|
|||||||||
|
Thread Tools | Search this Thread |
February 14th, 2005, 02:50 AM | #1 |
New Boot
Join Date: Jan 2004
Location: seattle, wa
Posts: 5
|
DVX manual lens conversion
Hey guys -
over the last three days I converted my panasonic DVX100 to a c-mount lens assembly. I plan on doing a write up on the project, but I would like to get some benchmark footage first. I have noticed a problem that has emerged since the conversion, but I am not 100% sure how to solve it. I have posted this in dvxuer but am hoping I might be able to get some suggestions from you guys. Basically the camera performs as it should in 30i, but in any progressive mode the camera suddenly cuts latitude dramatically, as if it jacked the master ped. to -15. All shadow detail is lost, and color is off. In addition the camera will not white-balance correctly in the progressive modes. It may somehow be related to the white balance system, however, since 30p looked ok until I performed the white-balance adjustment. Has anyone ever encountered a problem like this? I have all my settings to factory default, and have tried using the reset function. I have included a link to a video as well as some stills showing the effect. thanks in advance!! http://www.swingsetfilm.com/projects/dvxm/dvxm_test.mpg www.swingsetfilm.com/projects/dvxm/test30p.jpg www.swingsetfilm.com/projects/dvxm/test30p_2.jpg www.swingsetfilm.com/projects/dvxm/test24p.jpg www.swingsetfilm.com/projects/dvxm/test24p_2.jpg http://www.swingsetfilm.com/projects/dvxm/dvxm1.jpg http://www.swingsetfilm.com/projects/dvxm/dvxm2.jpg http://www.swingsetfilm.com/projects/dvxm/dvxm3.jpg
__________________
Sean Porter "progressive rocks" |
February 14th, 2005, 07:18 AM | #2 |
Inner Circle
Join Date: Dec 2003
Location: PERTH. W.A. AUSTRALIA.
Posts: 4,477
|
I suspect by using the "C" mount lens, the camera and its auto-exposure and light-metering system, which may depend upon zones within the CCDs image area has become confused. These things seem to be tailored for lens types in the still cameras.
You may have taken one element of auto-control out of the equation if your "C" Mount lenses are entirely manually operated. That element is iris or aperture setting. With the DVX lens, I don't know if there is a mechanical iris or the option of one, which is simply a switching over-ride of automatic circuits which emulates a manual lens iris. Many cams, although they notionally support a "manual" iris or aperture setting and neutral density filter setting, actually over-ride this when the circumstances suit them in order to continue to give an image to the user. I wonder if your camera is finding things rather dark when you set your "C" mount lens aperture for the sharpest image and is over-riding your choice settings in order to get any sort of image at all. You may well have to use your "C" mount lenses wide-open in order to give the camera's software some room to move in. I suspect the image is going to look pretty poor under those conditions unless you have good production lenses. What sort of "C" mount lenses are you using? Are they old Schneider Kreuznach or Kern Paillard primes for Bolex? If so, they might be a bit tight for light transmission onto three 1/3" CCDs. I suspect your original DVX optics are a far better performer in relation to light than "C" mount lenses. A Nikon f1.8 28mm lens on a 2/3" image intensifier looks brighter than a Cosmicar f1.8 22mm C-Mount lens. So it seems that not all aperture ratings for lenses across formats were created equal. Only a lens designer or technician can advise witrh quality here. Some speculations for thought is all the above is. I am no technician. When you modified your cam, did you keep the blue filter on front of the 3xCCD block? If you did not, maybe you have done some infra-red harm, either to the CCDs or to the image management system which might have been over-driven trying to compensate. This is entirely speculation on my part. If the blue filter is not there, this could be a reason for the white-balance going cockeyed as it will simply have run out of usable range. Gamma option on the JVC KY-F50 and the Kampro crimcam significantly darkens the image and adversely affects low-light performance. |
February 14th, 2005, 08:32 AM | #3 |
Major Player
Join Date: Dec 2003
Location: Brooklyn, NY
Posts: 636
|
Hope you get those issues solved -- otherwise, that's a great hack! Looking forward to the "how to"...
|
February 14th, 2005, 09:44 AM | #4 |
Major Player
Join Date: Dec 2004
Posts: 749
|
I am suspect alot of technical knowledge on the modifiers part, because I know I wouldn't want to do that to such an expensive camera!
|
February 14th, 2005, 01:04 PM | #5 |
Major Player
Join Date: Nov 2004
Location: San Francisco, CA
Posts: 613
|
Crazy man, crazy. Wish you'd done it for the XL1/XL2. Lots easier, I'd imagine. That would be a big missing link for some people here.
|
February 14th, 2005, 02:13 PM | #6 |
New Boot
Join Date: Jan 2004
Location: seattle, wa
Posts: 5
|
Bob-
Thank you for your suggestions. The lens is a cannon f1.4 7.5-90mm video lens for 1/2" CCD cameras. I can't be to sure of its overall quality, but from my experience with similar lenses in the past, its pretty nice. The aperature on the DVX is an actual iris, it is controlled mechanically via a coil. That circuit has been removed, it also had sensors for focus and a sensor for the ND filters. Do you think that the main board is expecting some sort of feedback from these devices to make a correct white balance adjustment? I would think this a possibility, but because the camera responds and operates as normal in 30i and not in the progressive modes, I am skeptical it is a feedback issue (unless of course the progressive system is completely different). atleast in 30i, it appears as if the camera responds well to the lens, white-balances correctly and does not affect lattitude. I have tried white-balancing at a variety of stops and lighting conditions, including full open. The blue filter is indeed still on the prism, so I dont think that is the issue. I suppose I could try plugging in the circuit harness for the lens control, and see if that changes anything - atleast that way the camera would be getting some "dummy" input which may be required for the progressive modes. Keep the suggestions coming - I really appreciate it thanks sean
__________________
Sean Porter "progressive rocks" |
February 14th, 2005, 09:59 PM | #7 |
Inner Circle
Join Date: Dec 2003
Location: PERTH. W.A. AUSTRALIA.
Posts: 4,477
|
That lens should be okay as 1/2" is a bigger patch than 1/3".
I can only speak from a PD150 perspective. That cam has a progressive mode but it can only manage 12.5 frames per second PAL. At 25 frames per second interlace it reproduces a marginally interior image, ie., an interpolatively enhanced half-field scan as a full two field 50Hz interlace will not fit whithin the time available for motion imaging. So it seems progressive scan places more of a data burden on a camcorder's brainpower. While the interlace scan works on your machine, the extra tasking created by progressive might be leaving your camera in a permanent state of catching up leaving some tasks undone on each frame. If you have a still-image or single-frame progressive facility like the Sony, try taking a single one of those. If a single-frame progessive image reproduces correctly, then a data bog is probably going on and you may have to explore means of reducing that load. Try using the Canon lens on f8 aperture on shutter speed 1/150 on gain 0db for starters. "C" mount. You can get a Nikon SLR to "C" mount adaptor. Just think of all the wonderful lenses you can put on that thing now including ridiculously long telephotos. If you have some contacts within the government security agencies, research agencies or NASA, maybe enquire with them for tech info. They may have already implemented similar hacks to enable back inventory of specialty lenses to be used but I suspect they might instead go for the "C" Mount camera head derivations of these cameras which are usualy to be found in laboratories attached to microscopes and the like. That's about the limit of my brainpower on this subject. Good luck. As for the innards of my camcorder seeing the light of day, you are a braver (possibly richer) man than I gungadin. I have tried to find specimens which have been run over or drowned to experiment upon, however the insurance companies take them away to some place of respectful interrment and I cannot get my hands on one. |
February 15th, 2005, 01:44 PM | #8 |
New Boot
Join Date: Jan 2004
Location: seattle, wa
Posts: 5
|
thank you very much for your response. I'll see if I can implement your suggestions. Unfortunately i have met little aid and alot of sarcasm on the other forums, so I guess I am on my own.
I will definately post once I get the camera operating correctly and still would appreciate any feedback or ideas. and I am definately not rich. possibly stupid but not rich. The project took three days and <$80 to complete. thanks
__________________
Sean Porter "progressive rocks" |
February 15th, 2005, 09:06 PM | #9 |
Major Player
Join Date: Mar 2004
Location: Melbourne, Australia
Posts: 223
|
Have a look at this hack done to a Sony FX1 http://www.eidomedia.com/hdve/ziess_fuji.htm. The video Fujinon lens compares way better than a Zeiss lens made for a film camera. I don't know if these (video) lenses are designed to compensate for the presence of the prism in the 3 chip cameras, but the "chromatic aberation" of the Zeiss lens is quite shocking.
|
February 15th, 2005, 11:40 PM | #10 |
Regular Crew
Join Date: Oct 2002
Location: Eldorado Hills CA
Posts: 68
|
That is a really awesome idea.
I hope you figure out the progressive problem, and can find a better lens with less aberation. I'd love to see the results with some nice primes on your system. |
February 25th, 2005, 02:43 AM | #11 |
Inner Circle
Join Date: Dec 2003
Location: PERTH. W.A. AUSTRALIA.
Posts: 4,477
|
There wasn an article on the VX1000 some time ago which addressed among many other things chromatic abberation. The article has since disappeared but what I recall is that the small size of modern CCDs brings forth a problem to do with the wavelengths of different colours of light, which is aggravted by small apertures.
Apparently, best practice with the VX1000 was to use an aperture of f5.6 which is contrary to accepted practice with larger formats of using smaller apertures unless going for a narrow depth of field. You might best set your lens to f5.6 or experiment in the aperture mid-range and let the camcorder automatic exposure look after the rest. The only other thing I can think of is back-focus of your "C"-mount lens. I assume your modification enables fine back-focus adjustments of the "C" mounted lens. I could not see from the pic if the lens has its own internal backfocus adjustment. If not, you might try screwing the lens forward a little on the "C"-mount thread to see if this worsens or improves the problem. |
September 9th, 2005, 02:41 PM | #12 |
Major Player
Join Date: Mar 2005
Location: Ventura, California, USA
Posts: 751
|
Did you ever get anywhere with this, Sean?
|
September 11th, 2005, 02:46 AM | #13 |
New Boot
Join Date: Aug 2005
Posts: 21
|
anything new with this? this is something that could be pretty huge if it can be made to work correctly in progresive scan modes. i hope the hard work continues and is paying off.
|
October 20th, 2005, 10:39 AM | #14 |
Regular Crew
Join Date: Mar 2005
Location: St.Louis, USA
Posts: 145
|
any updates
Sean,
All these technologies that we're using is based on the risk taken by someone at some point of time. Don't care about the sarcasm. Is there any new updates? Hari |
January 22nd, 2006, 07:15 PM | #15 |
Major Player
Join Date: Mar 2005
Location: Ventura, California, USA
Posts: 751
|
The site's down and the links are dead. I hope Sean puts it back up soon!
|
| ||||||
|
|