F35: F35 will still look better (8K Pana/Red)

V

VinceinLA

Carbonite Member
http://panavision.com/dxl

I bet the CCD will still be the better looking choice. And what on earth good is 8K anyway? I really dont get it and yes, I understand the arguments for it but the reality of all this ultra high-rez is pretty boring in the end, unless maybe you are making something for real IMAX screens but only .00000001% of us get to do that.
 
its a vista-vision camera, so they're using red's vista-vision sensor, which happens to be 8k. they went 6k for their super 35 cameras for additional repo options along with, in effect, giving you the ability to somewhat supersample. The cool thing is that its one more option thats making vista vision a thing again. from a vfx standpoint but also just from an ascetic standpoint, its kinda awesome. 8 perf is back baby lol
 
You mean 8-perf digital that will look the same as all the rest of the high-end digital motion images out there? It's more fluff for those who don't actually sit in theaters and watch films, than anything useful for making stories entertaining. I forgot the other thing: 98% of people watch movies on 40" TVs and iPads!

Honestly, who cares about resolution or sensor size anymore? No one can tell which is which anymore, who isn't sitting in front of an 80ft. IMAX screen, nor do they care. Give me more Kodak so I at least FEEL something anyday.
 
John Galt from Panavision 2009 essay on the resolution wars, in relation to the Genesis' CCD. He gets flamed in the comments.

https://library.creativecow.net/galt_john/John_Galt_2K_4K_Truth_About_Pixels/1


"But if you use the arithmetic that people are using when they are taking all of the photosites on a row and saying they're 4K, they are adding the green and the blue together and saying, "Oh, there are 4K of those, so it's 4K sensor." Now actually, in order to get RGB out of a Bayer pattern you need two lines. Because you only have green plus one color (red) on one line, and green plus the other color (blue) on the other line. You then have to interpolate the colors that are missing from surrounding pixels."

"So 4K is not these 8 mega pixel or 9 mega pixel or 10 mega pixel CMOS images for the Bayer pattern where they add up all the pixels in a row and say hey, we got 4K. The great perpetrators of that mythology have been RED and Dalsa. That's why I call these "marketing pixels." It's intentional obfuscation. Because they really do nothing to improve image quality. They may improve sales volume. But they don't do anything to quality."

"The most important issue from our point of view, is that we want to have equal resolution, TRUE edge resolution in red, green and blue. The most important thing is not to have interpolated information. You want to know that the edge is REAL."

"A number of years ago some IMAX engineers - and I don't think IMAX ever let these guys out of their lab again -- did this wonderfully elegant experiment at the Large Film Format Seminar at Universal Studios Imax theatre. They showed this film they made that began with 2 rows of 2 squares: black white, white black, as if you had 4 pixels on the screen."


"Then they started to double and double and double the squares. Before they got to 4K the screen was gray. Do you know what the means? There was no longer any difference between black and white, which is what allows you to see sharpness. It's the contrast that we see, not the actual information. Technically, the MTF (Modulation Transfer Function) was zero at 4K!"
 
Galt's explanations go to the heart of a dispute about "sampling" vs. "synthesizing" information. In Panavision's approach, they use a 5,760 x 3,840 (a nearly 6k sensor) to create a 1920 x 1080 purely-sampled image. Each of the pixels in the image is distinctly sampled from individual red, green, and blue pixels (actually 2 red, 2 green, and 2 blue). There is no interpreting, no Bayer pattern, no demosaic algorithm, it's true actual resolved detail. But it's limited to 1920 x 1080 - although it's a true 4:4:4.

It's in some ways similar to the original Canon C300's sensor, where a 4K sensor was used to make 1920 x 1080 HD by using four pixels (1 red, 1 blue, and 2 green) to make each HD pixel. In Canon's case, there was no demosaic done on the Bayer pattern either. Sony/Panavision used six pixels in a vertical stripe (red next to green next to blue, two rows per HD pixel) whereas Canon used 4 pixels in a conventional Bayer arrangement (red/green on one line, green/blue on the next line, all sampled to create one HD pixel).

Bayer-pattern sensors don't sample all that detail, they synthesize and they guess as to what the missing components are. They're pretty good at that guessing, and the results definitely do look good, but the fact of the matter is they're still synthesizing and there's no two ways about it. At any pixel location there's only one of the three primary colors, the other two primary colors for that pixel are guessed, projected, synthesized, uprezzed, or otherwise "fill-in-the-blanks". And that's one reason why Bayer pattern sensors don't deliver the resolution they're rated at; a 4K sensor (4096 pixels wide) generally resolves around 3.2k worth of detail once it's been demosaicked, etc. But it's done at an effective ratio of only about, approximately, 4:2:0.

Where Panavision lost the argument is in that even though a Bayer pattern results in inefficient, fuzzy, and synthesized resolution, it still creates more resolution than the same number of pixels distributed in the RGB stripe. Panavision used 6k x 4k pixels and ended up with a pristine, wonderful 1920 x 1080 at 4:4:4. Bayer cameras can use the same 6k x 4k sensor size, and end up with a crisp, clear, razor sharp 4K, at somewhere approaching 4:2:2-ish. The market has spoken, and regardless of tests or not, anyone who sees 4K footage next to 1080 footage can tell the difference when projected on a large-enough screen. The orientation towards 4:4:4 ensures pristine sampling in the Panavision/Sony approach, but the synthesized resolution of the Bayer system definitely results in the end-user's perception of more sharpness, and the tradeoff of actual sampling fidelity and color resolution have proven to be "no big deal" to the viewing audience.

I've always wondered what could be made out of the Panavision's sensor if you got ahold of the raw stream and applied a type of demosaic to it; the raw sampling is there (there's a 1920x2160 red matrix, a 1920x2160 green, and a 1920x2160 blue) that with some demosaic-style synthesizing you could probably make a perfectly good 4K out of it...
 
VinceinLA said:
http://panavision.com/dxl

I bet the CCD will still be the better looking choice. And what on earth good is 8K anyway? I really dont get it and yes, I understand the arguments for it but the reality of all this ultra high-rez is pretty boring in the end, unless maybe you are making something for real IMAX screens but only .00000001% of us get to do that.
Click to expand...



and kodak film will still be better than all of them and cheaper to work with
 
Barry_Green said:
... I've always wondered what could be made out of the Panavision's sensor if you got ahold of the raw stream and applied a type of demosaic to it; the raw sampling is there (there's a 1920x2160 red matrix, a 1920x2160 green, and a 1920x2160 blue) that with some demosaic-style synthesizing you could probably make a perfectly good 4K out of it...
Click to expand...
Well, ARRI uprez'es their final image and it's kind of hard to tell from a real 4K. It may be that "synthesized" 4K but it fools the eye into thinking that it's a real 4K.

That new Pentax K1 with the Sony 36 MPX sensor (basically, the one off D810) has a pixel shifting mode that not only gives a higher resolution but, off the in-camera interpolation, higher dynamic range as well. At the moment, it can only do this in the stills mode and even then only for landscape/interiors and not for portraiture. But one reckons that a few years down the road, this pixel shifting will travel down to video too. And then it all becomes "synthesized" in the sense that the final image will exceed the Raw data/straight-off-the-sensor capabilities. The only question will then remain whether to do this in camera or in post.
 
Barry_Green said:
The market has spoken, and regardless of tests or not, anyone who sees 4K footage next to 1080 footage can tell the difference when projected on a large-enough screen. The orientation towards 4:4:4 ensures pristine sampling in the Panavision/Sony approach, but the synthesized resolution of the Bayer system definitely results in the end-user's perception of more sharpness, and the tradeoff of actual sampling fidelity and color resolution have proven to be "no big deal" to the viewing audience.
Click to expand...

Barry I'm reading your thorough reply as a strong argument in favor of the RGB striped CCD. You've got me wondering if the "digital look" can be somewhat attributed to interpolated data.

Now in terms of your point above that the Bayer filter CMOS approach wins in the perceived sharpness category, that may be true. But ok big picture, according to Galt 4k doesn't even resolve on an IMAX screen:

"A number of years ago some IMAX engineers - and I don't think IMAX ever let these guys out of their lab again -- did this wonderfully elegant experiment at the Large Film Format Seminar at Universal Studios Imax theatre. They showed this film they made that began with 2 rows of 2 squares: black white, white black, as if you had 4 pixels on the screen.
Then they started to double and double and double the squares. Before they got to 4K the screen was gray. Do you know what the means? There was no longer any difference between black and white, which is what allows you to see sharpness. It's the contrast that we see, not the actual information. Technically, the MTF (Modulation Transfer Function) was zero at 4K!"
 
I really get tired of language that implies some "K" looks "better" than some other "K". I guess it's one of many ways to tell a pro from an amateur, although I have heard guys working in production for 20 years think they really have something because they bought whatever the latest marketing ploy told them to buy... so maybe that isn't a good method.

Maybe a better indicator is being able to spot the craftsman, from the guy who just has a button-pushing job.
 
When Barry brought up the "synthesized" look, I first began to think about the music recorded back in the 50'-60's that were transferred onto a CD. At first, there was a lot of hiss. Then, the audio engineers removed hiss. Then they began to add to that "synthesized" sound by turning those old monaural recordings into stereo. But then I thought, "Why even bother with audio? Two weeks ago, I saw a colorized episode of "I love Lucy". You can't get much more "synthesized" than that ... unless you include the purely CGI based cinema like many modern era kids' flicks".

BTW, the IMAX black+white=gray test was intellectually rigged. Real life images are not shapeless matrices of these black-and-white dots. In other words, what they were doing was a mathematical certainty but it doesn't translate to anything else.
 
I just whatched the Panavision showreel today at Cinegear. It looks very nice. It looks just like red footage.
 
MikalZ said:
Barry I'm reading your thorough reply as a strong argument in favor of the RGB striped CCD.
Click to expand...
Well, not necessarily. I think it was a brilliant accomplishment, and on its technical merits I believe it produces the crispest, sharpest, most accurate, most accurately-sampled 1080p you could get, at a pristine 4:4:4. It's a great design. But is it "the best"? Depends on what you're looking for. Red makes "4K" images out of fewer pixels than the F35's 1080p sensor. Their 6K sensor has around the same number of pixels as the F35's RGB stripe sensor, yet it makes a really, really solid 4K. So the RGB stripe sensor is an excellent way to go if you've got the pixels to spare, but you can get arguably higher-resolution footage using the bayer pattern even with many fewer pixels to work with.

I always liked the RGB stripe design, and the Canon C300's simple no-mosaic, 4:4:4 sampled approach to 1080 from a 4K sensor. They very much appeal to the engineering sentiment in me. But I also have to hand it to the Bayer, because it generates really good-looking footage, shockingly good when you consider the "guesswork" involved, and it does so with fewer pixels than the RGB stripe needs.

Now in terms of your point above that the Bayer filter CMOS approach wins in the perceived sharpness category, that may be true. But ok big picture, according to Galt 4k doesn't even resolve on an IMAX screen:
Click to expand...
Yes, there's definitely an argument to be made there. But in terms of simple black/white, the Bayer sensor with 4K pixels records a distinct, sharp difference in black and white lines up to 3.2K worth of resolution (about 1600 line pairs). The RGB stripe sensor in the Panavision is likely going to resolve a clear and distinct 1000 line pairs. Even a little GH4 can clearly show 4K/Bayer res charts that exceed the resolving power of the F35, so -- I'd have to revisit Galt's tests to see exactly what happened. I understand the appeal of the argument, certainly, but -- when I put a DVX200 up on a res chart and see it clearly resolving much higher than the Red One's original 4K, easily visible on a computer monitor (much less on an IMAX screen), then - I'm not sure how to reconcile that against the claims of the MTF being reduced to zero at 4K. You'd have to look at the projector resolution (or was it film projection?) etc.

A simple res chart comparison between an F35 and a 6K Epic Dragon should demonstrate the difference in actual resolved detail.
 
http://forum.blackmagicdesign.com/viewtopic.php?f=2&t=21212

It appears that sensors are hitting a wall on image return. The next gen of cameras will be repackaging of old cams as the style of how we use them changes. Sorta like how the F3-FS700-FS5-F5-FS7 chip has made its way around sony.

Wouldn't it be just wonderful if Sony decided to repackage the F65 with modern smaller components.

Kind of a bummer because the rumor of the 4k Panavision CCD camera probably dies with this camera.
 
yachacha said:
http://forum.blackmagicdesign.com/viewtopic.php?f=2&t=21212

It appears that sensors are hitting a wall on image return. The next gen of cameras will be repackaging of old cams as the style of how we use them changes. Sorta like how the F3-FS700-FS5-F5-FS7 chip has made its way around sony.

Wouldn't it be just wonderful if Sony decided to repackage the F65 with modern smaller components.

Kind of a bummer because the rumor of the 4k Panavision CCD camera probably dies with this camera.
Click to expand...

it will be interesting to see what the technology is that moves us beyond the CMOS sensor, and if it will be something that is in fact economical when compared to the CMOS type. It is a bit sad to see that panavision's entry is far less revolutionary than the Genesis, but that may simply speak to the cost and challenges in moving beyond the current hardware. It may well require a disruptive innovation, or perhaps a massive company with money to blow, to get us to the next generation
 
macgregor said:
I just whatched the Panavision showreel today at Cinegear. It looks very nice. It looks just like red footage.
Click to expand...

Saw that too and was.... underwhelmed. I just saw footage that looked like everything else, honestly. Maybe when some stupid hero movie is shown in a true IMAX theater it will show some slight advantage in clarity, if you want to call that an advantage? Whatever. I'm officially done going to camera demos and caring about "what's next"; it will not be anything worth paying attention to. It's all the same now!

I will buy an Arricam LT when they are cheap enough and digital will be an F35 or Amira, if I can help it.
 
yachacha said:
It appears that sensors are hitting a wall on image return.
Click to expand...

DLD said:
In what way?
Click to expand...


I heard (sorry, I do not remember where) about a year ago that sensors were already very close to unity output. That is, one photon enters, one electron goes out.

In terms of dynamic range capability, that would be one very important limiting factor. And many other IQ parameters.

The fact that Arri can't (yet) replicate their ALEVIII with smaller photosites (or, at least smaller enough to make a difference) certainly says something. I'm sure they probably will eventually, but it will be hard-won.
 
It's not just image quality but ergonomics and usability. Similar to the difference between stills lenses and cine lenses. Not so much always an image quality advance, but focus marks, breathing, rotation, smoothness, etc.

So so in my opinion the new Panavision camera is awesome in terms of what it provides for a cinema production environment. Very well designed, from the looks of it. And a sensor the red sensor which is the most popular look they could access...since Arri won't collaborate, probably.

I I applaud their efforts. I mean, controlling their lenses without external motors is great! Amongst the host of other options. Perhaps it's not as crazy a step as the Lytro, but still a camera I would love to use someday.
 
You must log in or register to reply here.
Back
Top