MTF 1/3" Nikon "G" Adapter

Steve Phillipps · April 27th, 2009, 01:31 AM

There was a thread on another forum ages ago about this, and this was the scientific explanation from BBC technical expert Alan Roberts:

"JVC have adaptor rings to fit larger format lenses, particularly 1"/2 and 2"/3 format. BUT, you must beware of fitting odd lenses, because these adaptors have no glass, they are just adaptors. This means that even a top-notch HD lens may well look soft on it, simply because the lens is computed for a larger format. This camera has pixels spaced 3.33 microns apart while a 2"/3 HD camera has pixels at 5 microns. So, a 2"/3 HD lens like a Zeiss prime (4 micron disc of confusion) will probably be pretty good on the HD100, but any lesser lens (i.e. any zoom) is going to look a bit on the soft side. Bear in mind that the limiting resolution of the HD100 is 300 lines/mm, that should tell you what makes sense in terms of fitting other lenses (e.g. a 35mm still lens making that resolution would produce 10,800 lines/pic width, and I know of no lens that will do that). A decent 2"/3 SD tv lens will have a disc of confusion of 15 microns or so, so can produce only about 66 lines/mm, a lot less than the camera wants."

This is the science I'm basing my comments on and that I'm not qualified to say myself!

Steve

Bill Ravens · April 27th, 2009, 06:47 AM

I came across an interesting article I'll share here:
Grain aliasing

The point of the article attempts to relate film grain size to pixel size. The author makes the point that apparent and actual resolving power is influenced by the aliasing that occurs between the discrete photon sites. It may well be an explanation of why a 35mm lens looks "better" than a 1/3 inch lens.

Steve Phillipps · April 27th, 2009, 07:05 AM

Well I don't have a 1/3" camera to test it with, but as I said, I tested the EX3 stock lens against some of the world's best ever Nikkors and the EX lens won in terms of sharpness and CA (although there was little CA in either). Once again, the science tells us that it should be even more prevelant with the smaller 1/3" chip so I am surprised at what you're finding.
Steve

Bill Ravens · April 27th, 2009, 07:17 AM

A quick calculation shows that for the HD110 with a 1/3 inch(diagonal) dimension and 1,110,000 pixels, the vertical pixel size is on the order of 8 micrometers. The result of a Kodak study of 35mm film grain size shows grain can vary from .8 to 3 micrometers in size. On this basis, alone, I would say the JVC resolving power is about 3x worse than 35mm film. Again, on this basis, the film lens should be higher resolution than the videocam lens.

Steve Phillipps · April 27th, 2009, 07:26 AM

Certainly getting beyond my scientific expertise now! But I do know there to be problems when relating film grain to pixels, it's not quite as straightforward as one set of number vs another. Otherwise Super 16 film would be many times more higher resolution than HD, and it isn't.
Steve

Bill Ravens · April 27th, 2009, 07:35 AM

Last year, I was involved in a film to digital project , in which super 16 film was digitized to Avid DNxHD. We went thru several digitizations(don't ask why) including DNx175 and DNx220. The improvement in image quality at DNx220 was approaching the images available from the raw film stock when projected. This is a VERY empirical conclusion, but, I would say 2K resolution is fairly equivalent to super16. THerefore, I would also assume 4K resolution is equivalent to 35mm film. Admittedly, there's many, many factors involved, here(production monitor quality, aliasing from the digitization process, etc., etc). I draw my conclusions on what my eyes tell me, not the physics.

One more caveat...I have fairly old eyes....whatever that may mean.

Steve Phillipps · April 27th, 2009, 08:38 AM

I'm sure I remember a seminar where the guys from Kodak were trying not to get swamped the the HD revolution and were quoting 35mm film as having a resolution equivalent of something like 30 million pixels (can't remember exactly). Then the next seminar by Sony would show how much more resolving power there was on an F900's 2mp sensor than 35mm filmstock.
Steve

Alex Humphrey · April 27th, 2009, 09:41 AM

interesting thread, I was going to comment earlier, but i've been under the weather. We are totally off subject by this point. The thread was started off asking about focal ranges from 70mm to 300mm with a MTF and Nikkor lens combo and NOW we are on comparing (assumed?) the same focal range comparing 3rd party lenses and video lenses (meaning does a Nikkor with MTF set at 40mm compare to the Fujinon 16x or 17x lens at or around 40mm?). There is a lot to be said for testing.

Anyway since the 16x is fairly useless beyond 50 or 60mm and worthless at 88mm unless you have it stopped down to f11, and the 17x should be much more servicable at the same focal length, but neither function 88-300mm. For that there are few options. The results of the MTF by several users (not mine) have been extremely good. a Nikkor 80-200 was as good as, or maybe better than a 17x at 40mm at f4. That should be all you need to know.

Now I'll to add to the fire, I have a used Nikkor 300mm f4 (shipping to me from an ebay seller) and I'll order the newest MTF adapter this coming month and give it a whirl. Should be insane magnification. Nice views of a quarter of the moon filling the screen rising over a city scape is what I have in mind. I guess I'll need a better tripod and a wind break for that shot. I'll also throw on some of my nikon e and nikkor 28, 35, 50 and 100mm as well as zoom lenses and give them a whirl compared to the 16x lens I have at similar focal ranges. Now I'm not saying I would want to switch lenses in general practice, but it would be interesting. You have to remember, even the 16x isn't bad (6mm-50mm or so) I mean how many Canon and Nikon lenses do you see that are 5.5-88mm f1.4 with electronicly driven zoom? Not many.

Now if I only had a resolution test chart other than printing out a page of periods. Well maybe someone in northern california can meet me for coffee one saturday next month and we can try out some of these ideas and post the pics.

Bill Ravens · April 27th, 2009, 11:04 AM

Alex...

I use that 300mm Nikon lens for wildlife videography. It gives an excellent, hi-contrast image that I find very acceptable, with no CA. Again, my subjective opinion, but, to my eyes, it looks sharper than the Fujinon 17x lens images. My only complaint is the thermals/shimmering in the air really distort the images at long distances. Not much a lens maker can do to fix this...;o). For anyone who knows what strehl is, some days are better than others.

Steve Phillipps · April 27th, 2009, 11:11 AM

Quote:

Originally Posted by Alex Humphrey

Anyway since the 16x is fairly useless beyond 50 or 60mm and worthless at 88mm unless you have it stopped down to f11, .

Now that's another interesting optical kettle of fish, it should start to deteriorate in sharpness past about f5.6 due to diffraction, and on a 1/3" chip f11 is well well below the diffraction limit. One conclusion to draw from this is that the lens is a real dog, and is so bad at the lower apertures that it looks better at f11 despite the massive diffraction problems.
Steve

Stefan Immler · April 27th, 2009, 12:24 PM

Quote:

Originally Posted by Alex Humphrey

Now if I only had a resolution test chart ...

Try this one that I have just created:

http://giganova.com/Misc/test_chart.pdf (PDF, 2MB)
http://giganova.com/Misc/test_chart.tif (high-res TIF, 32MB)

It's not a scientific test charts but should do the job.

Robert Rogoz · April 27th, 2009, 02:34 PM

Quote:

Originally Posted by Steve Phillipps

I'm sure I remember a seminar where the guys from Kodak were trying not to get swamped the the HD revolution and were quoting 35mm film as having a resolution equivalent of something like 30 million pixels (can't remember exactly). Then the next seminar by Sony would show how much more resolving power there was on an F900's 2mp sensor than 35mm filmstock.
Steve

Steve, I don't get it, why do you insist on ignoring the rules of physics. Of course, where the quality counts, film will win hands down. First of all pixels are squares and film is built with irregular multi-layer of grain (3D like). If you compare dynamic range and detail, film is still far superior to any digital imagining method. A new 24.5 MP Nikon D3X produces files about 75MB in size. On the other hand a drum scan of a slide will produce files about 80-120MB. This is the newest and latest camera from Nikon, costing 8K for the body. A medium format doesn't even compare as far as film and digital.
BTW, any Nikon or Canon 35mm lens will blow Fujinon 16 or 17 lens out of the water. You'll need good, external monitor for manual focus. Pure physics, that's all.

Steve Phillipps · April 27th, 2009, 03:13 PM

Quote:

Originally Posted by Robert Rogoz

Steve, I don't get it, why do you insist on ignoring the rules of physics. Of course, where the quality counts, film will win hands down. First of all pixels are squares and film is built with irregular multi-layer of grain (3D like). If you compare dynamic range and detail, film is still far superior to any digital imagining method. A new 24.5 MP Nikon D3X produces files about 75MB in size. On the other hand a drum scan of a slide will produce files about 80-120MB. This is the newest and latest camera from Nikon, costing 8K for the body. A medium format doesn't even compare as far as film and digital.
BTW, any Nikon or Canon 35mm lens will blow Fujinon 16 or 17 lens out of the water. You'll need good, external monitor for manual focus. Pure physics, that's all.

On the whole, haven't got a clue what you're talking about really. Are you agreeing or disagreeing with me?
BTW not true about the Nikon and Canon lens statement, don't know if you've got examples to show it, but my top-notch Nikkors didn't match the EX3's Fujinon.
External monitor for manual focus? Not in my game, we seem to do OK, can't carry monitors out in the hills.
I wish people wouldn't start posts with the likes of "why do you insist on...", there's no need for aggression in any of these forums, I thought we were all supposed to be just giving our observations and opinions to help each other?
Steve

Steve Phillipps · April 27th, 2009, 03:24 PM

Robert, just read your post again and it makes even less sense now. You start by saying that film wins hands down for quality then later say that medium format doesn't anywhere near compare to a DSLR. Make your mind up.
And did I ever say that film doesn't beat video for dynamic range? Don't think so.
And as for ignoring the rules of physcis, if you'll look at the posts it was actually me that wa quoting the physics and others were contesting it.
All in all one of the most baffling posts I've ever seen.
Steve

Matt Lehman · April 27th, 2009, 08:14 PM

Gentlemen (and ladies), the current thread is extremely interesting but I just wish no one would continue to point out the downfalls of my HD100 :)

Does anyone have footage, shots, stills, etc from this MTF adapter for the GY-HD series? I'm extremely curious about it and would love to have some advice about the adapter itself and the lens they have used with it.

Thanks!
Matt

April 27th, 2009, 08:14 PM	#30
Matt Lehman Regular Crew Join Date: Nov 2008 Location: Honolulu, HI Posts: 35	Back to the adapter itself and Nikon's lens Gentlemen (and ladies), the current thread is extremely interesting but I just wish no one would continue to point out the downfalls of my HD100 :) Does anyone have footage, shots, stills, etc from this MTF adapter for the GY-HD series? I'm extremely curious about it and would love to have some advice about the adapter itself and the lens they have used with it. Thanks! Matt

April 27th, 2009, 01:31 AM	#16
Steve Phillipps Inner Circle Join Date: Apr 2006 Location: Wales Posts: 2,130	There was a thread on another forum ages ago about this, and this was the scientific explanation from BBC technical expert Alan Roberts: "JVC have adaptor rings to fit larger format lenses, particularly 1"/2 and 2"/3 format. BUT, you must beware of fitting odd lenses, because these adaptors have no glass, they are just adaptors. This means that even a top-notch HD lens may well look soft on it, simply because the lens is computed for a larger format. This camera has pixels spaced 3.33 microns apart while a 2"/3 HD camera has pixels at 5 microns. So, a 2"/3 HD lens like a Zeiss prime (4 micron disc of confusion) will probably be pretty good on the HD100, but any lesser lens (i.e. any zoom) is going to look a bit on the soft side. Bear in mind that the limiting resolution of the HD100 is 300 lines/mm, that should tell you what makes sense in terms of fitting other lenses (e.g. a 35mm still lens making that resolution would produce 10,800 lines/pic width, and I know of no lens that will do that). A decent 2"/3 SD tv lens will have a disc of confusion of 15 microns or so, so can produce only about 66 lines/mm, a lot less than the camera wants." This is the science I'm basing my comments on and that I'm not qualified to say myself! Steve

April 27th, 2009, 06:47 AM	#17
Bill Ravens Join Date: Nov 2001 Location: santa fe, nm Posts: 3,264 Images: 10	I came across an interesting article I'll share here: Grain aliasing The point of the article attempts to relate film grain size to pixel size. The author makes the point that apparent and actual resolving power is influenced by the aliasing that occurs between the discrete photon sites. It may well be an explanation of why a 35mm lens looks "better" than a 1/3 inch lens.

April 27th, 2009, 07:05 AM	#18
Steve Phillipps Inner Circle Join Date: Apr 2006 Location: Wales Posts: 2,130	Well I don't have a 1/3" camera to test it with, but as I said, I tested the EX3 stock lens against some of the world's best ever Nikkors and the EX lens won in terms of sharpness and CA (although there was little CA in either). Once again, the science tells us that it should be even more prevelant with the smaller 1/3" chip so I am surprised at what you're finding. Steve

April 27th, 2009, 07:17 AM	#19
Bill Ravens Join Date: Nov 2001 Location: santa fe, nm Posts: 3,264 Images: 10	A quick calculation shows that for the HD110 with a 1/3 inch(diagonal) dimension and 1,110,000 pixels, the vertical pixel size is on the order of 8 micrometers. The result of a Kodak study of 35mm film grain size shows grain can vary from .8 to 3 micrometers in size. On this basis, alone, I would say the JVC resolving power is about 3x worse than 35mm film. Again, on this basis, the film lens should be higher resolution than the videocam lens.

April 27th, 2009, 07:26 AM	#20
Steve Phillipps Inner Circle Join Date: Apr 2006 Location: Wales Posts: 2,130	Certainly getting beyond my scientific expertise now! But I do know there to be problems when relating film grain to pixels, it's not quite as straightforward as one set of number vs another. Otherwise Super 16 film would be many times more higher resolution than HD, and it isn't. Steve

April 27th, 2009, 07:35 AM	#21
Bill Ravens Join Date: Nov 2001 Location: santa fe, nm Posts: 3,264 Images: 10	Last year, I was involved in a film to digital project , in which super 16 film was digitized to Avid DNxHD. We went thru several digitizations(don't ask why) including DNx175 and DNx220. The improvement in image quality at DNx220 was approaching the images available from the raw film stock when projected. This is a VERY empirical conclusion, but, I would say 2K resolution is fairly equivalent to super16. THerefore, I would also assume 4K resolution is equivalent to 35mm film. Admittedly, there's many, many factors involved, here(production monitor quality, aliasing from the digitization process, etc., etc). I draw my conclusions on what my eyes tell me, not the physics. One more caveat...I have fairly old eyes....whatever that may mean.

April 27th, 2009, 08:38 AM	#22
Steve Phillipps Inner Circle Join Date: Apr 2006 Location: Wales Posts: 2,130	I'm sure I remember a seminar where the guys from Kodak were trying not to get swamped the the HD revolution and were quoting 35mm film as having a resolution equivalent of something like 30 million pixels (can't remember exactly). Then the next seminar by Sony would show how much more resolving power there was on an F900's 2mp sensor than 35mm filmstock. Steve

April 27th, 2009, 09:41 AM	#23
Alex Humphrey Major Player Join Date: Mar 2007 Location: california North and South Posts: 642	interesting thread, I was going to comment earlier, but i've been under the weather. We are totally off subject by this point. The thread was started off asking about focal ranges from 70mm to 300mm with a MTF and Nikkor lens combo and NOW we are on comparing (assumed?) the same focal range comparing 3rd party lenses and video lenses (meaning does a Nikkor with MTF set at 40mm compare to the Fujinon 16x or 17x lens at or around 40mm?). There is a lot to be said for testing. Anyway since the 16x is fairly useless beyond 50 or 60mm and worthless at 88mm unless you have it stopped down to f11, and the 17x should be much more servicable at the same focal length, but neither function 88-300mm. For that there are few options. The results of the MTF by several users (not mine) have been extremely good. a Nikkor 80-200 was as good as, or maybe better than a 17x at 40mm at f4. That should be all you need to know. Now I'll to add to the fire, I have a used Nikkor 300mm f4 (shipping to me from an ebay seller) and I'll order the newest MTF adapter this coming month and give it a whirl. Should be insane magnification. Nice views of a quarter of the moon filling the screen rising over a city scape is what I have in mind. I guess I'll need a better tripod and a wind break for that shot. I'll also throw on some of my nikon e and nikkor 28, 35, 50 and 100mm as well as zoom lenses and give them a whirl compared to the 16x lens I have at similar focal ranges. Now I'm not saying I would want to switch lenses in general practice, but it would be interesting. You have to remember, even the 16x isn't bad (6mm-50mm or so) I mean how many Canon and Nikon lenses do you see that are 5.5-88mm f1.4 with electronicly driven zoom? Not many. Now if I only had a resolution test chart other than printing out a page of periods. Well maybe someone in northern california can meet me for coffee one saturday next month and we can try out some of these ideas and post the pics.

April 27th, 2009, 11:04 AM	#24
Bill Ravens Join Date: Nov 2001 Location: santa fe, nm Posts: 3,264 Images: 10	Alex... I use that 300mm Nikon lens for wildlife videography. It gives an excellent, hi-contrast image that I find very acceptable, with no CA. Again, my subjective opinion, but, to my eyes, it looks sharper than the Fujinon 17x lens images. My only complaint is the thermals/shimmering in the air really distort the images at long distances. Not much a lens maker can do to fix this...;o). For anyone who knows what strehl is, some days are better than others.

April 27th, 2009, 03:24 PM	#29
Steve Phillipps Inner Circle Join Date: Apr 2006 Location: Wales Posts: 2,130	Robert, just read your post again and it makes even less sense now. You start by saying that film wins hands down for quality then later say that medium format doesn't anywhere near compare to a DSLR. Make your mind up. And did I ever say that film doesn't beat video for dynamic range? Don't think so. And as for ignoring the rules of physcis, if you'll look at the posts it was actually me that wa quoting the physics and others were contesting it. All in all one of the most baffling posts I've ever seen. Steve