JVC GY-HD100-250 35mm Nikon-mount lens adapter comparison

Joachim Claus · March 15th, 2008, 12:16 PM

I think the "red flare issue" is not a result of the adapters. In my view the reason is the difference in the diameter of the pictures built by a standard 1/3 inch TV-lens and a photo-camera lens designed for 35 mm film. In the first case the diameter of the picture is some 6 mm, while in the case of the the "Nikon's" the diameter is 43 mm. As a result of the much bigger picture area, quite an amount of light is focussed besides the the active area of both, splitter and CCD-chip. This "vagabonding" light may cause reflections somewhere inside the video camera, which finally causes the red flare effects you have found. A measure to avoid it would be the inclusion of an aperture inside the adapter, which narrows down the light beam to the active area of splitter and CCD chip. May be you can test it with placing some black paper made apertures inside your adapter.
Joachim

Eric Gulbransen · March 15th, 2008, 01:44 PM

Sean, Mike Tapa (who responded here) IS MTF. I spoke with the US Zork rep (dealer/importer) about this flare problem a while back. Both he and Zork are aware of it (by the way, "Zoerk" spelling seems open to interpretation. Even Zork spells it both with and without the e). In fact it was Zork who told me the reflection was coming from the aluminum flange, and that I should blacken it with a marker (which I did, but it didn't seem to help). Good insight and ideas. I think it's time to get creative. Oh and I did look into the hue situation last night, against a pure white surface. The image I see is pure, even white, no matter the exposure (at least with the Nikons).

Joachim, thanks for the knowledge. You and Sean seem to be on a similar page, which sounds great because I have not heard these points before. It's such a great combination, the JVC and these telephoto lenses. I hope we can work this nagging detail out.

Eric Gulbransen · March 15th, 2008, 02:21 PM

Quote:

Originally Posted by Mario Cater

I'm having the exact same problem with my 80-200mm Nikon on an HD100...
On other threads I've seen posts by a guy who seems to be very knowledgeable about optics. His name is Ryan Damm. Maybe it'd be a good idea to invite him to give his 2 cents to our problem. I'm not a regular forum user and wouldn't know how to do this.

Mario, I emailed Ryan through the DVinfo channels. Maybe he'll chime in

Sean Adair · March 15th, 2008, 02:47 PM

The more I think about it, the more convinced I am that the 35mm sized image being projected in there is the problem. In some respects you could consider it an adapter problem, since it is a tool for the job. A mask or perhaps even a paper cone which cuts off the unused outer parts of the image, and blocks the image area from reflections could eliminate it.
Since there is a glass element in front of the ccd and nd filters, it is likely that this is the surface getting the reflected images.

We are definitely on the same track Joachim! (see my previous note).

Mike Tapa · March 16th, 2008, 01:19 AM

Sean, I think you hit the nail on the head.
I was just about to jump in with that theory, but you beat me to it.
Tomorrow, Im going to look at the design to see if I can incorporate a built in mask.

Tonight, I just sold the last adaptor of this batch, with the next batch arriving at months end-too late for the redesign.
If the mask idea works I may even be able to have some made up to send out to anyone already using my adaptor.

I should come to this forum more often instead of wasting precious hours watching F1 !!

Mike Tapa · March 16th, 2008, 01:47 AM

oops, sorry, you to Joachim deserve credit if the idea works

Cheers people.

Eric Gulbransen · March 16th, 2008, 10:00 AM

Mike, I'm game to dork around with a black paper version of a mask here in the mean time. Mario sent me a detailed email (with images) explaining what might be happening with the excess 35mm image area, the beam splitter, and the resulting phantom reflections. It seems we're all on a similar page. Or, you guys are on a similar page - I'm just reading along.

T E A M - together, everyone, achieves, more

and by the way Mike, MotoGP not F1. Too many tires..

Mario Cater · March 18th, 2008, 12:48 AM

Mike,
I have a feeling that a built-in mask will not universally work for all lenses. For any given lens the correct mask might depend on the location and size of it's exit pupil.

Ryan Damm · March 18th, 2008, 01:28 PM

Hi all,

Great discussion and thread -- good to see lens adapters have made such progress in the past year or so. My two cents -- the masks will probably solve the problem, at least the "lower third" red flare.

The flare is almost certainly a true image from the taking (Nikon) lens being reflected internally in the JVC's optics, because it's in focus and approximately the same size as the overall image -- suggesting the beam path isn't significantly longer than the light that should be hitting the CCD's. Also, it's an image, not some nondescript flare (according to the lower-barn-door test, and the monkey frame pulls), which suggests it's not a reflection off of the imaging optics -- it's a reflection off a flat surface (very likely the inside of the beam-splitting prism). (In the surfing shots, it might be the sand.)

I could write more (and might, when I have more time). In the meantime, some worthwhile reading:

en.wikipedia.org/wiki/Telecentric_Lens
en.wikipedia.org/wiki/Three-CCD

When you take into account the fact that the 35mm lenses will always cover a larger area than the beam-splitting prism, and the fact that the rays that strike the edges of the frame aren't perpendicular (as would be in a telecentric design, which is usually required by the 3CCD design), some stray rays are able to bounce around inside the prism. The good news: looks like they're non-imaging rays anyway. You'd see this much worse, by the way, if you got a non-telecentric, non-retrofocus medium- to wide-angle lens (where the peripheral rays are extremely slanted). Then, it might not be fixable.

So -- hopefully it works. And thanks for the vote of confidence -- I have to confess to falling off in my posts in the last year; I'm getting a Red so adapters aren't as mission-critical. Though I do have an exciting new project... 4-dimensional imaging (long story -- for fun, check out www.refocusimaging.com, a friend's startup). I'm in the early stages of a conceptual project that would use computational imaging -- way cool.

PS -- you may actually want a double mask: one right at the beam-splitting prism, and one further out to catch stray rays... it occurs to me that a double mask will effectively vignette the incoming rays -- if done properly you can vignette the flare-causing strays while leaving the imaging rays mostly alone. A single mask (particularly at the prism) probably won't do it.

Mike Tapa · March 18th, 2008, 03:34 PM

Hello again.

Looking at the design of my adaptor again, I propose to incorporate a mask but not close to the prism as this will not solve the problem, as Ryan and Mario pointed out
Ideally the mask needs to act as an aperture close to the rear element of whatever lens is being used. However, as the physical position of the rear element will vary depending on which lens you have, it looks like I will have to try a compromise position for the mask.

Approx 17mmØ @ around 9.5 to 10mm in from the mounting face of the Nikon mount.

I will let you know how I get on.

Ryan Damm · March 18th, 2008, 05:51 PM

It turns out you ideally want to mask it in both places -- the light leaking through is probably stray rays from near the outside of the aperture (which is why it disappears at small f-numbers) that then strike the beam-splitting prism at an oblique angle.

The mask behind the back of the lens will help -- but because (as an earlier post pointed out -- sorry, I can't remember the cite) it's not near the focal plane, it will effectively be out of focus. This might be okay -- again, you're probably talking about stray rays near the edge of the aperture, so if you clip them correctly, you might be occluding the rays you want to occlude.

The mask at the prism block will do part of it, but again, I suspect the light being admitted is along the prism block, so you can't block the bad light without blocking the desirable light.

Finally, a matte box in front of the lens will do the absolute most. Keep that light from falling onto the lens in the first place, and you're in the clear. Again, the fact that the light appears not to be coming from inside the frame is the saving grace, here.

Are there any other flares that are concerning, or is it just the red flare? (Incidentally, it's red because this light is coming in at such an extreme angle, the dichroic mirrors don't reflect it in the typical, designed way -- so all the stray light is ending up on the red sensor. The wikipedia page on 3-CCD cameras has a pretty decent diagram, you can see that the prism block is offset.)

Mario Cater · March 19th, 2008, 01:52 AM

mike, i think a rectangular (16:9) mask would be more effective than a round one.
I believe Eric is doing some tests with different size masks already.

Eric Gulbransen · March 21st, 2008, 12:01 AM

Sorry guys, I'm in NY right now NOT doing anything related to video. In the last days rushing to complete a framing job before I left I did not have time to experiment with creating a paper mask. I will get on that starting this upcoming Tuesday.

While on the subject of using this camera, these adapters, and those lenses - I thought I'd share my most recent exercise. A few times now, quite appropriately, the issue of focusing, and smoothly panning while using such telephoto ranges has come up during talks about using our cameras this way. This past Sunday I went out for more practice - facing the red flare problem, the steady panning challenges, and follow focusing. It was a horrible shoot, aiming directly at the sun, but that's what I had to deal with on this day. There IS some red flare, but I did well to avoid it most of the time by staying above f8. Thanks to the testing here, I was quite happy to get such little flare, especially considering I was shooting the high glare of water AND shooting toward the bright sun. As for the steady tripod work, I didn't nail that all the time either - but I have made some progress. I can't take the credit though, it's the new tripod..

http://www.reelsense.net/QT/Kiteboar...l_3-16-08.html

Jeffrey Butler · March 31st, 2008, 01:01 PM

Eric - I really enjoyed your CA coastline video...I mean really...thanks.

http://www.Reelsense.net/QT/GoforIT/GoForIT.html

Oh, and that link above - have you been like, practicing or something? Nice camera work on a tough shoot!

Eric Gulbransen · May 21st, 2008, 12:33 AM

OK guys, back to the red flare problem.

I have gone at this a few times by now and not gotten anywhere. Tonight though, I really tried. I created a mask from black paper, cut out a nice 16x9 hole in the center and fastened it to the inside of the adapter. I wasn't so concerned with size at first - just wanted to see what effect cutting all that unused image out of the equation would do. I set up the same light/barn door scenario and tried it out. Flare was still there, looking and behaving exactly the same. So I pulled the adapter and lens apart again and closed up the 16x9 opening in the mask. The first opening in the mask was about 3/4" wide. The second was about half an inch wide. Same result, flare everywhere just like before, only now the overexposed areas weren't so over exposed. Third time proved NOT to be a charm either. Finally on my fourth go around, which gave me a ridiculously small 16x9 opening in the mask (@1/4"x3/16"), the overexposed situation that I created so well with the 600 watt light shining against a white wall was no longer overexposed at all. But the red flare was still there, exactly the same look, size, and behavior.

I was really looking forward to reporting good things, as all the theories on this problem make great sense. But I think we might be barking up the wrong tree here. Did I do something wrong? Any more ideas?

March 15th, 2008, 02:47 PM	#19
Sean Adair Major Player Join Date: Dec 2006 Location: New York, NY Posts: 463	The more I think about it, the more convinced I am that the 35mm sized image being projected in there is the problem. In some respects you could consider it an adapter problem, since it is a tool for the job. A mask or perhaps even a paper cone which cuts off the unused outer parts of the image, and blocks the image area from reflections could eliminate it. Since there is a glass element in front of the ccd and nd filters, it is likely that this is the surface getting the reflected images. We are definitely on the same track Joachim! (see my previous note). __________________ Sean Adair - NYC - www.adairproductions.com JVC GY-HM-700 with 17x5 lens, MacPro 3.2ghz 8-core, 18gb. (JVC HD200 4 sale soon)

March 18th, 2008, 01:28 PM	#24
Ryan Damm Regular Crew Join Date: Dec 2005 Posts: 46	The Mask will probably work Hi all, Great discussion and thread -- good to see lens adapters have made such progress in the past year or so. My two cents -- the masks will probably solve the problem, at least the "lower third" red flare. The flare is almost certainly a true image from the taking (Nikon) lens being reflected internally in the JVC's optics, because it's in focus and approximately the same size as the overall image -- suggesting the beam path isn't significantly longer than the light that should be hitting the CCD's. Also, it's an image, not some nondescript flare (according to the lower-barn-door test, and the monkey frame pulls), which suggests it's not a reflection off of the imaging optics -- it's a reflection off a flat surface (very likely the inside of the beam-splitting prism). (In the surfing shots, it might be the sand.) I could write more (and might, when I have more time). In the meantime, some worthwhile reading: en.wikipedia.org/wiki/Telecentric_Lens en.wikipedia.org/wiki/Three-CCD When you take into account the fact that the 35mm lenses will always cover a larger area than the beam-splitting prism, and the fact that the rays that strike the edges of the frame aren't perpendicular (as would be in a telecentric design, which is usually required by the 3CCD design), some stray rays are able to bounce around inside the prism. The good news: looks like they're non-imaging rays anyway. You'd see this much worse, by the way, if you got a non-telecentric, non-retrofocus medium- to wide-angle lens (where the peripheral rays are extremely slanted). Then, it might not be fixable. So -- hopefully it works. And thanks for the vote of confidence -- I have to confess to falling off in my posts in the last year; I'm getting a Red so adapters aren't as mission-critical. Though I do have an exciting new project... 4-dimensional imaging (long story -- for fun, check out www.refocusimaging.com, a friend's startup). I'm in the early stages of a conceptual project that would use computational imaging -- way cool. PS -- you may actually want a double mask: one right at the beam-splitting prism, and one further out to catch stray rays... it occurs to me that a double mask will effectively vignette the incoming rays -- if done properly you can vignette the flare-causing strays while leaving the imaging rays mostly alone. A single mask (particularly at the prism) probably won't do it.

March 18th, 2008, 05:51 PM	#26
Ryan Damm Regular Crew Join Date: Dec 2005 Posts: 46	Two Masks are better than one It turns out you ideally want to mask it in both places -- the light leaking through is probably stray rays from near the outside of the aperture (which is why it disappears at small f-numbers) that then strike the beam-splitting prism at an oblique angle. The mask behind the back of the lens will help -- but because (as an earlier post pointed out -- sorry, I can't remember the cite) it's not near the focal plane, it will effectively be out of focus. This might be okay -- again, you're probably talking about stray rays near the edge of the aperture, so if you clip them correctly, you might be occluding the rays you want to occlude. The mask at the prism block will do part of it, but again, I suspect the light being admitted is along the prism block, so you can't block the bad light without blocking the desirable light. Finally, a matte box in front of the lens will do the absolute most. Keep that light from falling onto the lens in the first place, and you're in the clear. Again, the fact that the light appears not to be coming from inside the frame is the saving grace, here. Are there any other flares that are concerning, or is it just the red flare? (Incidentally, it's red because this light is coming in at such an extreme angle, the dichroic mirrors don't reflect it in the typical, designed way -- so all the stray light is ending up on the red sensor. The wikipedia page on 3-CCD cameras has a pretty decent diagram, you can see that the prism block is offset.)

March 31st, 2008, 01:01 PM	#29
Jeffrey Butler Major Player Join Date: Feb 2007 Location: Virgina, USA Posts: 276	Eric - I really enjoyed your CA coastline video...I mean really...thanks. http://www.Reelsense.net/QT/GoforIT/GoForIT.html Oh, and that link above - have you been like, practicing or something? Nice camera work on a tough shoot! __________________ Jeff Butler Butler Films

March 15th, 2008, 12:16 PM	#16
Joachim Claus Regular Crew Join Date: Jun 2006 Location: Seeheim-Jugenheim, Germany Posts: 132	I think the "red flare issue" is not a result of the adapters. In my view the reason is the difference in the diameter of the pictures built by a standard 1/3 inch TV-lens and a photo-camera lens designed for 35 mm film. In the first case the diameter of the picture is some 6 mm, while in the case of the the "Nikon's" the diameter is 43 mm. As a result of the much bigger picture area, quite an amount of light is focussed besides the the active area of both, splitter and CCD-chip. This "vagabonding" light may cause reflections somewhere inside the video camera, which finally causes the red flare effects you have found. A measure to avoid it would be the inclusion of an aperture inside the adapter, which narrows down the light beam to the active area of splitter and CCD chip. May be you can test it with placing some black paper made apertures inside your adapter. Joachim

March 15th, 2008, 01:44 PM	#17
Eric Gulbransen Major Player Join Date: Nov 2006 Location: Palo Alto, California Posts: 520	Sean, Mike Tapa (who responded here) IS MTF. I spoke with the US Zork rep (dealer/importer) about this flare problem a while back. Both he and Zork are aware of it (by the way, "Zoerk" spelling seems open to interpretation. Even Zork spells it both with and without the e). In fact it was Zork who told me the reflection was coming from the aluminum flange, and that I should blacken it with a marker (which I did, but it didn't seem to help). Good insight and ideas. I think it's time to get creative. Oh and I did look into the hue situation last night, against a pure white surface. The image I see is pure, even white, no matter the exposure (at least with the Nikons). Joachim, thanks for the knowledge. You and Sean seem to be on a similar page, which sounds great because I have not heard these points before. It's such a great combination, the JVC and these telephoto lenses. I hope we can work this nagging detail out.

March 16th, 2008, 01:19 AM	#20
Mike Tapa Sponsor: MTF Services Join Date: Mar 2006 Location: London UK Posts: 80	Sean, I think you hit the nail on the head. I was just about to jump in with that theory, but you beat me to it. Tomorrow, Im going to look at the design to see if I can incorporate a built in mask. Tonight, I just sold the last adaptor of this batch, with the next batch arriving at months end-too late for the redesign. If the mask idea works I may even be able to have some made up to send out to anyone already using my adaptor. I should come to this forum more often instead of wasting precious hours watching F1 !!

March 16th, 2008, 01:47 AM	#21
Mike Tapa Sponsor: MTF Services Join Date: Mar 2006 Location: London UK Posts: 80	oops, sorry, you to Joachim deserve credit if the idea works Cheers people.

March 16th, 2008, 10:00 AM	#22
Eric Gulbransen Major Player Join Date: Nov 2006 Location: Palo Alto, California Posts: 520	Mike, I'm game to dork around with a black paper version of a mask here in the mean time. Mario sent me a detailed email (with images) explaining what might be happening with the excess 35mm image area, the beam splitter, and the resulting phantom reflections. It seems we're all on a similar page. Or, you guys are on a similar page - I'm just reading along. T E A M - together, everyone, achieves, more and by the way Mike, MotoGP not F1. Too many tires..

March 18th, 2008, 12:48 AM	#23
Mario Cater Tourist Join Date: Mar 2007 Location: New Westminster, Canada Posts: 4	Mike, I have a feeling that a built-in mask will not universally work for all lenses. For any given lens the correct mask might depend on the location and size of it's exit pupil.

March 18th, 2008, 03:34 PM	#25
Mike Tapa Sponsor: MTF Services Join Date: Mar 2006 Location: London UK Posts: 80	Hello again. Looking at the design of my adaptor again, I propose to incorporate a mask but not close to the prism as this will not solve the problem, as Ryan and Mario pointed out Ideally the mask needs to act as an aperture close to the rear element of whatever lens is being used. However, as the physical position of the rear element will vary depending on which lens you have, it looks like I will have to try a compromise position for the mask. Approx 17mmØ @ around 9.5 to 10mm in from the mounting face of the Nikon mount. I will let you know how I get on.

March 19th, 2008, 01:52 AM	#27
Mario Cater Tourist Join Date: Mar 2007 Location: New Westminster, Canada Posts: 4	mike, i think a rectangular (16:9) mask would be more effective than a round one. I believe Eric is doing some tests with different size masks already.

March 21st, 2008, 12:01 AM	#28
Eric Gulbransen Major Player Join Date: Nov 2006 Location: Palo Alto, California Posts: 520	Sorry guys, I'm in NY right now NOT doing anything related to video. In the last days rushing to complete a framing job before I left I did not have time to experiment with creating a paper mask. I will get on that starting this upcoming Tuesday. While on the subject of using this camera, these adapters, and those lenses - I thought I'd share my most recent exercise. A few times now, quite appropriately, the issue of focusing, and smoothly panning while using such telephoto ranges has come up during talks about using our cameras this way. This past Sunday I went out for more practice - facing the red flare problem, the steady panning challenges, and follow focusing. It was a horrible shoot, aiming directly at the sun, but that's what I had to deal with on this day. There IS some red flare, but I did well to avoid it most of the time by staying above f8. Thanks to the testing here, I was quite happy to get such little flare, especially considering I was shooting the high glare of water AND shooting toward the bright sun. As for the steady tripod work, I didn't nail that all the time either - but I have made some progress. I can't take the credit though, it's the new tripod.. http://www.reelsense.net/QT/Kiteboar...l_3-16-08.html

May 21st, 2008, 12:33 AM	#30
Eric Gulbransen Major Player Join Date: Nov 2006 Location: Palo Alto, California Posts: 520	OK guys, back to the red flare problem. I have gone at this a few times by now and not gotten anywhere. Tonight though, I really tried. I created a mask from black paper, cut out a nice 16x9 hole in the center and fastened it to the inside of the adapter. I wasn't so concerned with size at first - just wanted to see what effect cutting all that unused image out of the equation would do. I set up the same light/barn door scenario and tried it out. Flare was still there, looking and behaving exactly the same. So I pulled the adapter and lens apart again and closed up the 16x9 opening in the mask. The first opening in the mask was about 3/4" wide. The second was about half an inch wide. Same result, flare everywhere just like before, only now the overexposed areas weren't so over exposed. Third time proved NOT to be a charm either. Finally on my fourth go around, which gave me a ridiculously small 16x9 opening in the mask (@1/4"x3/16"), the overexposed situation that I created so well with the 600 watt light shining against a white wall was no longer overexposed at all. But the red flare was still there, exactly the same look, size, and behavior. I was really looking forward to reporting good things, as all the theories on this problem make great sense. But I think we might be barking up the wrong tree here. Did I do something wrong? Any more ideas?