Canon reveals its first two 8K broadcast lenses, the UHD-DIGISUPER 51 and 7x10.7 KAS S

yeah like who buy this just to attach to their mirrorless,i mean..come on.

Nov 14, 2019permalink

onlyfreeman

Way overpriced. I'll wait for the Tamron version. ;)

Nov 12, 2019permalink

no,dude..no..NO...NOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

Nov 14, 2019permalink

Digitgio

In our work we use usually UHD lenses for 4K format.
First issue is that the eng lens style for hand held cameras are
heavier then previous HD lenses, second issue is that 4K
cameras with 2/3 sensors are darker than HD cameras.
No way fior a 2/3 8K camera working at 120 FPS.
So new sensors and let’s see about DOF on sport
productions... For the next years 8K will be only for Japan
broadcasting will face the problems very far in future.

Nov 12, 2019permalink

Won’t they just adjust the light levels to give the cameras the fc they require?

Nov 12, 2019permalink

LJ - Eljot

"compatible with 8K broadcast cameras that feature 1.25" sensors"
No word about 2/3". Fractions are not your core competence, or?

Nov 12, 2019permalink

There are no fractions involved. The 2/3" is in reference to vidicon tubes that have been used in decades. The actual sizes are in mm.

Nov 13, 2019permalink

Digitgio

2/3” comes from Plumbicon tubes, The sensitive surface was not
all utilisable because not full flat, only 2/3” was flat and good as sensitive surface. In broadcast cameras is still the reference standard
for the 3 sensors (R-G-B) even for UHD 4K.
Just now for UHD 8K is not enough for good
sensitivity, to have bigger photosite is changing at the new size.
DOF probably will be an issue as already happened
some years ago with first UHD 4K cameras equipped with S35 sensors.

Nov 13, 2019permalink

LJ - Eljot

Still the 1.25" tube would be the larger tube.

Nov 14, 2019permalink

So did each size of tube have a different name? I’m always just heard of them all referred to as vidicon.

Nov 14, 2019permalink

Digitgio

Plumbicon was the top level tube, Vidicon arrived after
as a low cost replacement of it. Vidicon suffers a lot of smearing
and burn in effect. The funny thing today is the fact about
size of broadcast sensors, still the same size in about 50 years.
All for keep using year after year same lenses and
accessory.

Nov 16, 2019permalink

https://www.usa.canon.com/internet/portal/us/home/about/newsroom/press-releases/press-release-details/2019/20191106-8k/20191106-8k

Thanks. My school was the last university in the world to use black and white Sony video cameras that shot on 1/2-inch tape. The equipment managers warned us over and over to NEVER point them at bright lights because they would burn permanently into the tube.

Nov 16, 2019permalink

cdembrey

The 1.25-inch image format has a diagonal of 18.5mm as compared to the smaller 11.0mm of the 2/3-inch format and the larger 28.2mm of the Super 35mm format.

The early experimental 8K UHD live television coverage of sporting events proved critical to determining the image format size that could deliver the Essential Depth of Field ... Paper-thin DOF is NOT wanted by Sports Broadcasters.

Nov 12, 2019permalink

real342

1.25 inch sounds really small, but put into mm it's easier to comprehend. I had no idea these huge lenses were attached to cameras with smaller sensors though. ISO is not an issue because the field/studio is usually well lit?

Nov 12, 2019permalink

How is "paper thin" DOF related to the choice of the format?

Nov 12, 2019permalink

Ed Ingold

Depth of field for the same field of view is inversely proportional to the sensor size (e.g., the diagonal).

Graduating from 1/3" to Super-35 (or FF), I can focus on someone 60' away (@ 600 mm) and the person directly behind them is OOF. Video is typically shot wide open or no more than 2 stops down, using ND filters as needed. Sports may need more DOF, but image quality is secondary.

(Focal length doesn't change DOF, which depends only on the f/stop and absolute magnification at the image plane.)

Nov 12, 2019*permalink

"Depth of field for the same field of view is inversely proportional to the sensor size (e.g., the diagonal)."

No, it is not. There is also a lens and an aperture involved.

Nov 12, 2019permalink

Ed Ingold

Okay, I left out the "same relative aperture" part of the effect of sensor size. I thought that was understood, as well as distance to the subject. If you have a larger sensor, you need a longer focal length for the same field of view. By the time you do all of the algebra, DOF is inversely, linearly proportional to sensor size.

The second part regarding magnification is also true, but most applicable to changing lenses with the same camera, and moving so that the subject is the same size on the image plane. With a longer lens, you move further away, which increases the DOF to exactly offset the effect of focal length. The background will be OOF to the same extent, except there is less of it (narrower FOV).

If these concepts seem counter-intuitive, I suggest you refer to "The Camera" by Ansel Adams. It has lots of pictures ;)

Nov 12, 2019permalink

"1.25 inch sounds really small"
It is four times the area of 2/3" used for many for 4K cameras, and also for many big budget movies a few years back.
E.g. it is the format Sony is using its first 8K broadcast camera: https://pro.sony/ue_US/products/4k-and-hd-camera-systems/uhc-8300

AFAIK, all 8K broadcast cameras so far are either this 1.25" [18.5mm diagonal] or 4/3" [22mm diagonal] format.

"I had no idea these huge lenses were attached to cameras with smaller sensors though."
These lenses are already massive (about 50 pounds/24Kg) and expensive (over $200,000); think how big a lens with this wide a zoom range would be for use with a camera using the 35mm film format?!

"ISO is not an issue because the field/studio is usually well lit?"
This format and even 2/3" format is also used in the field, for sports and such.
https://www.usa.canon.com/internet/portal/us/home/products/list/lenses/broadcast/4k-uhd-field-box-lenses
Big _apertures_ (front elements) is what give low-light speed.

Nov 12, 2019permalink

@Ed Ingold Yes you left out the assumption that the larger format uses a lens of larger effective aperture diameter in proportion to the increase in linear sensor size and focal length, and thus with an aperture area and front elements larger in proportion to sensor area. The usual thing of attributing shallower DOF and better low light/ high speed performance to sensor size instead of lens size.

These broadcast lenses give speed (and low DOF if needed) through big heavy front elements, leading to very low minimum aperture ratios; as low as f/1.15 on some Canon lenses for 2/3” format 4K broadcast cameras.

Nov 12, 2019permalink

@Ed Ingold: Why same aperture?

See also the post above.

Nov 12, 2019permalink

real342

@BJL Very interesting!

Nov 12, 2019permalink

Ziginox

JACS, I see you're showing your ignorance yet again...
The reason sensor size affects DoF is simple. Crop factor and distance. You get the same DoF on, say, a 200mm lens at f/8 at the same distance, but on a crop sensor you're standing farther away for the same framing. Thus, your DoF increases.

Nov 13, 2019permalink

"Ziginox JACS, I see you're showing your ignorance yet again."
No, you are showing your ignorance and lack of reading ability: although DOF can be computed from format size, focal length and aperture ratio, it is clearly not dictated by format alone: once one compares at equal FOV in different formats(adjusting focal length), you get the same DOF when using the same effective aperture diameter (focal length divided by aperture ratio) which very roughly means "equally wide front elements give equal DOF".

It is pointless to assume equal aperture _ratio_ (i.e. smaller aperture sizes with smaller formats) when comparing DOF; its is just a round-about way of saying that smaller apertures give more DOF.

BTW, these broadcast "box lenses" have huge effective aperture diameters of about 200mm!

P. S. Why do you assume a different subject distance with a different format?! Clearly these broadcast cameras will be used at the same subject distance regardless of sensor size!

Nov 13, 2019*permalink

@Ziginox: Read the post above.

Your thought process is so messy - I counted more logical errors/inconsistencies than the number of sentences in your post.

Nov 13, 2019permalink

I bet the more expensive one will be around $250,000 and the cheaper will be $175,000.
But what on earth is a 1.25-inch sensor camera? I thought those were only for astro.

Nov 12, 2019*permalink

well....no doubt but it's kinda small,actually...it's like the size of aps-c sensor,i think? with that "small" and the ability to record 8k movie,no s***,thats the best thing i have ever seen.

Nov 12, 2019permalink

GulliNL

For 8k you don't need a FF sensor because that would mean you have to downscale to the correct resolution. It is more a matter of processing power to have 120fps @ 8K resolution

Nov 12, 2019permalink

Ziginox

GuilliNL, it's possible to make a full frame sensor with exactly the right number of pixels for 8K footage. Resolution isn't entirely tied to sensor size.

Nov 13, 2019permalink

Photoman

Can you adapter to mirrorless???

Nov 11, 2019permalink

All camera for these lenses are mirrorless! Here's one from Sony (they seem to share a new "1.25-type" lens mount:
https://pro.sony/ue_US/products/4k-and-hd-camera-systems/uhc-8300

Nov 12, 2019permalink

Only if the image circle is large enough to cover a m4/3rds sensor.

Nov 12, 2019permalink

how about pentax q?

Nov 12, 2019permalink

LJ - Eljot

Finally a lens that could give anough resolution to make use of the 12MP on that tiny 1/2.3″ sensor.
( I know the newer ones have 12MP on a 1/1.7" one)

Nov 12, 2019permalink

I literally was mainly interested in the price.

Nov 11, 2019permalink

analoggrotto

Amazing how much better this lens will perform on a Sony A7S

Nov 11, 2019permalink

More likely on a Sony UHC-8300; see post above:
https://pro.sony/ue_US/products/4k-and-hd-camera-systems/uhc-8300

Canon has not yet announced a camera than can use this lens, so maybe it is going for the 8K 1.25" format lens market, at least for now.

Nov 12, 2019permalink

It will produce beautiful 2.2mp images.

Nov 12, 2019permalink

Ed Ingold

An A7S? Yeah, right!

I have a Sony 28-135 PZ, which is intended for video use. It weighs over 2 pounds and is 4" in diameter at the objective (95 mm filter ring). I use it with an A7iii, but in a rig, preferably on a tripod.

8K images are about 28 MP, not 2.2 MP.

Nov 12, 2019permalink

"8K images are about 28 MP, not 2.2 MP."

Hence my remark.

Nov 12, 2019permalink

fredmason

Adding that to my Christmas wish list. Would love one of those for my Canon Digital Rebel.

Nov 11, 2019*permalink

Battersea

I hate having to use adapters though. ;-)

Nov 12, 2019permalink

It seems that there is a new standard "1.25-type" lens mount for 8K cameras, with Canon making these lenses partly for use on other brands of camera (like Sony and JVC?) and Sony already making 8K 1.25-type cameras advertised as usable with lenses from other makers (like Canon?)

Does anyone know more? The motion picture world seems more amenable to standardized lens mounts than still camera makers.

Nov 11, 2019permalink

You can purchase the spec detail here- https://www.arib.or.jp/english/std_tr/broadcasting/desc/tr-b37.html (For about 3000 yen)

Nov 11, 2019permalink

Chris Flowers

It's has been that way since ever. Broadcasting world is just for a few players, on the camera side is JVC, Sony, Panasonic, and even Hitachi. From the lens side is mostly Canon and Fuji.

Nov 12, 2019permalink

Thanks Chris, you made me look around and see how Fujinon and Canon lenses dominate broadcast video, with standard mounts like B4. Sort of like the PL mount that dominates in cine-cameras.

Wouldn’t it be nice to have one open standard mirrorless mount for each still camera format size?

Nov 13, 2019permalink

The difference it traditionally most Pro video (and film) gear is rented from facilities companies, so interchange and consistency of controls matters more. Also the upgrade cycle is slower because the gear is costed and built to last and conform to broadcast standards (which change slowly too.)

Nov 13, 2019permalink

@AshMills, yes, and there is also a natural role for an industry standards body, the Broadcasting Technology Association (and it successor organization, the Association of Radio Industries and Businesses) where things are agreed; for example the adoption of the B4 mount as a standard for 2/3" broadcast cameras: https://en.wikipedia.org/wiki/B4-mount

Nov 13, 2019*permalink

nickolas84

With just over 1 inch sensor and constant f/2.8 aperture the big 51X zoom will be far from the diffraction limit for those worried.

Awesome stuff!

...and another reminder that 95% of the time what matters is the lens, not the size of the sensor.

Nov 11, 2019permalink

(unknown member)

It's impossible for either of those lenses to "resolve" 8K with their F/#s.

Nov 11, 2019permalink

What are the minimum f-stops? I do not see that information in either this article or the Canon page it links too. It would need to be about f/4 to have little visible diffraction softening (not the nonsense about needing pixel pitch = Airy disk dimeter).

But 8K seems to be mostly about oversampling, maybe for most flexible subsequent processing; we cannot resolve 8K when looking at a video screen from even the minimum comfortable range of about 1.5 picture heights.

P. S. If @nickolas84 is right about constant f/2.8 there will be no problems with diffraction in practice!

P. P. S. This 66X cousin for 4K in 2/3" format is f/1.7 over most of its zoom range, f/3 at the long end: https://www.usa.canon.com/internet/portal/us/home/products/details/lenses/broadcast/4k-uhd-field-box-lenses/uhd-digisuper-66

Nov 11, 2019*permalink

Kharan

What would be their theoretical maximum resolution then?

Nov 11, 2019permalink

The crop factor is about 2.32. 8K has about 32mp. The 90D/6D2 have similar number of pixels with a crop factor 1.6, and they still show clear aliasing even at f/9 equivalent.

Nov 11, 2019*permalink

@ J A C S Maybe, but there is no need to mess with crop factors and "effective aperture ratios": what matters is the ratio between pixel pitch (which limits sensor resolution) and aperture ratio (which effects lens resolution). The obsessive standard of "Airy disc no wider than a pixel" gives a threshold of roughy f-stop equal to the pixel pitch in microns; actual experience shows that diffraction only starts to affect resolution noticeably between 2 and three times this; to be cautions, when aperture ratio is twice the pixel pitch.

(And why mention f/9? This lens will surely be faster than that, and this format can get adequate DOF at lower aperture ratio than that.)

With the slightly over 2 micron pixel pitch of this sensor, diffraction comes into sight between about f/4 and f/6. Judging by previous lenses in this product line, the minimum f-stop is likely to be under f/2 over much of the zoom range, and no worse that about f/3 or f/4 at the long end — if so, no worries.

Nov 12, 2019permalink

Crop factors is one way to think about it, nothing wrong with that.

F/9 is the eq. aperture used by dpreview in the test scene.

I did not follow your calculations. It is a fact that the 90D aliases at least with two converters (Canon JPEG and Adobe). Increasing the pixel density by 50% or so linearly (which is what this video sensor does) could take it close to optimal density or not - but it is no way overkill. Unless this lens is really soft as you may expect a superzoom to be.

Nov 12, 2019*permalink

Do you mean at f/2.8? Or ever? I was reading about a company that shoots a lot of American football in 4K on 2/3-inch cameras and their engineers found they had to stop down to f/5.6 to get satisfactorily sharp images. Their previous standard with high def was f/4 iirc.

Nov 12, 2019permalink

@ J A C S, I did not give the calculations, but here are some.

The diameter of the Airy disk is 2.44(wavelength)*(aperture ratio) which for visible light is about 1.3(aperture ratio) for mid-spectrum green light.
So the Airy disk covers a square pixel very roughly when aperture ratio = pixel size in microns.

A 1.25-type 8K sensor has 7680x4320 pixels and 18.5mm diagonal, giving pixel width and height 2.1 microns, pixel diagonal 3 microns. The “Airy disk same size as pixel” _guideline_ for diffraction starting to affect resolution gives about f/2 to f/3.

The rest was observations from numerous professional photographers that they do not see much diffraction softening until the aperture ratio is a bit beyond twice the pixel pitch, which is f/4 in this case. BTW, that is roughly when the Airy disk is as wide as a “quad” of Bayer CFA pixels,
G B
R G

Nov 12, 2019permalink

Leonp

"I want it".

(Not a clue why I want it, bit still...)

Nov 11, 2019permalink

Stig Nygaard

Because of the model name?

Nov 11, 2019permalink

otto k

Yup, it just rolls off your tongue...

Nov 11, 2019permalink

Horshack

I always make it a point to stop by Canon's booth at NAB - every year they have a raised platformed area with all of their broadcast cameras and lenses open to attendees to try out. There aren't many other opportunities to walk up and play with a $200k lens.

Nov 11, 2019permalink

Battersea

They do an awesome job at the Albuquerque Balloon Fiesta also. They have cameras and lenses available for anyone to borrow for the day. They clean, update and tune up your existing camera and lens while you use theirs pretty cool.

Nov 12, 2019permalink

Interesting. There must be a lot of electronics in the square cross section housing. It looks like it has rubber feet in the front. Tripod socket must be underneath somewhere.

Nov 11, 2019permalink

You tend to build a camera pedestal around one of these, https://www.thebroadcastbridge.com/content/entry/2833/i-movix-delivers-10x-super-slo-mo-for-the-2015-paris-roubaix-race

Nov 11, 2019permalink

AshMills: Holy smokes, I didn't know they were THAT big!

Nov 11, 2019permalink

paulfulper

I wish I could use it for birding , but then again it probably doesn't do still photography......and it must cost $90.000

Nov 11, 2019permalink

Mike Evangelist

More like $200,000.

Nov 11, 2019permalink

Ninety bucks? I'll take two! ;-)

(pesky comma/decimal issue...)

Nov 11, 2019permalink

Roland Karlsson

Yes the decimal comma vs point issue is irritating. And this is the only one where I reluctantly agree that US is right. We German influenced countries will change to decimal point if you remove all of your bad habits, like moving fork back and forth while eating and calling milliard a billion. And please? SI-units? And what about your weird way of writing dates? Today is 2019-11-13! Most significant digit to least significant digit. How hard can it be?

Nov 13, 2019permalink

Roland, I agree on most of your points. Fork/knife is a preference. (I don't switch hands, but most Americans do, I think) I think people can do it however they want without harming anyone else.

SI units? Agreed. (including date format) I've never heard 'milliard'. That might be a language issue. 'Billion' has always meant x10^9 here.

Nov 13, 2019permalink

https://en.wikipedia.org/wiki/Long_and_short_scales

Actually, we should go with the people who invented the zero and start using crore (10 million) and lakh (10,000).

Nov 13, 2019*permalink

Roland Karlsson

The two naming conventions for e.g. billion are long and short scale.

There are maps there telling who uses which, e.g. US uses short scale and Europe long scale. I thought Britain used long scale, but here it says the opposite. Anyhow ...

Personally I prefer long scale, but the most important thing is to use the same scale. Today you do not know what is meant be a billion.

Nov 13, 2019permalink

Camera Conspiracies

I'm the guy holding Canon's beer. They did it. I drank your beer, sorry.

Nov 11, 2019permalink

GinoSVK

Did you hear that recent rumor about cripple hammer mark II?

Nov 11, 2019permalink

IamToddNorris

When will we see a squirrel video with this lens? 😉

Nov 11, 2019permalink

Sony Influencer

I hope it is good enough for my cat videos.

Nov 11, 2019permalink

ewelch

Your cat will need to be very interesting. Way more interesting than the average cat, man.

Nov 11, 2019permalink

Androole

Interesting, I hadn't heard of 1.25" sensors before.

From another article on this, it looks like they have an 18.5mm diagonal, meaning that even if it is 16:9 or 17:9 aspect ratio, it still has a smaller long-edge dimension than a 4/3" sensor.

That's a promising development if the ultra-expensive 8K cameras that are going to be used for high end commercial TV applications are actually smaller sensors than what Panasonic could end up putting into the next GH...

Nov 11, 2019permalink