Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Panavision videos...

Putting it in simple words sounds like their questions is: If one has a 4K bayer sensor sampling with 12Mpixels, how come one gets (3 x 12 Mpixels =) 36Mpixels color data out of it?

We know that Graeme says one gets about a 3.2K out from the Red One 4K bayer sensor. Still this 3.2K corresponds to about 22Mpixels of color data, so in this sense it's a reasonable question to ask where the additional (22 -12 =) 10Mpixels come from.

Evidently, on one hand, the guys on the video go too far when claiming the image has only a color space 4:2:0. For, as somebody said in another thread, anyone whose tried keying with a 4:2:0 image and a Red One image is able to tell RED are not 4:2:0. But, on the other hand, with no doubt, these guys are professionals and deserve to be taken seriously.

I suspect the explanation boils down to extrapolation where as in the Panavision videos they implicitly assume debayerian is strictly about interpolation. In simple words, if there was a hypothetical ideal 4K camera yielding prefect 4:4:4 images, the outcome of the 4K debayerian process image need not to be precisely one-to-one with such a hypothetical image. Instead, once both images are downconverted to 3.2K, it is enough one is not able to say the hypothetical is more right than the other.
 
Shawn Nelson said:
The problem is that with the merging of digital technology and art, it becomes easy to mislead people who are gods of art into believing really stupid technological lies. Then, if you dont technically tell lies but just throw out enough tech to confuse the artists and try to reassure their feelings, they may be successful at least for a while.
Click to expand...

So, in other words....
If you can't dazzle them with brilliance, baffle them with bullsh#t?
 
Hahaha, just ran into this thread... Jim sure likes to stir the pot ;o)
 
While Pana is busy Demystifying Digital Camera Specifications,
by the end of the summer there could be some 4000-5000 happy Red cameras taking-over the industry.
Keep busy Pana, keep busy:)
 
Lauri, I fully understand how people think that Bayer patterns are "magic" and hence the pixels you get out, are not real.

However, if you do the math, we get a true 2k in R, G, and B, (igorning the second green pixel in the bayer array). That gives us another 2k of green detail. To get full 4k on green we'd need 4 * 2k greens (remember it area of pixels). From 2k * 2 of green, that's enough pixels for 2.8k green. And that's without any bayer magic. Now, that doesn't take into account how the greens are interleaved and how the demosaic works and we're getting 3.2k out of that, of which I can assure you from the above math that 2.8k of it is at least very real.

I strongly disagree about calling a sensor a 4:2:0, as not only is it bad terminology for where it was designed to be used for Y'CbCr chroma sub-sampling, but absolutely does not apply to a Bayer pattern sensor, and indeed does not apply to any sensor configuration.

Graeme
 
Daniel Reichenbach said:
The only thing what counts for me is: Do I like an image or not. Can I deal with the technique of a camera or not. What I get out of my RED is more than what I got out of any other digital camera I was able to rent or buy in the past. RED helps me to transform the reality to the digital world in a very special way like I only got from 35mm film. I don't need explications for that, I don't need somebody who tells me, oh but, by the way RED 4K is not really 4K. I give a s/&%/&t on that. RED is a concept, RED is moving, every day, RED don't look back, RED flirts with the future. There will be better REDs, better EPICS soon, I'm convinced. But what I already have is perfect. I love this camera, because the pictures of RED ONE are closer by my heart than any other I got from digital cameras.
Click to expand...

I think this one summarizes all that matters.
How image looks is the most important thing.
Realized this one definitely last Saturday. I went to the movies (Before the Devil Knows You're Dead) and before the movie started we saw some commercials. They showed a commercial (shot on 35mm) i just finished post in SD blowed up to 35mm. I could not believe how good SD looked on 35mm , just because good cinematography and telecine CC.
I i wasn't the one who handled all post in SD i would have said it was 2K ... because it just looked good.
Sure resolution matters, but sure we think too much about megapixels as if it was the first thing we should care of ...
 
G.I. Joe in RED

G.I. Joe in RED

Speaking of which maybe this is not the right place or someone else has noticed this but....G.I. Joe according to IMDB is being shot fully on RED ONE.

G.I. JOE IMDB LINK.

G.I. JOE IMDB Technical
 
Jay A. Kelley said:
I like RED.. It's a cool camera, and it makes very pretty pictures.

I'm good with that.

jay
Click to expand...

Me like...

I get many pm's these days. It's really no secret sauce. It's just a cam... You'll beat me any day. Just hit record... and develop them nicely.
 
"A so-called “4K camera” that has an imager with only 8 or 10 million photo sites but outputs 25 or 32 megapixels is interpolating data without adding any image performance."

Hmmmmm...

LOL

:)
 
So Graeme, after watching the videos from Panavision I have questions, (BTW: I've highly respected Larry Thorpe for years, he's taught me most of what I know about lenses).

What's the OLPF compromise you find most appropriate for the Bayer pattern and is there hypothetically (or existing) a way to have an OLPF that filters R,G & B differently?

Also, very hypothetically speaking if one could design a Bayer sensor that has the B & R sensitive "photosites" physically displaced from the G plane does that again, hypothetically count as a way to avoid aliasing?

As a future technology, could the OLPF also be implemented directly on each "photosite, hehehe" along with the dye?
 
I i wasn't the one who handled all post in SD i would have said it was 2K ... because it just looked good.
Click to expand...
Contrast and sharpness make a huge difference in your perception of sharpness, more so than resolution (!).

http://www.cambridgeincolour.com/tutorials/sharpness.htm

But there is a way to check for real resolution. Try to read small text in the frame... you can't really fudge this like you can with sharpness.

2- This is why I'm more excited about displays with high contrast and real blacks than 4k.
 
The issue is almost a non issue. Currently as we have it, I've not seen chroma aliasing on real world images. I used to see it on Frankie before we had the OLPF. I can see it on RED as it is today on test charts, but we have the de-moiré tool that removes it without removing luma resolution. On luma we're more heavily optically low pass filtered than other cameras (because we're bayer and do need the OLPF to eat into the chroma somewhat) and because we want negligible aliasing on the luma. We get negligible aliasing on the luma, which is more than can be said for most other cameras.

Graeme
 
So far, the rare issues in real world applications has been neglible to me. In your opinion does the compromise affect heavy color correction and compositing work in any manner?
 
JJ chapter #6 ... did you notice the subliminal cue? There is an interesting contradiction by him first stating ( In religious terms no less!!! ) a claim to being "Agnostic" and then singling out Red from the usual suspects by having all the other cameras listed side by side in lavender and Red in red! ... A most telling detail.

Very sloppy and quite frankly beyond Burlesque.


DogDay.
 
Hahaha, maybe it was Red because the powerpoint designer thought he was being cleverly "creative.
 
All I can say is… statistics don’t lie but liars use statistics.

More to the point. Images don’t lie.

If you sat in Red’s cinema at NAB and watched “100”, the images from the first 100 cameras shot over the first 100 days (as I would suspect Panavision did) their arguments become less meaningful.

That presentation was stunning. This camera creates incredible images. It’s not surprising
that anyone (including that ASC member Jim mentioned) was converted.

That kind of quality, at that price point, would concern any company that sells cameras. It has certainly caught Panavision’s eye.
 
Graeme, do you think that one day it will be possiable to create debayer algorithm such that their will be NO way to tell the difference between a stripped RGB and bayer sensor? In reality I am seeing that as far as the image goes having a bayer sensor can improve the over all fluid feel of the image, causing it to feel more like film but without grain. Is what I am seeing right, whats the cause for the image to act in this fashion?
 
Videodisc said:
That presentation was stunning. This camera creates incredible images. It’s not surprising that anyone (including that ASC member Jim mentioned) was not converted.
Click to expand...



Videodisc, watch those double negatives; I think that phrase there is saying the opposite of what you're trying to state.
 
You must log in or register to reply here.
Back
Top