Saw the Panavision Genesis

[David Mullen ASC]

Finally found some time to go over to Panavision and take a look at the Genesis HD camera. Bob Harvey arranged a screening for me of the Daviau tests plus some tests in Paris that Salvatore Totino shot for Ron Howard's upcoming "The DaVinci Code."

 

Dean Semler was over there testing the camera for an upcoming Adam Sandler comedy called "Click". Heard that he shot an F900 vs. 35mm test for Rob Cohen's "Stealth" but felt that the HD was not adequate and they went with 35mm. But clearly he feels that the Genesis stands up pretty well.

 

The Daviau test was pretty impressive. The 35mm footage went through an all-4K D.I. while the Genesis footage was posted at 2K and then uprezzed to 4K for film out. It was a surprisingly close match, even in full sunlight. I was able to guess correctly which shots were Genesis and which were film but barely, it was so close. It was only in scenes of intense backlight where a certain SUBTLE edginess, clippiness to the halo in the backlit hair was more apparent in the Genesis version. There were other minor differences.

 

The Totino test for "The DaVinci Code" was more informative in some ways because Totino shot it in available light in the Louvre Museum during the day (but in dim conditions) and on the streets of Paris at night, both cameras shot at the same speed rating (500 ASA for the Louvre, push one stop for the night exterior on both cameras) I believe they were using 5218 in Super-35 and did a 2K D.I. for the blow-up to anamorphic; the Genesis footage was also cropped & stretched to anamorphic.

 

One difference between the two versions becomes apparent because the Primo lenses on the Genesis seem apparently sharper than the Cooke S4 lenses on Totino's personal Arri-435 (the lenses are his too.) They flare less from passing streetlamps and car headlamps -- with the Cookes, there is a subtle fuzzy red edge to the halations, while with the the Primos, there is less halation and it stays neutral. But Totino is still such a fan of the Cooke S4's anyway that he'd rather not shoot without them, meaning he'd rather not use the Genesis, which can't take them.

 

But to my eyes, the test showed the Genesis somewhat outperforming the Super-35 film in dim lighting, where the low-contrast nature of the light in the Louvre caused a somewhat soft, murky, graininess to the film version while the Genesis looked cleaner and even sharper, clearer. The same for the night stuff. Exposure latitude was about the same, which film having the slight edge for holding very bright detail, but the Genesis footage was cleaner looking, less muddy. Again, perhaps one factor is the Primo versus Cooke S4's, I don't know. It was an interesting test because in some ways, Totino was trying to show how film was superior but it didn't quite end up looking that way; I heard a rumor that Ron Howard is pushing him to at least shoot the Louvre scenes with the Genesis since it seemed to handle that diffused, dim lighting better. But that might not happen; it sounds like they are going to shoot the movie in Super-35 anyway using Totino's Cookes.

 

Another reason why perhaps the Genesis held up well in this test was the cropping & stretching to anamorphic. Although the crop & stretch was done digitally in both cases, I feel that when you resize film, you leave clues as to what the original film size was by the size of the grains, whereas digital images can be resized in limited ways more seamlessly in my opinion. It becomes less of an issue if you are using fine-grained film, but when using fast film, the grain is more visible.

[drew_town]

Interesting summary. Thanks for the input. Were you able to put your finger on what gives the Genesis an advantage over current HD models like the 900? Something to look for in future cameras? I'm curious to see the new Superman movie.

[David Mullen ASC]

Interesting summary. Thanks for the input. Were you able to put your finger on what gives the Genesis an advantage over current HD models like the 900?

<{POST_SNAPBACK}>

 

Well, you've got a couple of things at play, just as with the Arri D20:

 

HDCAM-SR 4:4:4 HD recording, which is much better than HDCAM

 

Single 35mm-sized sensor, so 35mm-type depth of field characteristics / standard 35mm optics used

 

There seems to be an improvement in exposure latitude but it's less clear how that was managed; the larger "wells" of the pixels in the 35mm-sized sensor may be helping; less image processing so more information going to tape, etc. But it seems that cameras like the Genesis, Dalsa, and Arri-D20 do have slightly more exposure latitude than the F950 or Viper even when all are going to the same recording format, HDCAM-SR, so perhaps there is something to making the sensors larger, I don't know.

 

Film still has even more exposure latitude in the brightest areas of the frame and still "burns out" to white more gracefully. Actually what the Genesis, Dalsa, and Arri-D20 images remind me of more than anything are good digital still camera images.

 

The problem is that even though we are getting closer and closer to parity with 35mm with each generation of digital movie cameras, it's obviously not enough to simply achieve parity... if it's just as expensive, or more expensive, less convenient, etc. We'd have to achieve parity and THEN become cheaper and more convenient than film for it to become a real threat to film, which is still many years away because of the sheer volume of data involved in improving the quality to match 35mm.

 

The rental on the Genesis is high enough that it only makes economic sense to use it instead of 35mm if you have a really high shooting ratio, where the cost difference between tape stock versus film really start to become the dominant factor. But most movies with budgets that high as to allow really high shooting ratios generally can afford the film costs anyway.

[Phil Rhodes]

Hi,

 

> 500 ASA for the Louvre, push one stop for the night exterior on both cameras

 

That would be +6dB of gain on the Genesis.

 

> the larger "wells" of the pixels in the 35mm-sized sensor may be helping

 

That's pretty much the only way of improving things as regards the makeup of the sensor itself, unless you're going to do a two-pass approach or multiple ND-masked pixels with the attendant issues of that. Big wells is how things like Northlight film scanners can exist, and not destroy the dynamic range; it's also one reason why you can't do that at 24fps, because there are inherent time penalties to big wells. That said, the dual goals of big sensor for low DOF and big sensor for big wells do complement each other; it's just quite difficult to make a semiconductor device that's an inch across with zero defects. I suspect that ever Genesis (and every Kinetta and every D20 and most certainly every F950) will have duff pixels they have to mask, perhaps to a greater extent that some cheap off the shelf handycam, just because of the enormous substrate you're asking for. Not that this is unacceptable so long as you don't get several in a row.

 

I would be interested to know what processing is going on in the Genesis - is it a traditional video camera with the usual controls and DSP, or is it more like a Viper, outputting an essentially unviewable image for later manipulation? I would consider the latter to be more desirable, but probably incompatible with even the smallest amount of compression given the amount you have to screw it around to get a reasonable image. In an ideal world you'd be offered both!

 

Phil

[David Mullen ASC]

The camera processes the signal like the Viper in HD Stream mode, not raw like in Film Stream, but even though you can turn on a color matrix, etc. Panavision recommends sending a fairly minimally processed signal to tape.

 

They have a box you can attach to the HDSDI signal going to the monitor with color looks that can be applied to the displayed signal so that the director doesn't have to see this flat, washed-out image.

[Max Jacoby]

What I don't understand is why all these tests involving the Genesis always have to pass the film through a DI, even if it is 4K. Although that is certainly instructive, I would it find much more interesting to compare the Genesis footage to a print taken straight from the original 35mm neg.

 

Why Panavision don't offer the Genesis with a PL mount is beyond me. Seems like a way to make people shoot on their Primo lenses. They seem to forget that there are Dops, like in Totino's case, who prefer to use different lenses. Especilly now that the Zeiss Master Primes are available, since their combination of speed (T1.3) and sharpness is superior to the Primos.

[Michael Struthers]

Great stuff David, thanks.

 

35mm will be hanging in there for a while methinks.

[David Mullen ASC]

What I don't understand is why all these tests involving the Genesis always have to pass the film through a DI, even if it is 4K. Although that is certainly instructive, I would it find much more interesting to compare the Genesis footage to a print taken straight from the original 35mm neg.

 

Why Panavision don't offer the Genesis with a PL mount is beyond me. Seems like a way to make people shoot on their Primo lenses. They seem to forget that there are Dops, like in Totino's case, who prefer to use different lenses. Especilly now that the Zeiss Master Primes are available, since their combination of speed (T1.3) and sharpness is superior to the Primos.

<{POST_SNAPBACK}>

 

Well, they don't offer PL mounts on any of their cameras as far as I know, but I agree that a PL mount should be an option these days. I would have disagreed a few years ago, but I'm one of those people dying to try the new Zeiss Master Primes and it seems silly that I'd have to go to an Arri camera house just for that reason.

 

At least they should consider buying sets of Master Primes and converting them to a Panavision mount, since they have no Primo equivalent.

 

As for the D.I., well, that's almost inevitable, otherwise you'd be trying to digitally time the color of the Genesis to match a contact print off of the 35mm version and then cutting neg, etc. Especially when you factor in someone shooting in Super-35 and planning a digital blow-up to anamorphic anyway, as in the case with "The DaVinci Code". Any mismatches in color would be distracting -- you'd have some idiot saying "film has better blue colors" or something caused by mismatches in color-correction.

 

But I'm sure one could shoot a test -- Panavision is not trying to kill the 35mm format by any means; they have a LOT more of those cameras to rent. Same with Arri's motivations in building the D20. They aren't "anti-film" people.

[Kevin Zanit]

I would not be surprised to see either a rehoused or just "Panavisied" version of the Master Primes down the road.

 

If you are on a big enough show, I would bet Panavision (specifically Dan Sasaki) would adapt a set of master primes with PV mounts. You may have to get the lenses from a different rental house, but Panavision tends to really jump through hoops as far as accommodating special requests.

 

 

Kevin Zanit

[Max Jacoby]

The problem with the PV mount is that the flange depth is longer than that of the PL mount, so one cannot really rehouse the lenses. They would have to change the mount on the camera instead.

 

As for the comparison tests, even if one did a DI, one could still cut the regular neg and print it. It would be interesting to see, because as good as DIs have become, they still are not 'invisible', meaning they still change the texture of the filmstock.

[Elhanan Matos]

It shouldn't be a problem changing the mount on the camera. We have a High speed HD camera (Cinespeed cam) that comes standard with a PL mount. One day we had a shoot for cadillac ( Cadillac -Mimic- ) The DP was Darius Wolski, everything except for our Cinespeed came out of panavision. During the prep at Panavision, Dan Sasaki came by and took a look at the camera and said that he can easily change the mount, it took about an hour or two for them to machine a new one for us, now we can easily change the mount whenever we want to.

 

I don't see why you couldn't just take the Genesis over to another rental house, clairmont for example, and ask them to put a PL mount in place of the PV mount.

[Kevin Zanit]

I don't know the physics behind it, but I know you can rehouse a PL mount lens, and I know Panavision has. The "Z" series Panavision lenses are just rehoused Zeiss lenses.

 

Panavision would be unlikely to change the mount on their cameras because that defeats the whole proprietary nature of their system. A Panaflex with a PL mount, you will never see it.

 

The difference between them changing the mount on your camera and just taking a camera to a different rental house for modifications was that you were there, and said it was okay. I doubt Panavision would agree to Clairmont even touching the camera, let alone take it apart. In fact, doing something like that without permission on such a proprietary/ secretive system is borderline cooperate espionage, not something worth being a part of. Thus I doubt Clairmont would even do it if you asked.

 

 

Kevin Zanit

[Elhanan Matos]

I don't see how modifying a camera to work with different lenses is corporate espionage, that's what Panavision does with all the equipment that they purchase. If Clairmont pulled the PV mount out and slapped CLAIRMONT or ARRI on it then thats something else, still not corporate espionage though.

[Kevin Zanit]

The Genesis, despite its origins (I know, I know they just gutted Sony gear, it?s an old imager, etc, etc, blaw, blaw) is in fact a proprietary camera that Panavision created (as far as they are concerned, so don't yell at me). It is a camera they are being very secretive about, and it is not released in the open market. You can buy an Arri camera and do what you want.

 

Panavision could construe their competitors taking apart their "top secret" item, which was designed to give them an edge in the marketplace as corporate espionage.

 

It is one thing to own the product you are taking apart, to see how it works, it is another thing to rent someone?s not-for sale camera and take it apart for your future gain. In fact, after the "Digital Millennium Copyright Act" was passed, taking a part something that is yours to see how it works to then change and improve on it (aka Reverse Engineering, aka, something that is vital to the development of technology, and was stupidly hindered by our government with this act) is now illegal/ a copyright violation (if you had to go around *any* sort of device guarding that information).

 

I just don't think that a company would take the risk and start problems with another company, let alone one as big as Panavision. It is not worth it.

 

I don't want this to turn into a "why Panavision is evil" discussion, so with all that said, I am not forming an opinion of Panavision (good or bad). I am simply stating the way it is, and the way things would most likely unfold.

 

(We can argue more about this at dinner tonight if you want . . . ;) )

 

Kevin Zanit

Edited May 18, 2005 by Kevin_Zanit

[Max Jacoby]

I don't know the physics behind it, but I know you can rehouse a PL mount lens, and I know Panavision has.  The "Z" series Panavision lenses are just rehoused Zeiss lenses.

It was always my thought that the 'Z' Series although Zeiss glass, was based on Panavision's design. If you look at some of the focal lenghts available (i.e a 14mm, a 24mm, a 29mm), there is no corresponding PL mount Zeiss lens on the market.

 

The flange depth for the PL mount is shorter than that of the PV Mount, and it is the same for lenses as well. That means that a PL mount lens put on a PV mount could not reach the film plane (and you wouldn't get an image), it being too short. Although on can put a PV mount on a PL mount camera (it has been done for the 435, Moviecam SL, etc..., one simply needs to extend the flange depth, which can be done with an adaptor) I have never heard of a PL mount being put on a PV mount camera. Changing the mount on a PV camera would involve reducing the flange depth and that is more tricky than the other way round. At least Panavision 35mm cameras are not designed to be adapted this way, but maybe the design of the Genesis mount took this into consideration.

[Kevin Zanit]

Interesting points.

 

I think that to counter the problems you point out; they used a corrective element perhaps. Thus, it could explain the oddities in focal length.

 

I only know one thing for certain about the "Z" series lenses, and it is that the glass inside of them is Zeiss glass. As I said, I have no idea how that was made possible. My other guess is that they took the Zeiss elements, and modified a lens barrel to shift the elements around, thus causing where the lens comes into focus to change.

 

As I said, I don't know for certain, so I would tend to go with your knowledge. I do know that it is Zeiss glass. Next time I am in Panavision I will ask Dan about the points you raised because you got me curious.

 

 

Kevin Zanit

[Max Jacoby]

I'd be curious too as to where exactely these lenses come from, so please keep me posted.

 

Panavision is a company that, as far as I know, design their lenses themselves, but then they have them manufactured by an outside company. This is the case for the Primos, which are Leitz glass.

[David Mullen ASC]

"Z" Series (Zeiss) predate Primos, back when they assembled their lenses from other companies' elements, rehoused by Panavision. These lenses date back to the 1970's/80's and were made from Super and Standard Speed Zeiss lenses.

 

They also have a non-Z Series of these old lenses which are harder to figure out where they come from, but I think they use Cooke elements, maybe mixed with their own. The old Ultra-Speeds flare like crazy, like old Cooke's. Actually, they are kinda of pretty if you like flares.

[Max Jacoby]

That's interesting, especially if Panavision's Super Speed lenses are rehoused Zeiss Standards, since the Panavision version of these lenses is a bit faster (T1.9 as opposed to T2.1). I wonder were they found that extra fith of a stop? <_<

[David Mullen ASC]

That's interesting, especially if Panavision's Super Speed lenses are rehoused Zeiss Standards, since the Panavision version of these lenses is a bit faster (T1.9 as opposed to T2.1). I wonder were they found that extra fith of a stop?  <_<

<{POST_SNAPBACK}>

The rehoused lenses are slightly physically larger and heavier than the original Zeiss, so perhaps that somehow allowed them a slightly wider aperture? I have no idea, actually...

 

I used to joke with my camera assistant, holding up a Z Series lens, that only Panavision can manage to make a Zeiss lens bigger and heavier....

 

But now, everyone went that route and made their lenses bigger and heavier! Now the Z Series are small compared to Cooke S4's and Zeiss Ultra Primes.

[Brad Grimmett]

Not only are the Z series heavy, but they're also all different sizes, which makes them a pain to work with. Every time you change a lens you have to use a different step-down ring so that the matte box will fit. Working with the Z series really made me appreciate modern lenses that are all the same size (or close to it).

[Tenolian Bell]

What I don't understand is why all these tests involving the Genesis always have to pass the film through a DI, even if it is 4K.

<{POST_SNAPBACK}>

 

I also wonder what scanners are Panavision using. 4K is not a ubiquitous standard and all 4K scans are not created equal.

 

Spirit 4K:

 

3656x2664 10 bit log scan

 

4096x3112 10 bit log output

 

 

 

Arriscanner 4K:

 

7296x4324 16 bit log scan (micro scanning, double exposure)

 

4048x3112 16 bit log output

[David Mullen ASC]

I believe the D.I. was done at E-Film, which uses an Imagica XE for scanning, I believe. But I'm not sure what the point is -- do you think Panavision is trying to hide something? To my eyes, the 35mm footage was not being compromised in any way.

[Tenolian Bell]

No I'm not trying to point out any conspiracy.

 

Only matter of factly stating there is no singular definition of 4K across all scanners.

[Phil Rhodes]

Hi,

 

Northlight agrees that a 4K scan is 4096 by 3112 but scans at 6K.... but so long as you're scanning at either an equal or higher resolution than the result by a significant degree (which the Spirit appears not to) then you should be OK.

 

Phil