot, PCs might have suitable infrastructure for supporting multiple cards without any special effort on the part of GPU manufacturers.
It seems likely that we'll see broader support for using GPUs for non-traditional tasks. If you think about it, the GPU makers stole a big piece of "instruction share" (% of total instructions executed in a PC) from Intel and other CPU makers, and along with it, they captured revenue growth that would have otherwise gone to CPU makers. Agreeing to some level of standardization in programming models (a la DirectX & OpenGL) and supporting the use of GPUs for non-traditional tasks will allow the GPU makers a chance at grabbing even more revenue growth away from the CPU makers.
A via isn't going to be able to do some things, like encode MPEG4 real time, at least not at their current clockspeeds, and it isn't going to do much for modern games. I'm also not sure it will handle playback of a HiDef MPEG4 source, or if its up to running something like descaler.
That said, its probably fine if you are just doing playback, or using a hardware MPEG2 encoder.
I'm actually debating whether to use a VIA for my HTPC. My plan is to use tuner cards with hardware MPEG2 encoders, so a Via should be fine, but i'm limited if I want to do anything different.
A big part of the motivation for Star Trek was that Moto was late late late with the 030 (due at least in part to a patent licensing issue) and there was a chance the CPU wouldn't ship at all, leaving apple a generation behind.
The story that Apple realizes profit from the iTMS not by selling songs, but by driving iPod sales is a good one, and no doubt there is a lot of truth to it, but I think it is foolish to take it at face value.
Right now, its in Apple's interest to make sure iTMS appears like an unattractive business to get into, because it discourages potential competitors, and the investors who might fund competitors. Meanwhile, the iPod story keeps Apple investors happy.
In time though, as volumes increace, as their initial investment is recouped, as they improve efficiency and lower costs, and as they negotiate better terms with record labels, their story will likely change.
I would wager that if LCD and DLP projection TVs really take off in volume, and there is a real attempt to standardize bulbs, that prices on bulbs will come down.
Right now, its a fairly small market, even taking into consideration all the conference room projectors. However, if a significant number of middle-class homes worldwide end up with these things, we are talking whole new economies of scale.
She was your aunt, and its your money, but if I wanted to honor somone for introducing me to technology, I don't think I'd put the money into hooking up a bunch of guys who already have ready access to tech (gamers).
I'd probably play "stone soup" with it and use it to attract other funds and direct them to a more community minded project that might actually introduce people to technology.
1. Why are you tied to DSL. I'm aware of some of its theoretical advantages, and in my market, its the only way to get more than 128 kbps up, but technology evolves, as does the way companies tune and deploy it, so it may be worth looking at a bundle of Cable IP & TV again.
2. Cable systems and satelites are both bandwidth constrianed, and HDTV takes a lot of bandwidth. Even before HDTV, satelite and digital cable providers were overcompressing stuff just so they could fit in a few more pay per view channels. One result is that their HDTV quality is often below that you could get with a decent broadcast antenna. Not only that, but it might get better or worse depeding on whatever else they are trying to cram down the pipe.
On the other hand, there is a tendancy for technologists to look at new applications of technology and say: Thats no big deal, its just a perl script, or, thats no big deal, its just a portable hard disk. Bittorrent is nothing, its just a python script. This jaded attitude can blind them to the implications of technology change.
Perhaps there is nothing new about this, but I'd suggest that it is an extreme illustration of a change that may have real implications.
With the rise of the iPod, and similar devices, people are now in the habit of carrying around a significant amount of storage, in many cases, a 10 GB iPod is already enough to carry their entire corpus of personal files and settings, along with a decent sized collection of music.
One thing to consider, is that this effectively gives people dramatically more effective bandwidth out of their homes. How might we, as technologists, make use of that fact to do cool and useful things?
Slashdot Mirror
If everyone listened to his advice, they'd at least have a good picture of the real requirements of the project they were undertaking, which would leave them in a much better position to learn while doing than just plunging in without a clear definition of problem.
And besides that, a smart person could learn alot from doing the requirements gathering and then working closely with an SI to define and implement a solution.
signal processing circuitry and instructions aimed at software radios, perhaps?
You do realize that every food crop/animal that man has raised since the dawn of time has been slectively bred to produce higher yields, disease resistance, and/or other physical traits (Dog and Cat breeds look nothing like their ancestors, neither does your baked potato). GE/GM crops are just allowing us to add factors that would normally take millenia to add.
And you do realize that a millenia gives a lot of time to identify and weed out undesirable side-effects?
With a decent blogging tool I can post a link to a page with a text excerpt and some brief commentary with minimal effort:
1. Select text
2. Click blogging bookmarklet.
3. Add any comments
4. Click the post button
I can't do that with notepad. Of course, I can do it with free software.
Other things that take more work with notepad.
1. Cycling stuff off my front page.
2. Creating archive pages.
3. Creating navigation by topic
4. Keeping a consistent template for all my blog pages
5. Syndicating blog content in RSS and/or other formats.
I tend to agree that Slashdot isn't the best place for this type of question, when there have to be thousands of real-world and online support groups out there. On the other hand, the person asking the question obviously has an affinity with the Slashdot community, and so experiences of Slashdotters may have more resonance for them than say that of a bunch of random AOL users.
Ummmmmmmmm, you are talking to me? Or a strawman?
It is really refreshing to see someone so willing to demonstrate their wrongheaded ignorance. Saves us all a lot of trouble.
I've found most of the people on Hydrogenaudio to be incredibly pragmatic. Perfection isn't the only parameter of importance. If it were, they'd not be wasting time testing codecs at 128kbps, except to demonstrate their unsuitability compaired to losless formats. They'd not be wasting time letting phillistines with their waxy untrimmed ears particpate in listening tests with their $20 sony earbuds.
As for the vendors lauding useless gear, um, what vendors lauding useless gear?
But hell, why let any of that get in the place of a perfectly good piece of ranting rhetoric. Still, it would be better if you'd unloaded at a deserving target. There are certainly enough of them out there.
While true that at 128kbps the differences will be more obvious to untrained ears, I wouldn't want people to think that the results of this this test will says anything about the relative performance of a given codec at a higher or lower bitrate.
A given codecs performance relative to other codecs at a given bitrate doesn't necessarily predict the performance of the same codec at another bitrate. For example, the codec implementation may be optimized for 128 kbps (something people suspect of the Apple AAC codec, for example).
The guy arranging the listening test at Hydrogen audio is well aware of this fact, so I think its rather likely that 128kbps wasn't chosen simply because its easier for untrained ears to hear the difference.
I think it was chosen because it is "disputed territory" where audio quality is acceptable to most users under causual listening conditions at a reasonably low bitrate.
I'd read the thread when they were discussing which version of Apple's ACC codec to use for the test, and concluded based on a few samples that the new version was subpar.
If they'd included both versions of iTunes/QuickTime in this test, perhaps they could have helped shame Apple into fixing what they broke.
Because when you are dealing with portable digital audio, storage still costs.
And how do you know what you are asserting? Have you done properly controlled listening tests with 128kbps encoding using a variety of codecs?
The fact is that for a lot of people, knowing the best codec at 128kbps is worth knowing because:
1) They are using portable devices where they are space constrained
2) They are using portable devices that may not have the perfect fidelity of a high-end sound system, but can go anywhere with them.
3) They are using their portable device in a somewhat noisy environment that overshadows any sound quality issues caused by a lower bitrate.
And, the bitrate isn't a moot point. Download speed is only one consideration. Bitrate also directly impacts filesize so if you are storage constrained, like say a solid state digital media player, lower bitrate performance is important.
Different codecs and implementations of those codecs may be optimized for different bitrates, so its important to test codecs at various target bitrates.
Dude,
1. Hybrids generate their own electricity from gas.
2. The post you were replying to is so clearly sarcastic.
SGIs early version of the Reality Engine used multiple units of the intel i860, a RISC chip and one of intels many attempts to move beyond the x86, to do the serious crunching. If one reads the paper in IEEE Micro when the i860 was released, one can see that it was positioned as a general purpose CPU for engineering workstations. I don't think it was used in said capacity in anything other than a reference design.
At this point there isn't much of a market for massive parallel floating point (or integer) performance for anything but 3D graphics, much to the chagrin of CPU makers, so its not clear to me what incentive GPU makers have to produce a more general purpose and more commodified product.
I'm trying to imagine how the market might evolve from here.
Clearly there are some computationally intensive tasks that could benefit from a more general purpose math unit, but not as general purpose as todays CPUs.
Some of those tasks are of interest to a relatively small number of people, but those people will buy as much as they can afford. If they could add four or six big vector units per PC in a commodity cluster, they'd do it. So, while it may be a relatively small market in terms of # of customers, they are likely to buy a disproportionate number of units.
Some of the those tasks are of interest to a much larger number of people, but those people may be served, and satisfied, with just being able to tap the lone GPU on their lone vid card for things like video or audio encoding.
Taken together, the combination of the two markets may be big enough to be interesting to one or more GPU vendors, and both markets would benefit from a standardized general programming model.
What is less clear is whether both markets really demand an architecture that supports multiple GPUs per system for non real-time 3d tasks. Of course, the successors to PCI and AGP may make this question moot, PCs might have suitable infrastructure for supporting multiple cards without any special effort on the part of GPU manufacturers.
It seems likely that we'll see broader support for using GPUs for non-traditional tasks. If you think about it, the GPU makers stole a big piece of "instruction share" (% of total instructions executed in a PC) from Intel and other CPU makers, and along with it, they captured revenue growth that would have otherwise gone to CPU makers. Agreeing to some level of standardization in programming models (a la DirectX & OpenGL) and supporting the use of GPUs for non-traditional tasks will allow the GPU makers a chance at grabbing even more revenue growth away from the CPU makers.
Well, for one thing, Sports Illustrated's photo department doesn't have to impress anyone with their gear to keep getting jobs.
Thos studios in NYC have to look the part.
A via isn't going to be able to do some things, like encode MPEG4 real time, at least not at their current clockspeeds, and it isn't going to do much for modern games. I'm also not sure it will handle playback of a HiDef MPEG4 source, or if its up to running something like descaler.
That said, its probably fine if you are just doing playback, or using a hardware MPEG2 encoder.
I'm actually debating whether to use a VIA for my HTPC. My plan is to use tuner cards with hardware MPEG2 encoders, so a Via should be fine, but i'm limited if I want to do anything different.
A big part of the motivation for Star Trek was that Moto was late late late with the 030 (due at least in part to a patent licensing issue) and there was a chance the CPU wouldn't ship at all, leaving apple a generation behind.
Sound familiar?
The story that Apple realizes profit from the iTMS not by selling songs, but by driving iPod sales is a good one, and no doubt there is a lot of truth to it, but I think it is foolish to take it at face value.
Right now, its in Apple's interest to make sure iTMS appears like an unattractive business to get into, because it discourages potential competitors, and the investors who might fund competitors. Meanwhile, the iPod story keeps Apple investors happy.
In time though, as volumes increace, as their initial investment is recouped, as they improve efficiency and lower costs, and as they negotiate better terms with record labels, their story will likely change.
I would wager that if LCD and DLP projection TVs really take off in volume, and there is a real attempt to standardize bulbs, that prices on bulbs will come down.
Right now, its a fairly small market, even taking into consideration all the conference room projectors. However, if a significant number of middle-class homes worldwide end up with these things, we are talking whole new economies of scale.
She was your aunt, and its your money, but if I wanted to honor somone for introducing me to technology, I don't think I'd put the money into hooking up a bunch of guys who already have ready access to tech (gamers).
I'd probably play "stone soup" with it and use it to attract other funds and direct them to a more community minded project that might actually introduce people to technology.
interesting observations, but, just because VirtualPC doesn't do those things now doesn't mean a version of the technology for the xbox wouldn't.
1. Why are you tied to DSL. I'm aware of some of its theoretical advantages, and in my market, its the only way to get more than 128 kbps up, but technology evolves, as does the way companies tune and deploy it, so it may be worth looking at a bundle of Cable IP & TV again.
2. Cable systems and satelites are both bandwidth constrianed, and HDTV takes a lot of bandwidth. Even before HDTV, satelite and digital cable providers were overcompressing stuff just so they could fit in a few more pay per view channels. One result is that their HDTV quality is often below that you could get with a decent broadcast antenna. Not only that, but it might get better or worse depeding on whatever else they are trying to cram down the pipe.
On the other hand, there is a tendancy for technologists to look at new applications of technology and say: Thats no big deal, its just a perl script, or, thats no big deal, its just a portable hard disk. Bittorrent is nothing, its just a python script. This jaded attitude can blind them to the implications of technology change.
Perhaps there is nothing new about this, but I'd suggest that it is an extreme illustration of a change that may have real implications.
With the rise of the iPod, and similar devices, people are now in the habit of carrying around a significant amount of storage, in many cases, a 10 GB iPod is already enough to carry their entire corpus of personal files and settings, along with a decent sized collection of music.
One thing to consider, is that this effectively gives people dramatically more effective bandwidth out of their homes. How might we, as technologists, make use of that fact to do cool and useful things?