Home Made HD Cinema Cameras - Technical Discussion

[Wayne Morellini]

Here is a link to a program reporting on a new artificial photsynthesis by using dye. Works in "lowlight" etc.

Now this may appear bot to have anything to do with cameras, but all sensors collect light and converts to electrical charge. So anything that ups the conversion rate of light to power, improves the sensors efficiency.

As I have discussed with Ronald before, there are a number of new technologies reporting potentials of over 60% efficency. It might be interesting to note that silicon solar pannels have around 20% efficency (it starts under, and I don't knoiw how high it now reaches, but probably no more than around 27%). Now if you go to a cmos sensor, how much are we getting between 10 to 20% efficency??. So a jump to 30-60% (realistically somthing might get above 70-90% one day) might do wonders.

http://www.abc.net.au/catalyst/stories/s1241478.htm

It sounds familiar, but I believe there are some technologies that could get closer to 100% (not actually 100% because of inefficencies).

Here is an interesting page with lots of links on microsdisplay, wearable computers, programmable silicon and processors etc:

http://216.239.57.104/search?q=cache...&hl=en&start=5

Fun Fun Fun for those itnerested in embedded stuff and electronics. Have alook at these:

http://www.devrs.com/e/
http://www.devrs.com/e/docs.php#docsmags
http://www.devrs.com/pic/
http://www.freertos.org/PC/

Wayne.

[Régine Weinberg]

Dear Wayne
I"ve been for years in the wearable group of
Steve Man lot of fun ...

[Wayne Morellini]

Would you believe it, attachable lens for Mobile phones (even a zoom ;):

http://www.amacrox.com/04english/p-camera.htm


And free Windows XP like Linux:

http://www.freedows.com/index.php?Lang=en


Low noise ATX poxer supply (22db):

http://www.amacrox.com/04english/p-fanless.htm

[Wayne Morellini]

Actually, looking at the back of that power supply it looks very simular to mine, probably a latter developement and a bit quieter (it also states 22db, full load, so it must adjust itself too).

A word of warning with these low niose power supply, I regard anything above 17db as too niosy for close proximity work.

Also note that there was some extremely small silent embedded power supplies for ITX boards I mentioned previously. I don't remember but I think there was ATX versions. Good if you don't require big power draws.

[Régine Weinberg]

Dear Wayne
I do have the catalog of a Frenc dealer www.edox.com
they do have a pc atx 350 W without any fan ...0 db

[Wayne Morellini]

Having trouble finding that, do you have link to the product?

Thanks

Wayne.

[Valeriu Campan]

Did anybody have a look at embeded SBC?
A sample here:
http://www.lannerinc.com/products/ac_em_sbc.asp

[Wayne Morellini]

Nothing extra special, some new micron models, no Altasens yet.

http://optics.sumix.com/products/cameras/index.htmll

Wayne.

[John Nagle]

I emailed Sumix and they say their new Altasens based camera will be out next March

[Wayne Morellini]

That is not very good, they were supposed to be out by December, so the Altasens manufacture delay should allowed them to have a design fully complete by the time the sensor became available. I can only summise that perhaps the altasens is further delayed :( . From a marketing point of view this does not bode very well for the projects here, are we going to have to wait for the Sony HDV cameras to go on discount before we get an Altasens. A few months delay in the release date for a silicon product because of manufacturing delays should not be counted as anything especially unusual, but this delay is much longer, and makes me wonder what's up.

Ronald, what do you think, you should have more upto date knowledge of general chip production problems?

[Régine Weinberg]

Well a guest you can read any posting in here, can you? With a real name you can too.

I do know that Intel for example is watching microcontroler and other forums on a day by day basis. So I do guess, that Altasens will deliver the ProCam market and does not talk about as this is common with CPU. If I comes true that we can do for 5k or 10k a flexible camera design with 1080i recording to dead cheap Sata array as disks , think of laptops from the shelf, used on fishing trawlers,a in Sport avionic, in artik reserch and so on, on the foolish sail race Vendee globe to name a fe. Tape is dead, to much mechanics ,a tape drive is vulnerable. Writing on any DVD the laser head is vulnerable, only a disk is sealed but they dont like it, as Cinema people stick to tape, only to take it out and pop it into a tapedrive elsewhere.

It is all politics, habits, tics,
have fun

Dear wayne and all the others:
http://linuxdevices.com/news/NS4887107636.html

that is only on of the board but is has RAID SATA, GIGE Ethernet, LOW power Fanless usb 2, Firewire, Graphics typical VIA stuff, they do have a dual fanless MOBO too and Linux ready and preloded

NO Noise CPU cooler, eg heatpipe

No noise CPU cooler
http://www.silverstonetek.com/products-nt01.htm

no noise power supply
http://www.silverstonetek.com/products-12a.htm

once found a powersupply no fan that
can be connectected to 24V DC

[Wayne Morellini]

Ronald, my applogies, I think you misunderstood my meaning. I was just asking general opinion on what long delays in chip manufacturing stage mean, nothing specific. In old days (20 years ago) delays were often enough, but 6 month plus delays were distraterouse, but that knowledge is out of date now.

Wayne.

[Régine Weinberg]

this is with GIGE Ethernet low, low power and fast the Dual CPU board just scroll down on the page a bit

http://linuxdevices.com/news/NS7787520286.html

[Wayne Morellini]

David, I would like to pass this by you and other compression experts). I have just had a sudden revelation, maybe we are going about the compression the wrong way around. The following is given on the assumption of masses of future computing power. Though I am unsure that present machines can do it. For a next gen device I would suggest the IBM/SONY CELL processor (with power CPU core for Mac people) with staggering power and for consumer electronics (16 Teraflops per rack). With Linux programming it should be relatively easy.

Compression:

- Can be built for different ussages

- for transmission the routine has to provide complete frames every now and then for drop out, and extra data to rebuild data incase of corruption.

- For tape you have simular overhead.

- But because the data has to be compressed one frame at a time this reduces the overall compression rate, as you can not compress to many frames at once.

- With file compession greater lossless compression rates are possible because the whole file is compressed at once.

We don't need to use standards for single frame transmission and tape recording.

As we are going to be using mainboards that can take 8 gigabytes of memory in future, that is a lot more like a file than 6 frames worth of data. So we can get around some comrpession bottle necks and achieve greater comrpession. Somewhere I posted a link to the leading file compression routine, from what I remember we could well and truely break the 2:1 lossless compression ratio. I also would advocate the filtering of unwanted niose to increase compression (even thought this is not regarded as lossless).


So what do you think, within the range of possibility?

[David Newman]

It is basically what CineForm does already, although we apply it to lossy compression (yes it's extendable to lossless -- one day.) It is why CineForm achieves 10bit compression at the same bit-rate as Avid's DNxHD 8bit mode. But the techniques of exploiting temporal redundancy has been known forever. The trick is to develop this style of compression suitable for post-production. You may have already seen our quality analysis at http://www.cineform.com/technology/quality.htm, showing how two codecs around 100Mb/s can differ so greatly in quality. Wavelets aren't our only edge.

David