16mm Cine Zooms (English y Espanol)

[Cortlan McManus]

I’m sure like the other 20k+ owners of the 1/3” JVC HD cameras, we all love the interchangeable feature of the kit. Yes, the stock lens, effectively given away with purchase is nice, effective and cheap.

There are now more choices with upgrading to other 1/3” lenses, all of which sound really nice. Unfortunately the prices in all but one case are expensive for most indies, and even 3k is nothing to smile about; maybe rental but to put 9-11k into a lens that can only be used on this lens port is a tough swallow.

We are offered lens relay systems. The best and most expensive has its accolades, but almost always require more than natural lighting and adds bulk and weight. I’m not familiar with the more economical versions, but they do definitely add length to the camera. The upcoming HZ-CA13 sounds great, seems to be only geared for primes, and again, the cinematographer will be spending approximately 4k for a lens (or optical relay system) that can only work on a 1/3” bayonet lens port such as the HD100/110/200/250 employ. Like the 1/3” lenses, this investment will not be carried over to your next camera (say SI, Red, or "_____").

Before NAB of this year when the HZ-CA13 was announced, we began a small project trying to get cine zooms to work to a high level on these cameras. We had always heard that you cannot directly attach a cine lens on a prism camera without major problems. Now after many months of shooting and comparing with the stock lens, I enjoy working with the cine zooms A LOT more in all most all cases. If not, swap out the lens and go on with my shooting.

I do come from a S/16mm background so I never did enjoy the position a video zoom requires (i.e., hand on the grip and a 45 degree angle et al). I always use a lens in manual mode so I do not miss at all the auto features. My work is mostly documentary, both hand held and tripod, but also narrative, psa's, and other work for hire.

This was our first attempt with a lens we thought would work.

Hoping it will not be our last.

If there is any interest, I could post some footage after the holiday.

Cortlan McManus
nYc

http://cine.ma/defCine/defCine.html

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Estoy segurisimo que como los otros veinte mil duenos de los JVC 1/3” HD cameras, todos nosotros nos encantamos el hecho que se puede cambiar los lentes. Si, el stock lente no es mal e efectivmente es dado como gracias con la camera, se funciona pero no es perfecto para na’.

Claro, existe opciones en usar otras lentes. Desafornudamente todos son caros para nosotros cineastes. A lo mejor alquilamos para una produccion tal y cual, pero para comprar algo en los 9-11 mil dolares estadounidenses es mucho para comer especilamente cuando solo se puede usar con esta camera (i.e., 1/3” bayo mount).

Existe “lens relay systems,” si, es cierto. El mini35/movietube son carisimos y son los mejores, pero hay que usar luz electrica y daban peso y tamano a la camera. No conozco directamente las mas economicas, pero seguramente da mas tamano a la camera y basicamente elimina la posiblidad de capturar como documentario. La HZ-CA13 parace buenisima pero solo se puede usar lentes primes y otra vez, el cinematografo va a gastar sobre cuatro mil pelas para un lente (o “lens relay system”) que solo se puede usar con la 1/3” bayo mount con la HD100/110/200/250.

Antes del NAB de este ano cuando se adversito la HZ-CA13, nosotros empezabamos un proyecto chiquito intentando usar “cine zooms” para trabajar con exito con esta camera. Desde siempre nos han dicho que no es posible usar cine zooms con una camera de prismo sin muchos problemas. Ahora desde meses de shooting y comparando con el stock lente, me gusta mucho mas con el cine zoom. Esto solo es mi opinion y no en todos casos, pero esto es el punto de tener una camera donde se puede cambiar lentes, no!

Tengo experiencia con el S/16mm o sea, nunca me gusto trabajar con una video zoom como tienes que apretar su mano tal y cual. Siempre uso un lente en manual y para na’ no pido de menos las opciones autos (“I don’t miss the auto functions.”)

Este fue nuestro primer ejercicio con un lente que pensababmos iba a funcionar.

Esperamos que no es la ultima.

Si hay interes podria postar footage despues de nuestros dias de fiesta aca en los EEUU.

(me ha pasado mucho tiempo en escribir en el castellano! Disculpame!)

Cortlan McManus
nYc

http://cine.ma/defCine/defCine.html

[Stephan Ahonen]

Originally Posted by Cortlan McManus

Like the 1/3” lenses, this investment will not be carried over to your next camera (say SI, Red, or "_____").

You can't buy Canon SLR lenses that fit on a Nikon either. As for the price, well, welcome to the world of video. The shooting style demands a high ratio zoom lens which is difficult and expensive to build.

The argument about carrying lenses over to your next camera is really just a legacy of your film background, where lenses are switched out often enoughto be considered separate from the camera. In video, a lens is really considered part of the camera, and the price of the lens is considered part of the cost of the camera, albeit a rather large part, usually at least half to two thirds of the cost of the two of them combined. An $11k lens for a $6k camera is not unusual in broadcast terms, and with the HD250 costing about $9k for the head, I honestly think an $18k lens wouldn't be out of line. Sure you've now spent $27k on the camera package, but we're talking about a picture that rivals cameras costing that amount before you even buy the lens.

Video is expensive! If you thought buying a $6k camera was going to change that you were sadly mistaken, even though $6k is an incredible deal for a camera that looks as good as the HD100.

We had always heard that you cannot directly attach a cine lens on a prism camera without major problems.

What kind of problems? The biggest one I see would be a larger back focus distance on a video camera due to the prism, anything else?

I do come from a S/16mm background so I never did enjoy the position a video zoom requires (i.e., hand on the grip and a 45 degree angle et al).

To be honest, the folks I consider to be the best 16mm shooters in the world, NFL Films, use cameras in exactly this configuration. If it's good enough for them it's good enough for me.

[Miklos Philips]

Yes, would love to see some footage shot with 16mm lenses. Please post.

[Keith Gruchala]

I spoke extensively with optical techs and did research because this also seemed like a great way to get cheaper high quality but the basic consensus came up with several problems:
1) because 16mm is designed without thought of 3 CCD's in mind, the probability of Red, Green, and Blue wavelengths focusing correctly together, especially with the altered rear flange focal distance is unlikely. There will most likely be chromatic aberations.
2) Because the size of the 16mm gate is much larger that 1/3" chip, the required resolution characteristics of the normal design of a 16mm lens are exceeded. In other words, the resolving power designed into a 16mm lens to provide sharpness over a 16mm gate, is not high enough when your only looking at a chunk that is fraction of the size and expecting to resolve high definition from that smaller piece. Some really sweet PL mount super 16mm lenses might actually be up to snuff, but the mounting issues make it necessary to destroy the standard JVC lens mount to work. There might be a few "C" mount prime lenses out there good enough (like the vario switar 10mm) but your basic Angenieux or Canon zooms will be soft.
3) Because the focal length is more than doubled, your standard 16mm zoom which starts at 10 or 12mm is already quite telephoto and therefore only good for telephoto type work. This becomes particularly problematic in narrative work when matching wide and tele shots is a must for seemless scene work and there is no way the look of a 16mm lens is going to match some kind of other wide video lens, and in 16mm the widest lens usually even available is 5.5, which we know in tight location work is sometimes not even wide enough.

All these things together made me feel it was a dead end for modifications.

[Steve Oakley]

I'm the guy who mounted a canon FD still lens onto the HD100. in shooting the lens a bit, its NOT as sharp as the stock lens. wide open it has major CA, but by F4 looks ok. Its a 50mm 1.8. in order to get the lens onto the camera, I had to build a 1/3" lens flange and attach that to the lens because machining an adaptor, while I think possible, would not be very durable. it would be along the lines of 1mm thick - hard to machine given its size & heat issues when machining. I do have a 50mm1.4 I want to try, but getting a flange attached to it may not be readily doable. besides all this, you have to pull the iris mechanism apart, and basically rewire the springs so the iris works as expected. canon FD lenses stop down about 1/2 way when not mounted, and a specific pin is not pushed in. then when mounted, they stay wide open unless another lever is pushed, hence the work on the iris mechanism. great hobby project/proof of concept , but not a great shooting lens.

another issue is that the 35mm lens does not need to resolve as well as a 1/3" lens because its got a lot more surface area to work with on a 35mm neg. I'd expect the very very best 35mm still lens might work ok, but hacking one up is a major proposition.


Steve Oakley

[Bill Ravens]

FWIW, Steve, I use a canon 100-400mm 35mm zoom lens on an xl2. the image is fantastic. no CA, and very good resolution.

[Steve Oakley]

an XL2 in SD requires less performance from a lens than does HD which will reveal flaws far more easily because the circle of confusion is 1/4 or 1/6th the size 720X480 VS 1280X720 is 4 times more pixels, or each pixel is 1/4th the size. so what may work well for SD may not work well for HD. oddly enough, the glass I have might do ok with a 1/2 or 2/3" chip lens and look pretty nice, but the small high res chip really requires a lens to focus light into a much small spot ( circle of confusion ) than does a larger imager, and in particular 35mm which is massively larger and much more forgiving.


Steve Oakley

[Stephan Ahonen]

I'm crunching some numbers here and this is what I've come up with.

The critical number for resolving power is the size of each pixel, so we have to take the resolution of the chip and divide it by the dimensions of the chip. The highest density chip in an EOS camera I could find was the Canon Rebel XTi, with a 10.1 megapixel APS-sized sensor for 175 pixels per millimeter. We'll use this as a maximum for how good a 35mm SLR lens can be expected to perform.

A 16x9 1/3" sensor is 7.64x4.3mm. XL2's chips are anamorphic 960x480 pixels for 125.57 px/mm horizontal by 111.62px/mm vertically, so that means an XL2 is far less demanding than a Rebel XTi for resolving power of the lens. So far so good.

Let's look at HD100's chips now. 1280x720 square pixels mean 167.43 pixels per millimeter. The chips are still just barely less dense than the chip in the XTi, though without taking into account the resolution loss caused by the bayer pattern on DSLR chips. Still, the pixel density is close enough that I think you could pretty much take a 1280x720 crop of a full-sized photo from an XTi and that'd be just about what that lens on an HD100 would look like in terms of resolution. See the attached images, they're crops of photos I found on photozone.de taken through a regular Rebel (the 8 MP, not the 10.1 MP XTi), but which should still give you an idea of what kind of resolution you can expect.

Where it get's really interesting is Canon's own XL H1. With a chip resolution of 1440x1080, we're talking about 188.35px/mm horizontally and 251.1(!) px/mm vertically. Canon's lenses for the H1 have to be even sharper than their 35mm lenses even with a longer zoom ratio and larger range of f-stops, though I'm sure the shorter back focus and smaller image circle help. But this pretty much means you can perish the thought of throwing one of your EOS lenses on an H1.

[Cortlan McManus]

Sorry for the longer delay…

Thanks all for your thoughts.

Keith, we must have been talking to the same people!

I figured why not though, right?

What follows is a mini-report of phase two of this three-phase project.

PHASE 1
As I’m sure many of you know, changing the lens port of the camera is not possible. The port and CCD apparatus are effectively the same part (source: my understanding of service manual). This would have of course facilitated experimenting with all cine lenses (or lenses w/ adapters) but would have made almost impossible the continued use of 1/3” bayo lenses. One might say, hey, but we’ve already got optical relay systems that will work with 35mm cine, 35mm still, and soon, 16mm primes! True (well when started HZ-CA13 had not been announced yet), but an optical relay system is something quite different than direct lens to capture plane (in this case our sensor). This phase has since been scrapped.

PHASE 2
Let me begin, by saying that there is nothing scientific about these shots. No test charts, no scopes, no gyrotrons, no pixel analyzing, no 30” HD monitors. Just shooting on location with mixed light sources, my eyes, the LCD/viewfinder, and some lenses.

LOCATIONS:
(I guess you’ll have to look at the pics!)

CAMERA SETTINGS:
(same for both lenses)

HDV – HD24p

MB: 2
DETAIL: NORMAL
VH: NORMAL
H Freq: MIDDLE
V Freq: HIGH
SKIN: OFF
BLACK: STRETCH 2
WHITE CLIP: 108%
KNEE: MANUAL
LEVEL 80%
MOTION: OFF
CINELIKE: ON
COLOR MATRIX: CINE (all NORMAL)
GAMMA: CINE, LEVEL 1
COLOR GAIN: 4
WHITE PAINT R/B: BOTH NORMAL
SHUTTER: STEP
FAW A

No matte box was used or lens filters. The only ND applied was in camera.

Footage was captured from tape via firewire into FCP. Each still was exported from QT Pro as an uncompressed Pict. These were than imported into iPhoto which compressed and rendered the jpg and web pages. No software color correcting, etc was applied to the shots at all. This was a quick export; there could be some random issues affecting individual stills (i.e., don’t know why some appear slightly smaller than the rest; why are the images slightly smaller than 1280x720?, et al, but nothing serious that detracts from the integrity of this already unscientific experiment ;-).

There are two links of interest:

http://cine.ma/lensCompare/lensCompare.html

-jpg photo gallery of roughly ~ 272kb;
-write down image name of interest;
-if desired, download uncompressed PICT from:

http://cine.ma/lensComparison/

-each of these files is 3.3MB;

Attempting to be objective, I will refrain from expressing my thoughts about the images until others have had a chance to view them.

Best,

Cortlan McManus
nYc

~ red.coop ~

[Miklos Philips]

Hey Cortlan,

Nice stuff and thanks! For me two vital bits of info missing:

1.) Just what are you using to get the 16mm lens onto the HD100 and
2.) What cine lens(es) are you/ were you using?

Thanks,

Some actual motion footage would be nice too... cheers

[Cortlan McManus]

Hi Miklos,

The lens used for this test was the Angenieux 9.5-57mm f1.6.

In the lens comparisons tests therefore, the shots are taken either at ~ 9.5mm or 57mm. The defCine Randoms are shot across range, as are the stills from the zoom tests.

Not the best lens in terms of zoom range (6x) but the fastest 16mm zoom available that I know with the best minimum focus (two feet). Many of us have noticed that the stock zoom loses something beyond 60mm or so. Measuring zooms in terms of T-stop have no relation unless shooting film, which I'm sure you know with your background.

The lens was purchased from eBay. I didn't own any regular 16mm zooms and I also wanted to measure the quality of available used 16mm cine zooms purchased via this avenue. Phase III (Lenses 2-5) is presently on going.

From my research, all 1/2" and 2/3" lens are expensive (i.e., in the many thousands of dollars). The used market for these lenses is much more limited than used 16mm cine zooms. Obviously if you already own these lenses or have free access to them, this is a non-issue.

It was originally mounted with a different mount of course (Cameflex if memory serves me correctly ~ i.e., for Eclair's).

This is not the same as simply using the 1/2" or 2/3" mount adapters.

Instead, a 1/3" bayo mount was machined and this is how the lens attaches to camera (can simply attach and mount as if it were the stock or any other 1/3" bayo lens).

I say not the same as 1/2" or 2/3" because it is also necessary to properly collimate the lens (i.e., cannot simply put custom mount on lens and you're set to shoot: collimating is a requirement with this route).

To overcome this, a custom mount was made for the collimator so that the lens could be properly set for the HD100/etc. This basically means that, this lens is only designed to be used with the 1/3" bayo cameras, it cannot be used with other cameras. This having the benefit of being properly calibrated for this sensor, and the negative being lens can only be used with this camera (i.e., a 1/2", 2/3", stills, or PL's are not taken to lens technician who spends time with lens making it properly focus at the required focal plane of 31mm ~ it seems quite possible that even if off by one micrometer, your focus is out ~ no room to play with slight offsets or improper collimation ~ but 1/2", 2/3"/stills/PL lens can be used with less precision on different cameras as well).

As the cinematographers know, each lens for any cine camera always must be collimated for proper operation. No exceptions. This will also be necessary with SI or R1 as it would with any PL hard-front lens port.

For a firmer understanding of focal lengths on particular formats, please refer to Panavision New Zealand Knowledgebase:

http://www.panavision.co.nz/main/kba...enseqvform.asp

The widest end of the stock (i.e., 5.5mm) has approximately a 48 degree angle of view. A 9.5mm would have an angle of view of approximately 30 degrees. A 60mm would have an angle of view of approximately 4.6 degrees.

For comparisons sake, a 9.5mm on a 1/3" bayo would have the same angle of view as a ~ 60mm on a 35mm still camera, and a 57mm would have the same angle of view as a 380mm. The widest achievable 1/3" bayo is the 3.5mm which would be equal to ~ 23.3mm with an angle of view of ~ 70 degrees.

Lenses used on the HD100/etc might actually have a different angle of view than the Panavision chart will show you. I say this based on Mr. Dashwood's invaluable research showing that the 16:9 1/3" chip of the HD100/etc is actually smaller than your typical 4:3 1/3" chip (i.e., HD100 is 2.739mm x 4.864mm whereas a 4x3 would be 4.8mm x 3.6mm). The full 16mm aperture is 7.5mm x 10.4mm (I say full aperture, and not size of negative).

I am going to shoot some musicians this coming weekend and will post footage in motion at that time. I plan on posting as .m2v direct from cam unless others desire some other format. Democracy will win and I will only be able to post one format.

BTW, I enjoyed many aspects of "Riding With Jack."

All the best,

Cortlan McManus
nYc

~ red.coop ~