Panavision 8K Camera

[Marc Wielage]

Probably should have called this thread Red 8k camera wrapped with extras. Only thing to me thats weird is no Red logo anywhere.

The 1980s Panavised versions of the old IVC video cameras and some of the 1990s/2000s Sony cameras only had Panavision logos, as I recall.

[/COLOR]The FTC, in of all cases a DP wanting to buy a Panaflex, ruled that if a company offers a product for rent, they must also offer the product for sale. Panavision puts such a cost prohibitive price tag on their equipment that nobody wants to purchase.

Show me in print where this is true. As far as I know, Panavision has never, ever sold cameras or lenses to anybody. If anything, the FTC is only concerned about truth in advertising and warranties, not about whether somebody will rent-only or sell-only a highly specialized product to a very narrow audience.

 

[Jeffery Anderson]

But what's a regular lens? S35? PV made 70s and needed to repurpose them. They are still 70 medium format sized and I'd guess slower as a result? PV isn't interested in renting this camera out without getting their glass booked too. My red 8k will be rocking full frame still glass but no one using PV will would be my guess. It's the only reason to book from them besides this camera basically.. It will be interesting to see this going against Arri 65

S35 definitely USED to be the standard for Cinema lenses

yoi would imagine Cooke, Hawk and others will start making more full frames cinema lenses

zeiss compact zoom is FF

 

[M.Halsell]

It's really crazy how well developed and mature the Panavision DXL configuration is. Just as comparison, the original RED ONE was 10lbs, body only. The DXL can be shoulder, drone or gimbal ready with minimal effort. Amazing. just amazing. Michael Cioni is a superstar. i've had good opportunity to sit in on 2 conferences and an early RED Education session. He really has a finger on the pulse of how to quickly migrate emerging tech into a smart workflow, particularly on huge projects. And has a really good talent for explaining dense information and concepts. And Jarred Land has moved RED from the outer circle to the inner circle with a good core of continuing innovation.

 

[M.Halsell]

Does anyone know the history of Primo prior to being acquired by Panavision?

 

[Christoffer Glans]

But what's a regular lens? S35? PV made 70s and needed to repurpose them. They are still 70 medium format sized and I'd guess slower as a result? PV isn't interested in renting this camera out without getting their glass booked too. My red 8k will be rocking full frame still glass but no one using PV will would be my guess. It's the only reason to book from them besides this camera basically.. It will be interesting to see this going against Arri 65

A regular lens in regards to sensor sizes based on both cinema and still lenses. There are plenty of full frame lenses in the still photography world and they all work on the 8K Vista Vision. I haven't tried, but my Canon CN-Es should cover the 8K while still having a T1.3 aperture. Lenses going above FF still lens-equivalent are slower and much more expensive. The reason why the Panavision lenses is a good choice has more to do with sharpness. Medium lenses are (or should be) more sharp since they've always been made for high density 65mm film, so you get better edge to edge sharpness and probably no vignetting on the 8K. It's the same as full frame lenses on a s35mm sensor size, you have nothing of the edge problems that could occur on some of them.

 

[Phil Holland]

...and why continue with that when you have a Vista Vision 8K sensor instead, that enables use of "regular" cinema lenses that cover full frame. I think the most underestimated thing about the 8K sensor is that you don't have to use cinema medium glass, which would be quite costly. I don't doubt that people who use this camera and the Alexa 65 can afford such lenses, but still. Also, it gives more options when picking lenses and you don't have to be locked into super slow apertures. You can still probably shoot with T1.4 lenses in full frame. As a format, Vista Vision 8K is quite final I would say. It's the largest sensor in a "normal" working film gear architecture and it's enough resolution for many years to come.

But what's a regular lens? S35? PV made 70s and needed to repurpose them. They are still 70 medium format sized and I'd guess slower as a result? PV isn't interested in renting this camera out without getting their glass booked too. My red 8k will be rocking full frame still glass but no one using PV will would be my guess. It's the only reason to book from them besides this camera basically.. It will be interesting to see this going against Arri 65

A regular lens in regards to sensor sizes based on both cinema and still lenses. There are plenty of full frame lenses in the still photography world and they all work on the 8K Vista Vision. I haven't tried, but my Canon CN-Es should cover the 8K while still having a T1.3 aperture. Lenses going above FF still lens-equivalent are slower and much more expensive. The reason why the Panavision lenses is a good choice has more to do with sharpness. Medium lenses are (or should be) more sharp since they've always been made for high density 65mm film, so you get better edge to edge sharpness and probably no vignetting on the 8K. It's the same as full frame lenses on a s35mm sensor size, you have nothing of the edge problems that could occur on some of them.

A great deal of FF35 lenses will and do cover Weapon 8K FF. The allure of larger format glass is playing within the sweet spot and not necessarily "on the edge". This actually in some cases leads to a different look to your out of focus image.

What is good? There's lots of glass options. I've kitted up with Super 35mm, FF35/VV, and 65mm glass as of this moment to explore all of those worlds of possibilities as each shoot presents itself.

And all I'll say is..... after using the Weapon 8K and the Zeiss Otus Primes in particular it makes you very hungry for more. I've never seen motion picture footage of that quality from any camera compared to that combo.

 

[David Battistella]

For those of you wondering about the effects of using large format glass. Panavision has this great video up on vimeo. Worth checking out.

https://vimeo.com/167052317

 

[Mike Nagel]

Panavision may have deeper access to the sensor - who knows. But the point is that if you give Light Iron your r3d they will transform it and develop it to its full potential using the experience and knoweldge of colourists and colour scientists. To my mind, RED have always just offered you the information in an unbiased way and said 'let's see what you can do with it' whereas ARRI have tried to make it look a certain way at the get-go.

Agreed, and absolutely possible they get access to the raw sensor data... Dragon is a fantastic sensor but does have issues with contamination in the skin tones... Fixable, but takes extra time... Every time

very possible Cioni does custom debayer and/or manipulates the raw data earlier in the stage to achieve higher color accuracy from the sensor data... Graeme does it with every color science update...

Cioni has graded tons and tons of RED stuff for a long time and he surely knows where .r3d footage deviates most commonly and what fixes to apply to bring it back to norm... Will be interesting to see how much more accurate the Light Iron image is...

 

[Christoffer Glans]

There's a great interview with Bill Bennett ASC who was there. Maybe someone already posted, but there's some answers to the questions posted here.
http://www.newsshooter.com/2016/06/02/audio-interview-bill-bennett-asc-goes-hands-on-with-the-panavision-dxl-8k-camera-with-red-sensor/

 

[Ugo Del]

Interesting things here too: http://www.liftgammagain.com/forum/index.php?threads/panavision-dxl.6817/
Ian Vertovec, from Light Iron, answered some questions about how they tweaked the camera.

"We profile the cameras and build color adjustments in the widest possible color space. Then that adjustment is embedded in the RED SDK. This way we work off the raw data and you can still output to rec. 2020, ACES, etc, and still maintain all the original camera latitude. And since it is embedded in the RED SDK, it's available in any software that supports RED. The .r3d's the DXL creates will have special meta data that only makes LI color available with DXL media. We are also planning on building our own OLPF and IR filter built to Panavision optical standards, so it really will be a completely different looking camera, but with all of the amazing advantages of Redcode and the 8K sensor."

"On the user side, it will work the same way "DragonColor" and "DragonColor2" work now. You just select "Light Iron Color" instead. So it should be totally seamless to the way you use .r3d's now. Under the hood, we are doing quite a few unique things to create most cinematic off-the-shelf color possible. I won't go into details on exactly how, but our top priority is good, clean skin tone and rich, vibrant saturated colors and not plastic or electric-looking colors. We are not creating a look, but a rich, cinematic starting point to create a look. Come to CineGear if you want to see it! I also want to make sure this isn't framed as a "DragonColor" vs. "Light Iron Color" conversation. We've worked on footage from all the major cameras and everyone handles the same problems in slightly different ways (skin tone, IR cut, gamut remapping, etc..). So Light Iron color is really a little bit of what we've observed from the market in general, plus A LOT of input from many, many cinematographers over the years on what they love and don't love from all the images they've made."

"There may be a reason for us to do our own SDK at some point, but for the foreseeable future it will be though RED. An easy implementation for third parties and users was a high priority. That is why we kept the .r3d format."

 

[Christoffer Glans]

Interesting things here too: http://www.liftgammagain.com/forum/index.php?threads/panavision-dxl.6817/
Ian Vertovec, from Light Iron, answered some questions about how they tweaked the camera.

"We profile the cameras and build color adjustments in the widest possible color space. Then that adjustment is embedded in the RED SDK. This way we work off the raw data and you can still output to rec. 2020, ACES, etc, and still maintain all the original camera latitude. And since it is embedded in the RED SDK, it's available in any software that supports RED. The .r3d's the DXL creates will have special meta data that only makes LI color available with DXL media. We are also planning on building our own OLPF and IR filter built to Panavision optical standards, so it really will be a completely different looking camera, but with all of the amazing advantages of Redcode and the 8K sensor."

"On the user side, it will work the same way "DragonColor" and "DragonColor2" work now. You just select "Light Iron Color" instead. So it should be totally seamless to the way you use .r3d's now. Under the hood, we are doing quite a few unique things to create most cinematic off-the-shelf color possible. I won't go into details on exactly how, but our top priority is good, clean skin tone and rich, vibrant saturated colors and not plastic or electric-looking colors. We are not creating a look, but a rich, cinematic starting point to create a look. Come to CineGear if you want to see it! I also want to make sure this isn't framed as a "DragonColor" vs. "Light Iron Color" conversation. We've worked on footage from all the major cameras and everyone handles the same problems in slightly different ways (skin tone, IR cut, gamut remapping, etc..). So Light Iron color is really a little bit of what we've observed from the market in general, plus A LOT of input from many, many cinematographers over the years on what they love and don't love from all the images they've made."

"There may be a reason for us to do our own SDK at some point, but for the foreseeable future it will be though RED. An easy implementation for third parties and users was a high priority. That is why we kept the .r3d format."

From that forum

We profile the cameras and build color adjustments in the widest possible color space. Then that adjustment is embedded in the RED SDK. This way we work off the raw data and you can still output to rec. 2020, ACES, etc, and still maintain all the original camera latitude. And since it is embedded in the RED SDK, it's available in any software that supports RED.

The .r3d's the DXL creates will have special meta data that only makes LI color available with DXL media. We are also planning on building our own OLPF and IR filter built to Panavision optical standards, so it really will be a completely different looking camera, but with all of the amazing advantages of Redcode and the 8K sensor.

:undecided: I guess that we won't be getting LI color science in our weapons then. Also seems confirmed that they will make a special Panavision OLPF. I guess we can hope that Red comes out with a new color science that is not proprietary to Panavision but with elements learned from the LI Color Science. Hopefully Light Iron could help us little guys out as well... you know, for the greater good of all :tongue_smilie:

 

[David Battistella]

very possible Cioni does custom debayer and/or manipulates the raw data earlier in the stage to achieve higher color accuracy from the sensor data... Graeme does it with every color science update...

Cioni has graded tons and tons of RED stuff for a long time and he surely knows where .r3d footage deviates most commonly and what fixes to apply to bring it back to norm... Will be interesting to see how much more accurate the Light Iron image is...

I'm not sure what it is deviating from? Your eye? Film? Other digital camera's?

To my eye, dragon with the STH and the DC/RLF combo yields very accurate color, perhaps not color we are "used to" with film but certainly very accurate.

Film and "other digital" camera's push toward skin tone and in feature work this is crucial, but they don't seem to produce the most accurate landscapes for example.

It will be interesting to see what happens with the light Iron Pipeline which surely will be geared to features and filmic looks. But I do rally appreciate what Graeme has done with the color science in terms of accurate color reproduction.

David

 

[Patrick Tresch]

Well if the "filmic look" DXL sdk won't be available for Redusers I think it's time to ask RED for an emproved "filmic look" RED WEAPON SDK!

I wonder how much FUD this new "color science" call is?

I imagine they allready must have done tons of DragonColor VS DXL colors tests if they can claim it's better. Hope they bring some 4k DCP tests at Cinegear.
I wonder how Graeme feels about it? He was the only one having the keys to the RED SDK!

Pat

 

[Michael Lindsay]

It makes absolute sense that PV needs to build their own OLPF stack (and then they obviously need their own colour to support it)

.. Their new lenses, which will be pushed for use with this camera, are uniquely designed to expect 7mm of glass ,with a specific refractive index, behind the lens. On its own (without the glass plug) the red weapon OLPF stack is too thin.... And the optical plug system seems far from ideal with it's extra glass to air surfaces. So to get the best out of their amazing optics they actually need their own maths in the decode.

Only Panavision had the business model to suport this... (Vantage may get there soon as well)

I think people should be pleased PV are going to be giving Red more feedback on their chips/camera from a demanding filmmaking Eco system...

Their is nothing wrong with red's current colour science coupled with Red weapons OLPFs. GN and the guys at red have given us multiple great starting points to build nearly any look imaginable... If you throw in the work John is doing at kippetie! (We should all be suporting those guys more) RED already has a very flexibile cinema camera..

 

[Patrick Tresch]

Well now I don't know wich infos to take from Ian Vertovec regarding their "color science":


- We profile the cameras and build color adjustments in the widest possible color space. Then that adjustment is embedded in the RED SDK. This way we work off the raw data and you can still output to rec. 2020, ACES, etc, and still maintain all the original camera latitude. And since it is embedded in the RED SDK, it's available in any software that supports RED.

- Under the hood, we are doing quite a few unique things to create most cinematic off-the-shelf color possible. I won't go into details on exactly how, but our top priority is good, clean skin tone and rich, vibrant saturated colors and not plastic or electric-looking colors. We are not creating a look, but a rich, cinematic starting point to create a look.

- There may be a reason for us to do our own SDK at some point, but for the foreseeable future it will be though RED. An easy implementation for third parties and users was a high priority. That is why we kept the .r3d format.


So at wich image processing stage do they apply their "Color Science"?
Pat

 

[Aaron michaels]

Digital cameras have always had the ability to shoot smooth shots...build in zoom abilities even if you change a lens. Why work harder when you don't need to right. Times have changed and i think they have the best of both worlds and now it doesn't look like hardcore video. Digital in the same bed as film??? its about time!

 

[Mike P.]

:undecided: I guess that we won't be getting LI color science in our weapons then. Also seems confirmed that they will make a special Panavision OLPF. I guess we can hope that Red comes out with a new color science that is not proprietary to Panavision but with elements learned from the LI Color Science. Hopefully Light Iron could help us little guys out as well... you know, for the greater good of all :tongue_smilie:

Yeah, it's kind of frustrating. Obviously I understand why they want to make it proprietary -- to maximize profits -- but it kinda sucks that it's merely an attempt to control it via a metadata flag. I mean, chances are it would do some wonky things with non-PV OLPF & IR stack anyway, but it certainly would've been cool to look/compare based off of the DXL look (whatever it may be... still would like to see some .r3ds and/or 4k footage). I guess the thing is, I'm fairly confident in LightIron ability to make a really solid look out of the camera.

Conversely, it'd be in Panavision's best interest to make the LIcolour/gamma available merely so gimbal/drone footage shot with the regular Weapons (1/3 the size/weight) matches the DXL acquired footage... I mean, one could hope, right? : )

Well if the "filmic look" DXL sdk won't be available for Redusers I think it's time to ask RED for an emproved "filmic look" RED WEAPON SDK!

I imagine they allready must have done tons of DragonColor VS DXL colors tests if they can claim it's better. Hope they bring some 4k DCP tests at Cinegear.
I wonder how Graeme feels about it? He was the only one having the keys to the RED SDK!

I already touched on this (either in here or the recon thread), but I think Graeme is (rightfully) more focused on accuracy/mathematical correctness, rather than a semi-subjective "look" (under the presumption that the look gets created off the raws in the grade anyway). Whereas LI, who's traditionally a post outlet, knows that 99.9% of the time people want a traditionally "film look" (in both colour and contrast), so it makes sense to have that out of the gate. Anecdotal evidence of this is seen in the Alexa/ALEV -- they presumably force slight-lowcon (as its on the OLPF stack), they have heavily boosted low-end, it's an ambiguous Log mode (rather than one identical to cineon), it's only 2k~3.2k, etc. -- yet, people still flock to it and the vast majority prefer it's "look" even compared to dragon.

But, yes, if a lot of dops flock to use the DXL over weapon, it'd be a smart move for RED to release something similar (especially now that prores/dnx are internal and baked)...

 

[M Most]

... I think Graeme is (rightfully) more focused on accuracy/mathematical correctness, rather than a semi-subjective "look" (under the presumption that the look gets created off the raws in the grade anyway). .

I can't speak for Graeme (although I'm sort of going to anyway), but color science has always been both an art and a science, balancing the physical principles of light, the human perception of it, the digital representation of it, and the art of what is pleasing in human visual experience. Graeme knows this very well, and while his earliest efforts might have been a bit more slanted to the pure science, he has balanced all of these considerations very well for a number of years now. Tne math is important, which is why (at least for the most part) colorists are not color scientists. And the art of color grading is not the same as the art of gamut mapping, writing transforms and matrices for various purposes, and creating LUTs for color space conversions. In both cases there is an element of technology and science, and an element of art. But looking at color science as "pure science," and casting Graeme in that light, is not really accurate or fair to the man, the position, or the results.

 

[David Battistella]

But looking at color science as "pure science," and casting Graeme in that light, is not really accurate or fair to the man, the position, or the results.

This is one of the most balanced statements I have read on Reduser since my time here and I could not agree more.

David

 

[Mike Finegan]

can you share what the objections were that have been addressed?