I think you're confusing duplexers with simple digital filters. DSP processors to handle roughly 50 kSa/s aren't really all that complicated and don't require any real tuning. All you'd need to do is load the right program. No need to worry about clock jitter either at those speeds considering how fast the clock your DSP processor is probably working (few megahertz) . Actually, the DSP part of bluetooth connectivity will be far more complicated than the actual audio processing. Programming said filter could easily be done by bluetooth once the correct firmware is in place, so actually you wouldn't even need any specialised equipment past product testing in the factory. The main cost really is due to the miniaturization.
You assume one has to be skilled in signal processing to configure a filter?
If you put some work into it you can make a program that even a 5 year old could use. And for these sorts of filters the stability criteria aren't all that complicated either. So you won't have much trouble on that front.
Do you realise how much it "costs" to configure a digital filter? The answer is, next to nothing.
Programs like MATLAB can calculate the necessary coefficients on the fly. Any skilled engineer can write a program to convert a given frequency and phase characteristics to a list of filter coefficients. Even if you want to do fancy things like lattice filters it's still not very complicated as most of these are designed by converting the direct form of the filter using a recursive algorithm. Again something a computer can do.
The actual cost I'd say comes from the size of the hardware combined with a good battery. If you wish to miniaturize it'll cost you big time. And quantities in the hundreds of thousands aren't quite enough to warrant the cost of miniaturization as these sort of things will take some custom hardware. Mask fabrication and set-up costs for a run of wafers; well it isn't cheap.
Contrary to popular believe many of the drop outs don't make it. They fail, end up without a single penny left in their pocket.
Additionally, comparing academic achievements in the US with how good the educational standard is considering it's largely based on immigrants. And these people are not a product of US education...
The current university ranking systems and judgement of quality is highly biased. To give an example: Many European universities are far ahead of US universities in educational level, yet they score way worse. The simple reason is that all these rankings assume the American system and introduce ratings on non relevant factors. I've seen some really sad examples, they actually managed to use factors like " openness ". I got rather fed up with it and contacted the authors of said ranking system. And the moment I asked that question the author just rambled on a bit and got off the phone quickly. Afterwards I figured out it means who they let in. Fact is that in most European universities a big fat wallet or knowing somebody won't get you in. It's either knowledge or being fast enough with the necessary high school degrees. Then they also mark said universities down due to the socialist system employed. Pretty much everybody has a change to start without having huge expenses. So on average the failure rate is a lot higher (I recall we lost over half the students in engineering over the summer break).
Another factor they will take into account is publications. Universities that focus on education will automatically score lower in this cause the professors and their assistants actually bother to teach the students and help them with their projects instead of focusing on their own research. Yet this does not mean their research is of lesser quality, they simply publish less and use less resources in doing so.
And the case with Chinese universities is simple, some indeed are of inferior quality. I'll give you that. But a lot of the good publications aren't in English. Or are of poor quality cause of their translation. That is absolutely no reason to call most of them crappy and their degrees useless. And if you're worried about the degree being recognized, each country has some committee that approves foreign degrees and says what the local equivalent degree is. If you really worry about how much a foreign degree is going to be worth, simply contact them.
I don't think you quite understand the point of my comment.
I generally find Weibo far more fun to use cause there is generally a lot less drama and whining going on there. Additionally they didn't bother to hide everything in a 20 layer menu structure like Facebook did. The result is rather easy to use. And frankly if I use a social network site I'm looking for a very specific set of services. I'm not interested in your revolutions, your whining, your political views,... . If you want to argue about that go and do it somewhere else but don't annoy me with it during my spare time.
And before anybody whines about handing over credentials. Fact is that if they really want to know your personal information, they'll get it no matter what. Most likely scenario is that they don't really give a damn about what you do. Governments work like that. So that argument is also very much flawed.
True story!
As it is right now with the huge pipelines the indicated clock frequency doesn't matter all that much. Especially as the clock is often divided into sub cycles. The real issue still is - as it always has been - talking to the memory. And while ever smarter compilers combined with better out of order execution does help. It's still a hassle that you can't directly talk to the memory and need to wait. It really drags down the efficiency of your pipeline if you made a wrong prediction. The actual value we should be looking at is the amount of instruction launched per second. If possible half floating point, half integer. That should lead to far more interesting performance evaluations than the just the clock frequencies. Might lead to funny results and a lot of sad gamers.
Actually, making your own motherboard with off the shelf components isn't all that complicated as you might think (at least to get a basic system that boots). Intel provides great documentation with its processors that pretty tell you exactly what to get and how to connect it to the processor. Most of the parts are available from distributors like Farnell and Digikey. Others you can actually get as samples straight from the manufacturer. The issues start once you want a complete system with support for modern peripherals. Another problem might be mounting all the BGA components they're so fond of today. But it's doable with a toaster oven if you're really into doing things at home.
Yeah, cause right now it doesn't sound all that impressive.
Fact is that any engineering student who took fluid dynamics has had to solve similar problems like "A ball > was shot out of a cannon > into a gas with next properties: >. Calculate...." At least I remember I had to do that, and I'm not even a mechanical engineering major... So unless he found some super exact way to do it I'm not very impressed. And even if he found a very exact way to do it I'm still not impressed, it just means he bothered to waste more time on it and decided to not simplify irrelevant factors.
Same with an object hitting a wall, you can go as far in calculating that on paper as you want. It's just questionable if there's a point to it cause the analytic solutions are often very generalised and useless for actual modelling of physical systems.
In fact exact analytic modelling of physical systems sort of became a joke with us EEs. Cause you always end up using statistical physics when dealing with semiconductors or very gross over simplifications; Like using Kirchoff and Ohm instead of Maxwell's equations. And at the point where you introduce statistics and probability into a problem -as is always the case with very complex realistic modelling- you can just as well use a computer program to verify the solution. The computer is less prone to errors (if configured correctly), faster and will be able to visually represent the results without having to spend hours of additional work. The resulting time is better spent sniffing solder fumes, playing minesweeper on logic analysers, working on your own personal projects in the lab, drinking coffee,...
Conclusion: Either the article must be mentioning something very wrong or this is another case of blowing up a highschool student's achievements to make a feel good story.
You might want to check out newhavendisplay their TFTs: http://www.newhavendisplay.com/
Their interface chips aren't that expensive either. Documentation could be a bit better but considering the price most of their products are a pretty good deal.
They actually added the same system to DVDs for some reason. Assuming that it'd still work.
The thing is, no matter what they'd try, its always a minor annoyance. It's easy to detect fake frames.
Sure it can be used as a first line of defence. But it should never be the only. And it should never have complete control. But hardware safety features should ALWAYS be implemented cause the software can (and will) fail at some point. Additionally it's important to design the hardware so that a failure of the safety system itself also disables the machine.
It's still very easy for software or hardware to detect such tricks. Remember what they tried with their analogue content protection system in the day for DVDs?
You can't exactly call that a huge success, and that's considerably more tricky to eliminate than what you're suggesting now.
First of all. Believe it or not, engineers make up a large part of the scientific community. And you're very mistaken in that though. MATLAB is often the best tool for the job. Go and try to prototype any form of digital signal processing system or algorithm in anything else than MATLAB and they'll point and laugh at you. If you do any of that in python or another language you'll find yourself reinventing the wheel simply cause all those functions already exist in MATLAB. And at least you can be certain those will work compared to your own writings.
Actually, it's not a small minority. FORTRAN is still the tool of choice in many fields. Especially where massive parallel computing comes into play (think weather forecasting). Another largely unknown language that is used in large systems world wide is Ada. Again it's not very well known. But it's one of the few languages reliable enough to deal with systems like weapon systems, flight traffic control systems, autopilots,... It's not cause you didn't run into it yourself that it's never used. FORTRAN still makes its daily appearance in most research facilities. And the older researchers will often teach the younger ones about the wonderful language it is.
High level languages are slow and inefficient. You don't need objects for scientific calculations, what you need is a good mathematical library that is capable of handling large vectors/matrices quickly with efficient memory usage. Neither Python nor Ruby are capable of those feats (their libraries are crappy at best compared to the established values). For FORTRAN and MATLAB that's a walk in the park actually. Side bonus is that MATLAB code can be translated into C meaning it can be compiled resulting in near native performance as well. Again something Python and Ruby can't hope to achieve properly cause of their dynamic workings.
MATLAB wins in the engineering community, and that's by a long run.
But everybody keeps forgetting the unsung hero called FORTRAN. Several decades later it's still alive and kicking and it's still used on a daily basis in the scientific community. We don't need another language for scientific computing. And the last thing we need is a language like Ruby. Ruby is a memory hog and inefficient at the best of times. And I sure as hell wouldn't want to be tasked for writing a good optimizing Ruby compiler. FORTRAN on the other hand has a large list of available compilers. And some of them produce simply amazing results. See Intel's FORTRAN compiler if you want to see what optimization can do if done properly.
The real mindf*ck for most computer programmers comes when they realise half their common constructs are completely useless.
Looping is generally a bad idea (unless if it's a generation loop), recursion shouldn't even be considered, etc... . But VHDL is one of the better languages out there actually. I've written large snippets of VHDL that worked instantly. Simply cause of the fact that it's very hard to make a mistake in VHDL. But the learning curve might be a bit steep without a good tool chain (Xilinx ISE + ModelSim should get you started though).
And I'm willing to bet that is the case here. But the fact is, it's common in the social "sciences" for such things to occur.
Never had to read some of their papers? The quantity of citations made my eyes bleed.
The thing is though, they'd have to put a citation tag after almost every single sentence if they did. Hence they start dropping some of them. Resulting in it being called plagiarism.
It's impossible to write anything in the social sciences field without some level of plagiarism. Since it's near impossible to make hard arguments you need to cite other works.
The real problem here is that some of these journalists and "film-know-it-alls" have their heads so far up their behinds that they refuse any new technology. They think the motion blurring effects to fix the problems of 24fps look artistic.
Slashdot Mirror
I think you're confusing duplexers with simple digital filters. DSP processors to handle roughly 50 kSa/s aren't really all that complicated and don't require any real tuning. All you'd need to do is load the right program. No need to worry about clock jitter either at those speeds considering how fast the clock your DSP processor is probably working (few megahertz) . Actually, the DSP part of bluetooth connectivity will be far more complicated than the actual audio processing. Programming said filter could easily be done by bluetooth once the correct firmware is in place, so actually you wouldn't even need any specialised equipment past product testing in the factory. The main cost really is due to the miniaturization.
You assume one has to be skilled in signal processing to configure a filter?
If you put some work into it you can make a program that even a 5 year old could use. And for these sorts of filters the stability criteria aren't all that complicated either. So you won't have much trouble on that front.
Do you realise how much it "costs" to configure a digital filter? The answer is, next to nothing.
Programs like MATLAB can calculate the necessary coefficients on the fly. Any skilled engineer can write a program to convert a given frequency and phase characteristics to a list of filter coefficients. Even if you want to do fancy things like lattice filters it's still not very complicated as most of these are designed by converting the direct form of the filter using a recursive algorithm. Again something a computer can do.
The actual cost I'd say comes from the size of the hardware combined with a good battery. If you wish to miniaturize it'll cost you big time. And quantities in the hundreds of thousands aren't quite enough to warrant the cost of miniaturization as these sort of things will take some custom hardware. Mask fabrication and set-up costs for a run of wafers; well it isn't cheap.
Contrary to popular believe many of the drop outs don't make it. They fail, end up without a single penny left in their pocket.
Additionally, comparing academic achievements in the US with how good the educational standard is considering it's largely based on immigrants. And these people are not a product of US education...
The current university ranking systems and judgement of quality is highly biased. To give an example: Many European universities are far ahead of US universities in educational level, yet they score way worse. The simple reason is that all these rankings assume the American system and introduce ratings on non relevant factors. I've seen some really sad examples, they actually managed to use factors like " openness ". I got rather fed up with it and contacted the authors of said ranking system. And the moment I asked that question the author just rambled on a bit and got off the phone quickly. Afterwards I figured out it means who they let in. Fact is that in most European universities a big fat wallet or knowing somebody won't get you in. It's either knowledge or being fast enough with the necessary high school degrees. Then they also mark said universities down due to the socialist system employed. Pretty much everybody has a change to start without having huge expenses. So on average the failure rate is a lot higher (I recall we lost over half the students in engineering over the summer break).
Another factor they will take into account is publications. Universities that focus on education will automatically score lower in this cause the professors and their assistants actually bother to teach the students and help them with their projects instead of focusing on their own research. Yet this does not mean their research is of lesser quality, they simply publish less and use less resources in doing so.
And the case with Chinese universities is simple, some indeed are of inferior quality. I'll give you that. But a lot of the good publications aren't in English. Or are of poor quality cause of their translation. That is absolutely no reason to call most of them crappy and their degrees useless. And if you're worried about the degree being recognized, each country has some committee that approves foreign degrees and says what the local equivalent degree is. If you really worry about how much a foreign degree is going to be worth, simply contact them.
I don't think you quite understand the point of my comment. ... . If you want to argue about that go and do it somewhere else but don't annoy me with it during my spare time.
I generally find Weibo far more fun to use cause there is generally a lot less drama and whining going on there. Additionally they didn't bother to hide everything in a 20 layer menu structure like Facebook did. The result is rather easy to use. And frankly if I use a social network site I'm looking for a very specific set of services. I'm not interested in your revolutions, your whining, your political views,
And before anybody whines about handing over credentials. Fact is that if they really want to know your personal information, they'll get it no matter what. Most likely scenario is that they don't really give a damn about what you do. Governments work like that. So that argument is also very much flawed.
Weibo is still a thousand times better than Facebook on any day. :)
Less 13 year olds shouting drama about how much their life sucks in random groups.
True story!
As it is right now with the huge pipelines the indicated clock frequency doesn't matter all that much. Especially as the clock is often divided into sub cycles. The real issue still is - as it always has been - talking to the memory. And while ever smarter compilers combined with better out of order execution does help. It's still a hassle that you can't directly talk to the memory and need to wait. It really drags down the efficiency of your pipeline if you made a wrong prediction. The actual value we should be looking at is the amount of instruction launched per second. If possible half floating point, half integer. That should lead to far more interesting performance evaluations than the just the clock frequencies. Might lead to funny results and a lot of sad gamers.
Actually, making your own motherboard with off the shelf components isn't all that complicated as you might think (at least to get a basic system that boots). Intel provides great documentation with its processors that pretty tell you exactly what to get and how to connect it to the processor. Most of the parts are available from distributors like Farnell and Digikey. Others you can actually get as samples straight from the manufacturer. The issues start once you want a complete system with support for modern peripherals. Another problem might be mounting all the BGA components they're so fond of today. But it's doable with a toaster oven if you're really into doing things at home.
Yeah, cause right now it doesn't sound all that impressive. ... ." At least I remember I had to do that, and I'm not even a mechanical engineering major... So unless he found some super exact way to do it I'm not very impressed. And even if he found a very exact way to do it I'm still not impressed, it just means he bothered to waste more time on it and decided to not simplify irrelevant factors. ...
Fact is that any engineering student who took fluid dynamics has had to solve similar problems like "A ball > was shot out of a cannon > into a gas with next properties: >. Calculate
Same with an object hitting a wall, you can go as far in calculating that on paper as you want. It's just questionable if there's a point to it cause the analytic solutions are often very generalised and useless for actual modelling of physical systems.
In fact exact analytic modelling of physical systems sort of became a joke with us EEs. Cause you always end up using statistical physics when dealing with semiconductors or very gross over simplifications; Like using Kirchoff and Ohm instead of Maxwell's equations. And at the point where you introduce statistics and probability into a problem -as is always the case with very complex realistic modelling- you can just as well use a computer program to verify the solution. The computer is less prone to errors (if configured correctly), faster and will be able to visually represent the results without having to spend hours of additional work. The resulting time is better spent sniffing solder fumes, playing minesweeper on logic analysers, working on your own personal projects in the lab, drinking coffee,
Conclusion: Either the article must be mentioning something very wrong or this is another case of blowing up a highschool student's achievements to make a feel good story.
I'd say bullies are the main problem.
You might want to check out newhavendisplay their TFTs: http://www.newhavendisplay.com/ Their interface chips aren't that expensive either. Documentation could be a bit better but considering the price most of their products are a pretty good deal.
They actually added the same system to DVDs for some reason. Assuming that it'd still work.
The thing is, no matter what they'd try, its always a minor annoyance. It's easy to detect fake frames.
Sure it can be used as a first line of defence. But it should never be the only. And it should never have complete control. But hardware safety features should ALWAYS be implemented cause the software can (and will) fail at some point. Additionally it's important to design the hardware so that a failure of the safety system itself also disables the machine.
It's still very easy for software or hardware to detect such tricks. Remember what they tried with their analogue content protection system in the day for DVDs? You can't exactly call that a huge success, and that's considerably more tricky to eliminate than what you're suggesting now.
Hard-coded isn't enough. Critical safety operations should NEVER be given to software. Remember what happened with the Therac-25 machines?
How does this relate to the subject? It's all about using the best tool for the job. And FORTRAN and MATLAB are the best tools.
First of all. Believe it or not, engineers make up a large part of the scientific community. And you're very mistaken in that though. MATLAB is often the best tool for the job. Go and try to prototype any form of digital signal processing system or algorithm in anything else than MATLAB and they'll point and laugh at you. If you do any of that in python or another language you'll find yourself reinventing the wheel simply cause all those functions already exist in MATLAB. And at least you can be certain those will work compared to your own writings. ... It's not cause you didn't run into it yourself that it's never used. FORTRAN still makes its daily appearance in most research facilities. And the older researchers will often teach the younger ones about the wonderful language it is.
Actually, it's not a small minority. FORTRAN is still the tool of choice in many fields. Especially where massive parallel computing comes into play (think weather forecasting). Another largely unknown language that is used in large systems world wide is Ada. Again it's not very well known. But it's one of the few languages reliable enough to deal with systems like weapon systems, flight traffic control systems, autopilots,
High level languages are slow and inefficient. You don't need objects for scientific calculations, what you need is a good mathematical library that is capable of handling large vectors/matrices quickly with efficient memory usage. Neither Python nor Ruby are capable of those feats (their libraries are crappy at best compared to the established values). For FORTRAN and MATLAB that's a walk in the park actually. Side bonus is that MATLAB code can be translated into C meaning it can be compiled resulting in near native performance as well. Again something Python and Ruby can't hope to achieve properly cause of their dynamic workings.
MATLAB wins in the engineering community, and that's by a long run.
But everybody keeps forgetting the unsung hero called FORTRAN. Several decades later it's still alive and kicking and it's still used on a daily basis in the scientific community. We don't need another language for scientific computing. And the last thing we need is a language like Ruby. Ruby is a memory hog and inefficient at the best of times. And I sure as hell wouldn't want to be tasked for writing a good optimizing Ruby compiler. FORTRAN on the other hand has a large list of available compilers. And some of them produce simply amazing results. See Intel's FORTRAN compiler if you want to see what optimization can do if done properly.
The real mindf*ck for most computer programmers comes when they realise half their common constructs are completely useless.
Looping is generally a bad idea (unless if it's a generation loop), recursion shouldn't even be considered, etc... . But VHDL is one of the better languages out there actually. I've written large snippets of VHDL that worked instantly. Simply cause of the fact that it's very hard to make a mistake in VHDL. But the learning curve might be a bit steep without a good tool chain (Xilinx ISE + ModelSim should get you started though).
Fortran, anyone? Or is this the wrong crowd?
And I'm willing to bet that is the case here. But the fact is, it's common in the social "sciences" for such things to occur.
Never had to read some of their papers? The quantity of citations made my eyes bleed.
The thing is though, they'd have to put a citation tag after almost every single sentence if they did. Hence they start dropping some of them. Resulting in it being called plagiarism.
It's impossible to write anything in the social sciences field without some level of plagiarism. Since it's near impossible to make hard arguments you need to cite other works.
The real problem here is that some of these journalists and "film-know-it-alls" have their heads so far up their behinds that they refuse any new technology. They think the motion blurring effects to fix the problems of 24fps look artistic.