Give JVC and their Lens a break - It passed a tough test

[Barry Green]

Actually, if I'm not mistaken, they already showed a mini35 on the JVC at NAB.

Also, Bill Turner of Century Precision Optics weighed in about the quality vs. price debate, and he said something interesting -- he said that the physical size of the lens is a big factor in how expensive it is. The bigger the glass, the higher the price. And as he said that, I thought back to how tiny the JVC stock lens was... and thought maybe, maybe that's a good sign? If it's so much less expensive to make a small lens, then maybe the low price isn't directly a reflection of low quality, but perhaps is just saying that the lens is fine, and the cost is a reflection of its small size?

[Michael Pappas]

That's true Barry! Good point.

Is that picture I posted, the same design that was at NAB on the JVC? This one looks like bit smaller too. It could just be the picture though. This one is a B4 mount version posted I believe....

Originally Posted by Barry Green

Actually, if I'm not mistaken, they already showed a mini35 on the JVC at NAB.

Also, Bill Turner of Century Precision Optics weighed in about the quality vs. price debate, and he said something interesting -- he said that the physical size of the lens is a big factor in how expensive it is. The bigger the glass, the higher the price. And as he said that, I thought back to how tiny the JVC stock lens was... and thought maybe, maybe that's a good sign? If it's so much less expensive to make a small lens, then maybe the low price isn't directly a reflection of low quality, but perhaps is just saying that the lens is fine, and the cost is a reflection of its small size?

[Tim Le]

Originally Posted by Barry Green

If it's so much less expensive to make a small lens, then maybe the low price isn't directly a reflection of low quality, but perhaps is just saying that the lens is fine, and the cost is a reflection of its small size?

Good point Barry.

Admittedly I'm no optical engineer either, but looking at our counterparts in the digital still photography world, I don't understand what the hang-up is with producing lenses that can resolve HD images for small sensors. For example, the Canon S1 IS has a 10X optical zoom lens and resolves images for a 1/2.7" sensor (just slightly larger than a 1/3") at a max resolution of 2048 x 1536. Sure, it's not as fast as an f/1.9 lens but that's why the Fujinon has to be bigger. Other than that, I don't see why it should be a major problem to properly resolve an image at the JVC's resolution. Thoughts anyone?

[Michael Pappas]

Tim, that is an excellent point. Very good!

Just talking about the actual glass and not the mechanical differences.

Originally Posted by Tim N Le

Athe Canon S1 IS has a 10X optical zoom lens and resolves images for a 1/2.7" sensor (just slightly larger than a 1/3") at a max resolution of 2048 x 1536. Sure, it's not as fast as an f/1.9 lens but that's why the Fujinon has to be bigger. Other than that, I don't see why it should be a major problem to properly resolve an image at the JVC's resolution. Thoughts anyone?

[Steven White]

The primary reason for building a large lens is to gather more light. One of the downsides though, is that those outer rays have to be "bent" significantly further than their on-axis counterparts. More bending = more optical path length, and this in turn means that all the "thin lens approximations" you know so well from high-school fall apart. The only lens that's really easy to manufacture is a spherical lens. The more aspherical the lenses get, the more difficult it is to manufacture them, and the more components are required to counteract the aberrations associated with the major components.

When you look at a still camera you have a significantly smaller aperture. The more expensive (i.e., larger) still camera lenses also cost a fortune.

-Steve

[Steve Gibby]

Originally Posted by Graeme Nattress

As far as I can tell (looking back through my R&D notes from working with Panavision) the 2/3" CCD on the HDCAM F900 (I think it was that at the time) has pixels of 5x5 microns, and a corresponding resolution of 100lp/mm. Does anyone have the CCD dimensions of the JVC HD100 1/3" camera? I took a guess based upon the above dimensions of a 2/3" 16:9 sensor and figured out that the pixels on the JVC are going to be around 3.7 microns square. That equates to a resolution of 133lp/mm.

Now, Fujinon don't have any MTF specs for the lenses, and I don't see any on the JVC site, but I did see the specs for the Zeiss DigiPrimes:

rated at 100lp/mm, but with a 90%mtf at 56lpmm. Usually, for the rated resolution they take the 5% mtf point, so I'm guessing that it has no resolution beyond 100lpmm.

So, for a 1/3" 720p camera, we need a lens rated at 133lp/mm, which to me means that..... Well, I've said my piece already, but I think anyone with a little mathematics can draw their own connclusions, but simply put, there's going to be quite a few figures after the $ sign.

For what it's worth, again at a rough guess, the Z1 CCD is about 100lp/mm horizontal but 200lp/mm vertical (again if anyone has exact CCD measurements, I can refine these figures).

Given this, I don't hold much hope of any 1/3" CCD camera not being lens limited for the forseable future, no matter whether it's Sony, Panasonic, or JVC, and whether it's lens comes from Zeiss, Fuginon or Leica.

Graeme

Hey Graeme,

I've been on location shooting for a few days and haven't been able to get the time to contribute to this thread. I think we're all in the same boat in seeking answers to our tech questions - in this case questions on the 16x lens for the HD100, and more broadly, questions about possible resolutions from 1/3' CCD camcorders. In looking for answers I've begun a more constructive track than before. Here's my progress:

Sony HDW F900 2/3" CCDs

I use the F900, so this data was easy to get.

The actual image size (active area) = 9.58mm (H) x 5.39mm (V) with the image diagonal = 11mm.

H Size of individual pixel = H Size of CCD/ Number of Active Image H Pixels =
9.58mm/1920 = 0.005mm = 5 microns

V Size of individual pixel = V Size of CCD/ Number of Active Image Pixels =
5.39mm/1080 = 0.005mm = 5 microns

Your memory was correct (5x5 microns).

JVC HD100 1/3" CCD

I've been totally milking every possible source to see if that data is obtainable. So far, no luck. I found a group in England (www.prohd.co.uk) that ahs a working prototype of the HD100 in their hands. I asked them to measure the CCDs and create an MTF chart for the 16x lens. They said they will at some point in the future. I'm assuming an NDA blocks that for a little while longer. I'll keep looking. Once we have that data it will be pretty straighforward in analyzing the potential of the camera/lens combination.

1/2" SD lenses on the HD100

I've placed inquiries directly to Fujinon and Canon to get the MTF charts for any of the 1/2" SD lenses listed on the JVC System Chart as being usable on the HD100 ( Link: http://www.jvc-victor.co.jp/english/...d100/sys.html). There are three Fujinon lenses listed: S20x6.4B12U, S17x6.6BRM, S14x7.3B12U. There are two Canon lenses listed: YH19x6.7K12U, YH16x7K12U. I own two of these lenses, the Fujinon 20x and 14x. I hold no illusions that the 14x could resolve adequately on the HD100. The 20x is a much better lens though. For some bizarre reason Canon, Fujinon, Nikon, and other lens manufacturers routinely publish MTF charts for their still photography lenses in print and on the Internet - and yet they steadfastly refuse to publish MTF charts for their motion media lenses! I don't hold out much hope for getting MTF charts on the five 1/2" SD lenses. I do have one workaround. I recently had my 20x lens serviced at Focus Optics in L.A. It's possible that they did an MTF chart for the lens while it was in their shop. I'll check. If any of these five 1/2" lenses could resolve a decent image on the HD100 the Fuji 20x or the Canon 19x would be it. It may not be a great HD image, but for some applications it may suffice. The Fujinon engineer did tell me that the 20x has good glass, with some properties similar to the new 16x lens.

If smaller glass elements cost less, and with advances in lens coatings, and good CCD properties, perhaps this could all dance well in unison, and the HD100 could surprise everyone with it's imaging capabilities. Let's keep this thread going and get to the bottom of all this - through teamwork.

[Luis Caffesse]

Could one of you super smart guys give me a link where I might be able to do some reading on MTF measurements, what they mean, and their relation to other lens factors?

This is all new to me...and rather fascinating.
If anyone knows a good resource, it would be really appreciated.

Thanks.

[Tim Le]

Here's one place that explains MTF:

http://www.photodo.com/art/Reso8.shtml

This site also has a large collection of MTF charts for still camera lenses.

[Graeme Nattress]

Steve, that's excellent research. I got ever-so-slightly different figures from Panavision for the CCD dimension (by about 0.01mm different) but they lead to the same figures - 5 microns square and 100lp/mm.

You're right that specs of broadcast lenses are terribly difficult to come by, and this is making this all quite difficult to research. If you have any luck with your sources, then I'd be very happy to see the results. As I mentioned on another forum, I'd be very, very happy to learn that I'm wrong here, as that gives us all the possibility of real HD from affordable 1/3" chips and affordable lenses.

One thing that struck me though, is that lens cost is proportional to the square of the radius or something - bigger lenses cost more. But the "radius" of a lens for a 16:9 chip will be greater than that of the radius for a 4:3 chip, it being more compact, so we need to use bigger lenses on these 1/3" 16:9 HD cameras than we would on SD cameras, and this again adds to the price? Sound reasonable or am I on another planet again?

I really appreciate the constructiveness of this current discussion - I think some really good research and learning is going on now.

Graeme

[Chris Hurd]

Hi Steve,

Originally Posted by Steve Gibby

I found a group in England (www.prohd.co.uk) that ahs a working prototype of the HD100 in their hands.

That outfit is OpTex in the UK and the fellow who's most likely responding on their board is my old friend Simon Beer. That's a great bunch over there. The prototype HD100 they have on hand is what's called an MT sample, or Market Test sample. MT samples are commonly about 80 to 85 percent complete and usually are not entirely fleshed out. I guess my whole point here is that such samples are frequently subject to any variety of changes, so the prototype that Simon has on hand over there should not be considered as a finalized version by any stretch.