A 23-min long Youtube video shows the recorded presentation of Transverse Field Detector (TFD) - a pixel that is said to allow for imaging without color filters, proposed by Politecnico di Milano, Italy:
I always like to see people doing new stuff and having fun. I think the "tunable" color response has been know for a very long time, and I know for sure was patented albeit probably not for the first time, by Cyrus Bamji within the last 10 years ago or so. The principle seems to be almost the same. As Foveon and others have proven, good color is possible, but generally expensive for computation. I could not sit still for the whole 23 minutes but I did not see a lot of references to earlier material. Hopefully I just missed it in this talk.
Yes, the idea is not new and, to me, seems quite obvious. Politecnico di Milano has been presenting it for the last 3-4 years and has invested a significant effort to implement it and explore many practical aspects of such sensor. I'm not aware of any other group spending that kind of effort in this idea.
That is the problem in EU research actually. They spend a lot of money without know what could be served by their "technology". I agree that in fundamental science, this is necessary. But in applied science and in so generalized way, I'm scared.
Einstein developed the special theory of relativity purely out of his passion and thirst for understanding the nature of time. There was no engineering project that sponsored him, and he did not have any commercial interests. We are finding application for his "abstract" theories now in applications such as GPS. So I doubt whether you can really segregate "fundamental" and "applied" sciences in a really B&W way in which you describe. In my humble opinion, such kind of a thinking pattern is the major stumbling block that engineers/scientists are facing today. It is a pity that some are even enthusiastic to douse such burning passions and enjoy the dark smoke that thus arises, in order to actively keep development and progress in check.
"I am awaiting the day when people remember the fact that discovery does not work by deciding what you want and then discovering it..." David Mermin http://en.wikiquote.org/wiki/David_Mermin
I am in violent agreement with the person above who wrote about Einstein. There a lot more relevant examples, the one that comes to mind immediately is NMR imaging.
At the same time applied research is what it is, and when it is funded it should account for what came before it. It would be difficult to justify funding "pure curiosity" research with tax dollars today. This is primarily due to the bloated size of the research establishment, as evidenced by the proliferation of journals and the uncontrolled expansion of the size of existing ones. A lot of the great discoveries that were made for pure curiosity's sake which we like to wax poetic about were done by what would today be described as the "idle wealthy". The best we can hope for are lucky accidents --- that someone smart working on widget X to please funding agency Y stumbles on something new and recognizes it for what it is.
No one is here so stupid here to support a B/W vision between the fundamental research and applied science. It's just a problem of proportion. Any researchers need a large vision before engaged into some "abstract" project. Even for Einstein, he didn't do his research work without objective. Relativity, quantum concept, etc. they are all invented for resolving some un-explicable phenomena at that time.
Foveon and also some researchers before worked on double junction detectors because it can potentially resolve the aliasing problem related to discret and spatially sumsampled CFA approach. They did a great job even there is no so big commercial success, but from scientific point of view, it's a big success.
This research work didn't bring something new in fundamental physics. The proposed approach suffers from aliasing problem. May I ask you one question: What is the fundamental problem you would like to resolve? It's not clear in your presentation.
Anonymous above: No one is stupid here either to fund 100% fundamental research; everything has a natural distribution curve. This is what I meant as B&W thought process. We develop and invest so much effort on developing sensors for outer-space and for various other space missions: are those (missions and sensors) at all needed according to you? What benefit do these space missions have to an average "tax payer"? How does knowing whether a distant planet has certain mineralogical content even help the society here back on earth? How about the multi-billion dollar particle colliders? What are they trying to solve according to you? How does unifying the fundamental forces for a grand unified theory help an average industry? I can extrapolate to include subjects such as history, literature, etc etc. Maybe we better scrap those and invest on improving the QE of a PPD by 2.73%, noise by 1.23% and crosstalk by 4.75%. Will that satisfy you? (BTW, I am not talking of the above presentation: it one's sole personal decision to approach a path in a way he likes; it may or may not have an immediate impact on the industry. You never know, it may be useful if certain special circumstances, don't jump too fast to conclusions)
Funny to mention the space program. The modern CMOS APS device came out of space instrument R&D at JPL. Sometimes you just don't know where technological inquiry will lead you.
I'm in moderate agreement, however, we must note that Einstein still needed employment at the patent office. Technology development today needs to see the hot dog. One of Malcolm Gladwell's books talked about how Fleetwood Mac cut many albums before the went platinum. No label today would have held on so long- it's a return on investment world... However, counterpoint to what I've just stated, antibiotics came out pure research, and work got under way in earnest on LCD's in 1964 but we did not see the first product until 1972.
I always like to see people doing new stuff and having fun. I think the "tunable" color response has been know for a very long time, and I know for sure was patented albeit probably not for the first time, by Cyrus Bamji within the last 10 years ago or so. The principle seems to be almost the same. As Foveon and others have proven, good color is possible, but generally expensive for computation. I could not sit still for the whole 23 minutes but I did not see a lot of references to earlier material. Hopefully I just missed it in this talk.
ReplyDeleteYes, the idea is not new and, to me, seems quite obvious. Politecnico di Milano has been presenting it for the last 3-4 years and has invested a significant effort to implement it and explore many practical aspects of such sensor. I'm not aware of any other group spending that kind of effort in this idea.
ReplyDeleteThat is the problem in EU research actually. They spend a lot of money without know what could be served by their "technology". I agree that in fundamental science, this is necessary. But in applied science and in so generalized way, I'm scared.
ReplyDeleteEinstein developed the special theory of relativity purely out of his passion and thirst for understanding the nature of time. There was no engineering project that sponsored him, and he did not have any commercial interests. We are finding application for his "abstract" theories now in applications such as GPS. So I doubt whether you can really segregate "fundamental" and "applied" sciences in a really B&W way in which you describe. In my humble opinion, such kind of a thinking pattern is the major stumbling block that engineers/scientists are facing today. It is a pity that some are even enthusiastic to douse such burning passions and enjoy the dark smoke that thus arises, in order to actively keep development and progress in check.
ReplyDelete"I am awaiting the day when people remember the fact that discovery does not work by deciding what you want and then discovering it..." David Mermin
ReplyDeletehttp://en.wikiquote.org/wiki/David_Mermin
I am in violent agreement with the person above who wrote about Einstein. There a lot more relevant examples, the one that comes to mind immediately is NMR imaging.
At the same time applied research is what it is, and when it is funded it should account for what came before it. It would be difficult to justify funding "pure curiosity" research with tax dollars today. This is primarily due to the bloated size of the research establishment, as evidenced by the proliferation of journals and the uncontrolled expansion of the size of existing ones. A lot of the great discoveries that were made for pure curiosity's sake which we like to wax poetic about were done by what would today be described as the "idle wealthy". The best we can hope for are lucky accidents --- that someone smart working on widget X to please funding agency Y stumbles on something new and recognizes it for what it is.
No one is here so stupid here to support a B/W vision between the fundamental research and applied science. It's just a problem of proportion. Any researchers need a large vision before engaged into some "abstract" project. Even for Einstein, he didn't do his research work without objective. Relativity, quantum concept, etc. they are all invented for resolving some un-explicable phenomena at that time.
ReplyDeleteFoveon and also some researchers before worked on double junction detectors because it can potentially resolve the aliasing problem related to discret and spatially sumsampled CFA approach. They did a great job even there is no so big commercial success, but from scientific point of view, it's a big success.
This research work didn't bring something new in fundamental physics. The proposed approach suffers from aliasing problem. May I ask you one question:
What is the fundamental problem you would like to resolve? It's not clear in your presentation.
Anonymous above: No one is stupid here either to fund 100% fundamental research; everything has a natural distribution curve. This is what I meant as B&W thought process. We develop and invest so much effort on developing sensors for outer-space and for various other space missions: are those (missions and sensors) at all needed according to you? What benefit do these space missions have to an average "tax payer"? How does knowing whether a distant planet has certain mineralogical content even help the society here back on earth? How about the multi-billion dollar particle colliders? What are they trying to solve according to you? How does unifying the fundamental forces for a grand unified theory help an average industry? I can extrapolate to include subjects such as history, literature, etc etc. Maybe we better scrap those and invest on improving the QE of a PPD by 2.73%, noise by 1.23% and crosstalk by 4.75%. Will that satisfy you? (BTW, I am not talking of the above presentation: it one's sole personal decision to approach a path in a way he likes; it may or may not have an immediate impact on the industry. You never know, it may be useful if certain special circumstances, don't jump too fast to conclusions)
ReplyDeleteFunny to mention the space program. The modern CMOS APS device came out of space instrument R&D at JPL. Sometimes you just don't know where technological inquiry will lead you.
ReplyDeleteI'm in moderate agreement, however, we must note that Einstein still needed employment at the patent office. Technology development today needs to see the hot dog. One of Malcolm Gladwell's books talked about how Fleetwood Mac cut many albums before the went platinum. No label today would have held on so long- it's a return on investment world... However, counterpoint to what I've just stated, antibiotics came out pure research, and work got under way in earnest on LCD's in 1964 but we did not see the first product until 1972.
ReplyDelete