4:4:4 12-bit Uncompressed DVX100

Juan P. Pertierra · March 24th, 2004, 12:31 PM

Assuming those Rockwell chips can be obtained for a decent price, a simple camera can be made from them very easily.

The cool thing about the rockwell chips, is that they have built-in timing/AD conversion, so the chips actually output digital signals. So right out of the chips I am getting the signal that I am getting out of the DVX's A/D converters.

You can make a camera by obtaining some casing that can house optics, and mounting the chips accordingly(only hard part is obtaining the 3-way splitter). After that, we can handle the output from the chips just like the output from my mod.

Of course, the bandwidth will be much larger since its HD, but we can always go dual FW800 or maybe SDI. I've tried to stay away from SDI because the recording equipment is expensive, and besides i haven't found the specs anywhere.

I emailed rockwell but haven't gotten a reply regarding low-volume sales...not sure if they will be carried by any online retailer...CCD's/CMOS sensors of this kind are generally hard to come by in this way.

Juan

Adam Burtle · March 24th, 2004, 01:08 PM

would it be feasible to just mount them in an XL1S.. rather than making a new casing.

i can stop by their booth next month and talk to them if you want.. they might be more receptive in person.

Juan P. Pertierra · March 24th, 2004, 01:13 PM

Assuming you gut all the electronics(which you won't need), the main constraint is the fact that the 3-way beam splitter in the XL1s(or any 3 chipper) is probably made to the chip size/aspect ratio that is being used. Therefor, since the rockwell chips are larger, chances are they will not quite work well with the beam splitter.

However, if you find another camera casing, like say from an old ENG system that uses similarly sized(16:9?) chips, it should be incredibly easy to mount the rockwell chips and even put hard disks and the rest of the electronics in the camera, since ENG shoulder bricks tend to be pretty large.

Juan

Michael Pappas · March 24th, 2004, 03:08 PM

This is a very interesting topic! Here is a system that is HD. Take a look!

http://www.siliconimaging.com/SI1280%20spec.htm

http://www.siliconimaging.com/

Michael Pappas
PappasArts Entertainment
http://www.Pbase.com/arrfilms
http://www.pbase.com/pappasarts15

Adam Burtle · March 24th, 2004, 03:17 PM

so then to "build your own" hd camera using those chips (assuming they were financially viable) and the dual FW800 setup (or similar).. one would still need a casing, beam splitter, a way to mount a lens to the casing the appropriate distance from the film plane, power supply for the chips or any other onboard electronics.. and what else?

Juan P. Pertierra · March 24th, 2004, 03:56 PM

That's pretty much it...it depends on how you intend to capture the data. If you want on-board hard disks, then you can use some sort of programmable logic chip to interface the data from the chips to, say, a serial ata controller. likewise, for a firewire 800 interface, you can have the programmable chip interface the data to firewire link layer->physical layer. The programmable chip can also selectively interface to both, making the system more flexible.

There's not that much to it. Of course, there is some PCB work to be done, in order to design and solder the board that has the programmable chip and interface chips, but the circuitry is extremely simple and there is little if any 'glue logic', as long as you use the rockwell chips.

if you use standard CCD chips, for example, then you need independent sample-and-hold chips for each CCD, together with timing chips and A/D converters...this complicates things a bit.

Juan

Mark Grgurev · March 24th, 2004, 05:08 PM

When are you gonna post the uncompressed pics?

Juan P. Pertierra · March 24th, 2004, 05:39 PM

Right now I am literally working on the bad connection issue that is causing noise in the capture...once i get a good connection i will post the 10-bit uncompressed color captures.

Obin Olson · March 25th, 2004, 08:39 AM

juan, what is the thing called that splits the light into RGB colors? it's not a beam splitter is it? or do you need a beam splitter AND color filters to do it? this is with 3ccd or cmos chips that are Black and White

Thomas Smet · March 25th, 2004, 10:06 AM

I have read about the chroma phasing problem with the dvx100 during pans which can make keying difficult with the camera. Do you notice this problem still using your method of pulling the image from the chips or does that seem to clean it up?

Juan P. Pertierra · March 25th, 2004, 10:19 AM

Obin:

Afaik, there is a three-way beam splitter which basically splits the light coming from the lens into three(i.e. the exact same image x3 with 1/3 the original intensity) and in front of each monochrome CCD, then there is a color filter for R,G or B.

Thomas:
I am not aware of the details, but it is very possible that this solves the problem. I think the chroma phasing could be a side-effect of how the electronics deals with the signals, perhaps when they are transformed to Y-C.

Unless this chroma phasing is a physical defect in the design, or in the timing characteristics of the chips(doubt it) it probably won't be present in the raw output.

Juan

Obin Olson · March 25th, 2004, 03:41 PM

ok in that case edmund optics has what you need a RGB filter set ...cost, about thirty bucks and I think they have a beam splitter too......http://www.edmundoptics.com/

Juan P. Pertierra · March 25th, 2004, 03:49 PM

Thanks! That's fantastic, they do have the beam splitters....now the only hurdle is obtaining the rockwell chips at a decent price, no reply from them yet.

An update on the DVX:
Last night i fixed the false connection problem, i was still getting bad connections but i found out that the culprit was a surface mount soldering i did that was getting loose. Tonight i will resolder it.

I managed to capture one out of focus full frame before the connector completely fell off, but the software was set to single channel so it captured a red 10-bit frame...looks noise free...

Juan P. Pertierra · March 25th, 2004, 07:55 PM

I got a response from Rockwell...it seems what was said on this thread before is true! For a 1-10 unit order, the ProCamHD 2560 is a bit over $500, while the ProCamHD 3560 is about $1500.

The ProCamHD 2560 would be the PERFECT choice for making a 3CCD HD box camera...it is 1280x1024, and can do up to 60fps progressive so slow motion is definitely in :)

Juan

Obin Olson · March 25th, 2004, 11:02 PM

dude that is soo great! I sure hope that this can be done and we can bypass the $60,000 + lenses why not bypass tape too? and use some sorta firewire800 capture ??

don't get ahead of yourself yet though! we all still want dvx100 4:4:4 !!!!!!

what about computer processing ?? don't you need lots of stuff to get the data from 3 cmos chips into a color image format? that can be saved?

I could not find a 3 way beam splitter...only 2 way...I sent an email to find out if they have what you need ;)

March 25th, 2004, 07:55 PM	#194
Juan P. Pertierra Major Player Join Date: Jul 2003 Posts: 479	Rockwell chips I got a response from Rockwell...it seems what was said on this thread before is true! For a 1-10 unit order, the ProCamHD 2560 is a bit over $500, while the ProCamHD 3560 is about $1500. The ProCamHD 2560 would be the PERFECT choice for making a 3CCD HD box camera...it is 1280x1024, and can do up to 60fps progressive so slow motion is definitely in :) Juan

March 24th, 2004, 12:31 PM	#181
Juan P. Pertierra Major Player Join Date: Jul 2003 Posts: 479	Assuming those Rockwell chips can be obtained for a decent price, a simple camera can be made from them very easily. The cool thing about the rockwell chips, is that they have built-in timing/AD conversion, so the chips actually output digital signals. So right out of the chips I am getting the signal that I am getting out of the DVX's A/D converters. You can make a camera by obtaining some casing that can house optics, and mounting the chips accordingly(only hard part is obtaining the 3-way splitter). After that, we can handle the output from the chips just like the output from my mod. Of course, the bandwidth will be much larger since its HD, but we can always go dual FW800 or maybe SDI. I've tried to stay away from SDI because the recording equipment is expensive, and besides i haven't found the specs anywhere. I emailed rockwell but haven't gotten a reply regarding low-volume sales...not sure if they will be carried by any online retailer...CCD's/CMOS sensors of this kind are generally hard to come by in this way. Juan

March 24th, 2004, 01:08 PM	#182
Adam Burtle Regular Crew Join Date: Oct 2003 Location: Los Angeles, CA Posts: 173	would it be feasible to just mount them in an XL1S.. rather than making a new casing. i can stop by their booth next month and talk to them if you want.. they might be more receptive in person.

March 24th, 2004, 01:13 PM	#183
Juan P. Pertierra Major Player Join Date: Jul 2003 Posts: 479	Assuming you gut all the electronics(which you won't need), the main constraint is the fact that the 3-way beam splitter in the XL1s(or any 3 chipper) is probably made to the chip size/aspect ratio that is being used. Therefor, since the rockwell chips are larger, chances are they will not quite work well with the beam splitter. However, if you find another camera casing, like say from an old ENG system that uses similarly sized(16:9?) chips, it should be incredibly easy to mount the rockwell chips and even put hard disks and the rest of the electronics in the camera, since ENG shoulder bricks tend to be pretty large. Juan

March 24th, 2004, 03:08 PM	#184
Michael Pappas Regular Crew Join Date: Sep 2001 Location: California Posts: 667	This is a very interesting topic! Here is a system that is HD. Take a look! http://www.siliconimaging.com/SI1280%20spec.htm http://www.siliconimaging.com/ Michael Pappas PappasArts Entertainment http://www.Pbase.com/arrfilms http://www.pbase.com/pappasarts15

March 24th, 2004, 03:17 PM	#185
Adam Burtle Regular Crew Join Date: Oct 2003 Location: Los Angeles, CA Posts: 173	so then to "build your own" hd camera using those chips (assuming they were financially viable) and the dual FW800 setup (or similar).. one would still need a casing, beam splitter, a way to mount a lens to the casing the appropriate distance from the film plane, power supply for the chips or any other onboard electronics.. and what else?

March 24th, 2004, 03:56 PM	#186
Juan P. Pertierra Major Player Join Date: Jul 2003 Posts: 479	That's pretty much it...it depends on how you intend to capture the data. If you want on-board hard disks, then you can use some sort of programmable logic chip to interface the data from the chips to, say, a serial ata controller. likewise, for a firewire 800 interface, you can have the programmable chip interface the data to firewire link layer->physical layer. The programmable chip can also selectively interface to both, making the system more flexible. There's not that much to it. Of course, there is some PCB work to be done, in order to design and solder the board that has the programmable chip and interface chips, but the circuitry is extremely simple and there is little if any 'glue logic', as long as you use the rockwell chips. if you use standard CCD chips, for example, then you need independent sample-and-hold chips for each CCD, together with timing chips and A/D converters...this complicates things a bit. Juan

March 24th, 2004, 05:08 PM	#187
Mark Grgurev Regular Crew Join Date: Dec 2003 Location: Fairview,nj Posts: 137	When are you gonna post the uncompressed pics?

March 24th, 2004, 05:39 PM	#188
Juan P. Pertierra Major Player Join Date: Jul 2003 Posts: 479	Right now I am literally working on the bad connection issue that is causing noise in the capture...once i get a good connection i will post the 10-bit uncompressed color captures.

March 25th, 2004, 08:39 AM	#189
Obin Olson Trustee Join Date: Jan 2003 Location: Wilmington NC Posts: 1,414	juan, what is the thing called that splits the light into RGB colors? it's not a beam splitter is it? or do you need a beam splitter AND color filters to do it? this is with 3ccd or cmos chips that are Black and White

March 25th, 2004, 10:06 AM	#190
Thomas Smet Trustee Join Date: Mar 2004 Location: Milwaukee, WI Posts: 1,719	I have read about the chroma phasing problem with the dvx100 during pans which can make keying difficult with the camera. Do you notice this problem still using your method of pulling the image from the chips or does that seem to clean it up?

March 25th, 2004, 10:19 AM	#191
Juan P. Pertierra Major Player Join Date: Jul 2003 Posts: 479	Obin: Afaik, there is a three-way beam splitter which basically splits the light coming from the lens into three(i.e. the exact same image x3 with 1/3 the original intensity) and in front of each monochrome CCD, then there is a color filter for R,G or B. Thomas: I am not aware of the details, but it is very possible that this solves the problem. I think the chroma phasing could be a side-effect of how the electronics deals with the signals, perhaps when they are transformed to Y-C. Unless this chroma phasing is a physical defect in the design, or in the timing characteristics of the chips(doubt it) it probably won't be present in the raw output. Juan

March 25th, 2004, 03:41 PM	#192
Obin Olson Trustee Join Date: Jan 2003 Location: Wilmington NC Posts: 1,414	ok in that case edmund optics has what you need a RGB filter set ...cost, about thirty bucks and I think they have a beam splitter too......http://www.edmundoptics.com/

March 25th, 2004, 03:49 PM	#193
Juan P. Pertierra Major Player Join Date: Jul 2003 Posts: 479	Thanks! That's fantastic, they do have the beam splitters....now the only hurdle is obtaining the rockwell chips at a decent price, no reply from them yet. An update on the DVX: Last night i fixed the false connection problem, i was still getting bad connections but i found out that the culprit was a surface mount soldering i did that was getting loose. Tonight i will resolder it. I managed to capture one out of focus full frame before the connector completely fell off, but the software was set to single channel so it captured a red 10-bit frame...looks noise free...

March 25th, 2004, 11:02 PM	#195
Obin Olson Trustee Join Date: Jan 2003 Location: Wilmington NC Posts: 1,414	dude that is soo great! I sure hope that this can be done and we can bypass the $60,000 + lenses why not bypass tape too? and use some sorta firewire800 capture ?? don't get ahead of yourself yet though! we all still want dvx100 4:4:4 !!!!!! what about computer processing ?? don't you need lots of stuff to get the data from 3 cmos chips into a color image format? that can be saved? I could not find a 3 way beam splitter...only 2 way...I sent an email to find out if they have what you need ;)