Panasonic unveils 4K sensor

[Bruce Watson]

If so, then why would Panasonic bother making a 2/3" 4K chip, if it will never be an optimal sized sensor?

Depends on how you define "optimal" doesn't it?

I suspect they'll work on 2/3" chips because of the huge acceptance of the 2/3" equipment in the broadcast industry. If they can make 4k work acceptably in a 2/3" chip format, they can more easily maintain their user base, and their vendor base. IOW, all the 2/3" lenses still work. Cameras still shoulder mount as before, operators still uses them as before, etc.

If they try to go to a bigger chip size, they also have to make everything else bigger, especially lenses. Which makes the lenses cost more, makes the cameras cost more, make the cameras heavier to support the heavier lenses, and pulls the weight balance toward the front. None of these things will win much favor with the on-the-shoulder users.

Then you factor in the decrease in DOF as they go to bigger chips, and the ENG crowd might start to rebel outright.

So I can see why they might just want to find a way to cram 4k resolution into the 2/3" chip form factor.

 

[vanvideo]

Personally, I think there will be a big market for a 2/3" 4K camera. As has been stated, the 2/3" platform already exists in large numbers. Panasonic could be capitalizing on a niche the other camera makers have yet to realize. This could be their "baby." Obviously, Panny has invested in this 2/3" chip, so they must believe the market is there.

My question of the validity of a 2/3" 4K camera was more rhetorical, as it would seem Panny has found a way to make it work. Perhaps not the optimal chip size compared to an S35 sensor, but one that can certainly deliver satisfactory images.

Or, the chip could one day quietly disappear, who knows?

 

[harddrive]

Depends on how you define "optimal" doesn't it?

Very much so, I was following on from the article you referenced which was primarily using the term in connection with quality and optical matters. For any given resolution, then over a certain range a doubling of sensor size gives big benefits - up to a certain point, then it sort of plateaus off. In that context, that's what's considered optimal. Factor size, legacy lenses etc into the equation and yes, the definition of "optimum" may change.

If they try to go to a bigger chip size, they also have to make everything else bigger, especially lenses. Which makes the lenses cost more, makes the cameras cost more, make the cameras heavier to support the heavier lenses, and pulls the weight balance toward the front. None of these things will win much favor with the on-the-shoulder users.

Then you factor in the decrease in DOF as they go to bigger chips, and the ENG crowd might start to rebel outright.

For given zoom ranges, the lenses will only be bigger, heavier etc if you have to maintain the max f stop of current 2/3" lenses (typically around f2), and that's not necessarily the case. There's another Roger Clark article which deals with this very well regarding still cameras - http://www.clarkvision.com/articles/dof_myth/

In this context, then it's feasible to imagine designing long zoom range lenses for s35 with smaller f stops than 2/3" users are used to - and giving size/weight, dof/lowlight performance equivalent to current 2/3" lenses and cameras. The smaller f stops are compensated for in low light terms by the higher base ISO figures of bigger sensors. Obvious question then is that if the performance is then going to be similar to 2/3", why bother with the large sensor!?

Answer is that with large sensor you have the option, it gives versatility. Either use the long range/small aperture lenses and get 2/3" performance OR use primes or short range zooms with wider f stops when low light/shallow depth of field is paramount. Use a 4K 2/3" chip and you don't have the latter option.

It may not even be necessary to consider getting all new lenses, if you can use existing high quality 2/3" glass on 4K s35 cameras via an adaptor. If you read the description of Sony's trials next year for FIFA at the World Cup in Brazil, using 4K cameras for live football, they cover exactly these points - http://www.tvbeurope.com/main-content/full/sony-details-4k-tests-in-brazil#.Ug-HpaNwZ9A :

One of the key things under test are the lenses. These include a prototype lens convertor that permits standard 2/3-inch box lenses to be used with the F55's Super 35mm sensor; and new 4K Fujinon Cabrio lenses, including the Premier PL 14-28mm Cabrio wide angle.

“The question becomes - if it works well and people like the price point [£22,000] and the director likes the images it creates - then do we need specific 2/3-inch cameras?” asks Grinyer. “The cost of building a new 4K three-chip camera is a serious discussion point."

 

[Bruce Watson]

In this context, then it's feasible to imagine designing long zoom range lenses for s35 with smaller f stops than 2/3" users are used to - and giving size/weight, dof/lowlight performance equivalent to current 2/3" lenses and cameras.

Your missing a key point. As the size of the sensor goes up, so does the focal length of the lens to provide the equivalent angle-of-view. And the lens also has to increase in size to provide the increased coverage required by the larger sensor. Even if you go to a smaller f/stop, you're lenses still get bigger. It's not negligible either.

Bigger sensors require bigger lenses to get similar performance. The Laws of Physics will not be denied.

 

[harddrive]

But think of the prototype lens convertor described in the TVB Europe link. What it basically must do is take the output light from the rear of the lens and spread it to cover about 8x the area (the difference between 2/3" and s35), whilst maintaining resolution.

It's effectively trebling the focal length for not that much increase in physical size. This is no different in principle to the telephoto principles that have been used in lens design for a long time, in proper usage "telephoto" really means a lens whose actual length is shorter than it's focal length. Principles are given at http://en.wikipedia.org/wiki/Telephoto_lens Some of the principles are employed anyway in most modern TV zoom lenses, there's rarely now a direct relationship between actual length and focal length.

The optical principles are similar to using the doubler on a normal broadcast lens. Engage, and it has the effect of doubling the focal length, but obviously the dimensions of the lens don't change.

So where's the catch? You are quite right in that the Laws of Physics can't be denied.

It's simply that although the physical size of the iris is unaltered (as with the length of the lens) then the effective f stop changes together with the focal length. If it's f2 when used for normal 2/3" use, it will appear to be f5.6 when adapted to s35 for both dof and exposure.

 

[dracalat]

Such adaptors already exist:

http://www.ibe-optics.com/HDx35UMS_MARKII

Here's a video shot with the F55 and said adaptor:

http://vimeo.com/71675069

The Sony adaptor will be exactly the same, but it will have the appropriate electronics to allow B4 lenses 'talk' to the camera. I also expect it to be a bit more expensive than the IBE (which is not cheap). These adaptors are great, but they do take away 2 1/2 stops (not a big problem thanks to the high sensitivity of S35 sensors) and the pictures can appear a bit soft around the edges when wide open. I still think it's a great solution when maximum coverage and reach are required.

 

[Samuel H]

There's more to lens size than all this. For example: focal flange distance matters too. I was truly amazed the first time I saw a set of Leica primes for the Leica-M mount. Those are some very sharp, very fast primes, and they cover a fullframe sensor. And they are TINY!!

 

[Bruce Watson]

But think of the prototype lens convertor described in the TVB Europe link. What it basically must do is take the output light from the rear of the lens and spread it to cover about 8x the area (the difference between 2/3" and s35), whilst maintaining resolution.

Forgive my skepticism, but I doubt they can "maintain resolution" across the entire coverage area with an adapter. All practical lens designs experience brightness and resolution fall off in the corners, how much depends on the design of the lens, but it's always greater than zero. Adding an adapter will be additive -- it won't be able to maintain the existing fall off, but will instead add to it.

A 4k sensor is going to be really unforgiving of this. If it's there, you'll see it. Maybe if we keep the camera constantly moving... but then, what would be the point of 4k?

 

[harddrive]

Your scepticism is understandable, and probably well founded! No, I don't like adaptors on principle though how good or bad they are I suspect depends a lot on how much they cost! But it sounds as if dracalat has first hand experience.....?

I really mentioned them more to prove the principle that a purpose designed long zoom range for s35 need not be much heavier, bigger or more expensive than the comparable range and quality for 2/3" - if you're prepared to allow for smaller f stops and let the higher sensitivity of s35 compensate. It's the comment about "If they try to go to a bigger chip size, they also have to make everything else bigger, especially lenses" that I can't agree with without qualification. It's not true if you're prepared to accept smaller apertures.

Very simply, a zoom lens consists of two basic parts - the front system of lenses which primarily determines angle of view, and a rear section that focusses the output of that on to the sensor. In essence, you could largely keep the front system from a current 2/3" lens, and redesign the rear to make the coverage suitable for s35. Compared to the original, that would seemingly keep the angles of view and zoom range the same and give full s35 coverage - whilst having little effect on overall size/weight. And because you're replacing one back section with another (rather than adding extra glass at the rear end) the quality should be comparable with the lens with original back end, as designed for 2/3".

OK, that's a gross oversimplification of a zoom lens optics, but the core point is valid.

 

[Samuel H]

There's a chance they can *nearly* maintain resolution in this case, because resolution is the same in both sensors, big and small. It would be impossible if the bigger sensor had more pixels (which is what happens when you put a 2x extender on a long lens: you get a central crop from a hypotetical sensor that is twice as wide and has 4 times as many pixels)

 

[ripupthehwy]

Personally, I think there will be a big market for a 2/3" 4K camera. As has been stated, the 2/3" platform already exists in large numbers.

The Television Station I freelance for just purchased all new HD cameras (with 2/3" sensors), over $30k each. And new 2/3" lenses for them that cost $10k each. All the control rooms are being gutted and rebuild with HD gear. No way are they gonna have even the slightest interest in bigger sensor cameras, much less even 4 K in the broadcast world for years to come.