4k by 2k or Quad HD...lots of rumors? thoughts? - Page 32

Where though did you get this information?

I read the book "Defining Vision" by Joel Brinkley about the development of ATSC based on interviews from those who made it. From what I read ATSC was based on the technical limitations of the time it was made. In fact there were strong efforts by certain people to promote progressive over interlace for ATSC even though the best progressive system at that time could only do 787.5p (here is a link to it).



3.16 PH is based on 20/20 vision. I would mention that 20/20 vision is nominal visual acuity so there are people who can see better than that (for example Ted Williams had 20/10 vision).

Once again guys: The point here is NOT, NOT, CCD vs. CMOS! The point is 3-chip vs. single chip tech. CMOS advanced to the point it is better than CCD, no doubt about it. However, 3-chip tech has some inherent advantages which are relevant when one it talking about esotheric details of 4K vs. 2K res . This is simply resulting from the analysis of Bayer vs. 3-chip (or Foveon). There is a difference if the RGB subpixels are arranged precisely in the same spatial location or if they are dispersed over and area in a RGGB fashion, processed in overlapping and RGB is restored digitally. This was shown many times using Foveon sensor. Btw, I am not claiming Foveon is overall better since it has its own problems and it is difficult to judge it they are ultimately inherent or could be eliminated. The Foveon main problem is blooming, at high light intensities there is dispersion of carriers to neighboring pixels. But in terms of sharpness Foveon is unparallel and this comes from the RGB subpixels arrangement.

Obviously the single chip CMOS is much cheaper than the 3-chip system. Moreover, the 3-chip system is not practical for digital cameras. This means the 3-chip systems are finished. But is does not mean the 3-chip is not principally superior, it means that single chip is just good enough. This is precisely the same type of problem like edge-lit vs full locdim TVs. Full locdim is principally better /especially with high number of zones is hard to beat/ but since edge-lits are good enough and cheaper the locdim moves into an extinct species.

That's not true at all, Foveon sensors are considerably lower resolution than CMOS, and Foveon is a layered design, not an "RGB subpixel" design. What this means is you have worse sensitivity to light, and colour contamination when light "leaks through."


3-chip designs are not without compromises either. There are additional optics required which can only degrade image quality. There will be more heat generated from three sensors, which means more image noise unless the chips are actively cooled. Most camera sensors in single-chip cameras, are not correctly aligned with the lens mount. In most cases, it won't show up, but now you expect them to get it right with three sensors, when they can't get it right with one?

If you are oversampling the image, bayer filtering really doesn't matter or reduce image quality at all.

Even you have said yourself that 3-chip systems do not seem to have a future, so why do you keep bringing this up?

And there is no resemblance between any of this and Edge LED vs Backlit LED at all. That's a bizarre comparison to make.


All of the highest resolution cameras available, even those using CCD chips, are using bayer filtering, even PhaseOne's $44,000, 80 megapixel IQ180 digital back. It's not the severely compromised design you seem to be making it out to be, 3-chip brings far more problems than it's worth and considerably increases costsand not simply because you're paying for another two sensors.

Miracles??

Hmmm lets see when I was growing up flatpanels were Star Trek only, ditto for worldwide library data in a computer and talking on a wireless telephone.

Today we have the internet, LCD and LED flatpanels and cellphones.

Two years ago a bulbless 720p, compact, low cost, quiet projector was a miracle waiting for some day and right now my Son has a Qumi LED 720p projector miracle.

Miracles do happen thanks to evolving technology but the trick is you have to be patient a year or two or five and boom!! out of left field here comes the next miracle.

The less complex a design the less failures and increased reliability/adherence to specs.
Thats why in the single chip versus three chip debate my vote goes to SINGLE chip every time.

2K to 4K is not a miracle. Sorry.

What are you really expecting from 4K that 1080p didn't already improve from SD?

I know what i do want to and it doesn't involve more pixels. Which is all anybody in this thread ever mentions. You can improve things with 2K that will give better PQ.

I take it you've never seen a 4k source on a 4k panel? I sometimes have a hard time telling the difference between 720 and 1080, but 1080 to 4k is like night and day. Almost 3D without any actual "3D". Once people see that in Bestbuy or the like, I believe it will stop the "race to the bottom" in terms of media quality. Almost like a tipping point where the average joe is now a videophile because he wants that clarity and amazing image that 4k provides and the piss poor netflix stream just won't cut it anymore.

Also, I'm not sure what you're on about, but going from 2k to 4k, assuming proportional bitrates, will provide a bigger improvement to PQ than anything else will. It's what everybody mentions because that is what will provide the biggest leap in PQ, not increasing the color space of whatever else you had in mind, although, that would be nice as well.

Though if the TVs get bigger and higher resolution, and they are fed 4K/24p, the strobing due to 24 fps will get worse, which in that way will be a decrease in PQ.

Still again, for you and the other guys, since you are missing the point: There is no doubt single chip has tons of advantages and as I said the 3-chip is becoming dinosaur. This however does not mean - if we are talking about stratospheric aspects of 4K res in which people here, trying to prove their point, show zone images with slight aliasing in the end region of vertical frequencies - that the arrangement of sensor matrix in which subpixels are perfectly aligned is not superior. It is evident that it is superior since Bayer sensors have an inherent error in creating pixels. Bayer is simply good enough. As I said, but people are blind to it, the Foveon sensor has many problems but it was shown that with respect to resolution and detail rendering it is clearly superior to Bayer /in normal light conditions/.

Eight megapixels. That's all 4K is.

People's phones can create 8MP images now. There is nothing "stratospheric" about it.


And yet the highest-end cameras available today, are all using a single chip design. Why is that, I wonder?


The highest resolution Foveon sensor available today, is 15.36MP. That pales in comparison to the resolution of "inferior" bayer sensors. (such as the new 36.8MP Nikon D800)

You do not follow precisely my text: 'stratospheric' was used in the context of the 4K vs. 2K debate here. For proving one camera is better than another somebody used zone images which were showing slight aliasing in the extreme range of vertical frequencies. Comparisons of this kind are beyond the atmosphere, they are surreal if one takes how the video is processed and watched by consumers.


Here one can see your lack of understanding of even elementary basics so no wonder there is big waste of time in this discussion. Foveon is 15 megapix BUT these are FOVEON FULL RGB PIXELS. Don't you know that in Bayer they count each R G B subpixel as pixel ? One could say then that Foveon is 45 megapix in the Bayer sense but this is not the point. The point is that in the Foveon, due to its layered structure, subpixels of a pixel are taken from precisely identical spatial locations. In Bayer this is not the case, Bayer recreates RGB positions by interpolation. Now you can say the resulting error is small. However, if one is talking about the stratrospheric differences between the different 2K/4K cameras and 3-chip vs. 1-chip designs this can not be neglected. The only way for Bayer to reduce its inherent error is by increasing pixel density, this why Red is using 20 megapix chip to get 4K.

They are not 45 megapixel equivalent in all three channels.


And it is due to this spatial difference that Foveon's 15 megapixel sensor is not at all equivalent to a 45 megapixel bayer camera.


But oversampling has benefits far beyond resolution. You greatly reduce the amount of pixel-level noise in the image for starters.

In a world where sensor quality between three chip cameras and single chip was identical, there was no degradation from splitting the image to three sensors, you were still oversampling with a three chip design, and you didn't mind paying three times the price, sure it makes sense. But that is not the world we live in. There is considerably more expense with three chip designs, and it compromises image quality.

I thought with more pixels in the same area, you reduce light sensitivity too, which means more noise unless you use more lighting? Or are you talking about bigger sensor areas? Or would the pixel level noise when downsampled to the original res be less than using the same size sensor and bigger pixels?

This is only true to a point, modern sensor design still manages to show improvements as resolution increases. We aren't talking about phone-sized sensors here. (though even they are improving as resolution increases)

Pixel-level noise may increase as resolution goes up, but image noise goes down.

Seeing actual 4K is a WOW!!! moment much more impactfull then seeing 720p or 1080p video.

Its a rock your socks moment when I said to myself imagine all tv shows, movies and video games in 4K and my face had a WOW!!! moment like this

For me 4K video is a miracle that sooner or later will become the defacto standard and if it takes a few years to get here then so be it.

Irkuck:

These are the only things that I care about for cameras:
Cheaper, faster and better and Red camera has that and next six months we get a taste of a better and improved camera and every year the tech improves!! and your point??? becomes what??? a smaller non issue than it is this year.

I drool for Epic Red cameras.

And where have you seen these 4K sources? Not compressed to a BD, but only in demo form showing live video? Yeah those are made to look the highest quality. Not saying 4K won't look good but i wouldn't really judge anything based off those. They're running of hard drives probably uncompressed with bitrates we will never see.

And you have waaaaaaaay to much faith in the average consumer. The consumers who since forever don't care for good quality anything and let garbage succeed. The consumers who have yet still to fully embrace bluray. These consumers are somehow going to fall for 4K and turn the tides for quality and become videophiles????????????

Excuse me, i laughed my ass off while i typed that.

http://www.youtube.com/watch?v=uNHFY...layer_embedded

How's that for timing? Nokia's new phone uses a 41MP sensor:

http://www.tested.com/news/nokia-pur...e-photos/3647/

Edit: more samples: http://www.gsmarena.com/nokia_808_pu...-news-3905.php

"What the hell are we going to sell them now"
I heard about that when the big deal was going from mono to stereo and in receivers from stereo to Pro logic.

The big deal was purchase a pro logic.
Then it was more power on the front channels and guess what it was buy an upgrade.

The big deal was equal power on all pro logic channels and the big deal was buy an upgrade.

The big deal was look! Dolby Digital! Upgrade NOW!!
And an upgrade to 5.1 DD was done

The big deal was upgrade to DD 7.1
And folks are starting to do that.

In HDTV sets the
"What the hell are we going to sell them now"

The big deal was upgrade to 720p
Then the big deal was upgrade to 1080p

"What the hell are we going to sell them now"

Is going to be 2160P

Then its going to be 4320P

To think that CE Producers are just going to abandon?? BILLIONS of DOLLARS invested to produce HDTV sets is infantile, naive and contrary to historical precedent.

"What the hell are we going to sell them now"

Thats the operative sentence for these guys and folks that produce cars, planes, motorbikes, helicopters, jets, guns, amplifiers, HDMI cables, speakers, refrigerators, etc.

Get the picture??

It is not so that you can always push anything onto consumers. People were trying to push 24-bit audio with miserable results and even Blu-ray had a bumpy road still far far from replacing DVD. Consumers buy new tech if it offers some useful difference or if the price is same as the old tech. 4K will not offer useful difference and there will be no content for it. The only possibility then is the same price. This is possible in long run but in the beginning the price will have to be higher. With companies too deep in red to subsidize prolonged introduction of new tech, and consumers indifferent to the PQ there is every condition satisfied the 4K will backfire.

In turn, 45 megapixel bayer is not a 45 real megapixel camera. Real pixel is composed of 3 RGB subpixels taking light from the same spatial position. In Bayer this is not the case and it is a convoluted design: there are four RGGB subpixels which could be roughly taken as corresponding to one real pixel. But real pixels are reconstructed in Bayer from overlapping RGGB matrix. Thus, in terms of real resolution Bayer 'pixels' should be divided by 2 in the most optimisic case and by 4 in the pesimistic case to correspond to the real RGB pixels. Foveon 15 megapixels are real RGB pixels, Nikon 36 bayer megapixels correspond then at best to around 18 real RGB megapixels and still there is an error from the creating of those pixels. Now, this error is practically small but in comparison with Foveon it can be seen. When talking about esotheric differences between the 3-chip and 1-chip designs and 4K vs. 2K this error can not be neglected.



First point is that you say oversampling (...with Bayer) is is reducing pixel level noise which is true to some extent - one has to include the sub pixel size too. Second, you notice correctly that 3-chip also makes oversampling plus subpixels are taken from the same spatial position. Thus, when comparing 3-chip vs. 1-chip and 4K vs. 2K, and the 2K decimated from 4K vs. native 2K one has to be very careful: Bayer will get equivalent to 3-chip with much higher number of (bayer) pixels. No question though Bayer is cheaper and will replace the 3-chip.

That is not correct at all. You have 36 megapixels worth of spatial resolution spread across the three colour channels, because each is taken from a different sampling point. With Foveon, you only have 15MP because all three channels are taken from the same sample location. (because it's a layered design)


A single chip 4K camera is going to be far higher resolution than a three chip design, where you cannot afford to oversample as much due to the fact you have three sensors in the camera.



So do you now agree that it is best to over sample? Earlier on, you were saying that you should be capturing at your target resolution, which was clearly wrong.

Wow, as to the technical, one side must be VERY WRONG in this thread. I have no idea which.

It does occur to me that it took only a few weeks (or was it months?) before 1080p became "meh". Certainly still nice and all, but the newness wore off quickly, the thrill gone. It became just the new normal. Sometimes I would even try to pretend to myself that I was seeing 1080p for the first time all over again, to try and recreate the initial fascination. But that never really worked. I'm sure the same will occur with 4k, 8k, etc. Those super duper resolutions will also quickly become no big deals after just a few weeks in use.

So I guess it all circles back to content quality, which by far takes precedent over screen eyecandy. soprano, you are excused here of course.

Sure, but this thread is about 4K which is why it is the main topic. If I was to list the picture quality improvements in order of what I would prefer for consumer video they would be frame rate (60 fps), resolution (2160p), color depth (10-bit), and getting rid of color subsampling. As such I would love to see 60 fps used in major movies but as far as I know the only major movies in production that use higher frame rates are The Hobbit movies and they are at 48 fps (which is a good choice for movie theaters but a horrible choice for 60 Hz displays). There is no question that there are other improvements that can be made to picture quality for consumer video but for 24 fps video I think increasing resolution is currently the most noticeable improvement that can be made.

Ough, thie enters stage of sickness. Pixel evidently means light at different wavelengths is measured and collected from single spatial location. The layered design of Foveon is precisely creating pixels. Bayer takes measurement of different wavelenghts from different spatial locations. There is thus an inherent error in the creation of pixels from the Bayer sensor. Saying that Foveon has 15 mp and Bayer 36 mp is sick. Difficult to imagine how one can come to such idea.


Another nonsense. A perfect 4K 3-chip camera would be made of 3 4K chips, one separate for each RGB channel. In total 12 mpix of RGB subpixels. These subpixels would be perfectly aligned collecting light from the same spatial position. Equivalent 4K Bayer camera needs more than 20 mpix. But of course Bayer is cheaper.


Since you lack basic I can tell you only it is not so simple as oversampling or not. What is of primary importance is proper sampling and bandlimiting of the signal. When these are done badly oversampling is no cure, it may be even worse. One can say that with proper system designe there is absolutely no need for oversampling. What you read in Red PR/marketing materials is directed to those who are ignorant. Ther write about oversampling since the Bayer sensor requires much higher pixel count to accommodate for its inherent error. Then, by digital manipulations they recreate original signal. Instead of writing this, they say that they do 'oversampling' to get better pics.

HEVC, BDXL, etc. will ensure that 4k video will look amazing. There isn't that much difference between a 1080p bluray and a 2k DCI copy using JPG2k compression.

As far as consumers, you completely ignored my argument. 4k just blows me away, something that 480p, 720p or 1080p, 3D, etc. have never done. It may be the tipping point as I said. It is to me what I think the industry wanted 3D to be, and that is that "I must have that and I'm going to do everything in my power to get it". It is also for that reason that I think 4k will be shoved out faster than everyone is expecting, media that is. Once people see the hobbit or avatar 2 4K@48fps, they're only going to want more, something that crappy compressed cable or netflix simply can't provide. Sort of like if a blind man were given sight for a day. He'd no longer be satisfied with being blind.

Personally, 3D was one of those "must have it" things for me. But so was 1080p, and so likely will be 4K. And of course 4K 3D @ 48fps

Exactly

dvd was introduced in North America in 97, overtook vhs sales in 2003, so 6 years.
D-Theatre started selling 1080 tapes in 2002, I think.
BD was introduced in 2006 and 6 years later (well, closer to 5 1/2) , is no where near overtaking dvd sales. Sales are still roughly 3-1 dvd-blu ray, are they not? Up perhaps 5% in the last year?
There was a LOT of expectation by the movie industry concerning the success of BD, and while BD is successful, I think most would agree that after 6 years it has fallen short of expectations. Consumers, didn't seem to get the WOW factor about the quality difference that we were expecting (we being us videophiles), so they weren't as interested in buying BD as had been thought.

The point is, I think the average consumer will find the bump from 1080 to 2160 even less compelling. I know some of you disagree, but I think you are overly optimistic.

To the best of my knowledge no large company or consortium has announced any plans to market a 4k player or provide downloads/streaming, and no studio or consortium has announced plans to release movies in 4k.

Because BD didn't live up to expectations, I don't think either the CE companies or studios are all that interested in selling 4k to the public.

Personally, I want a 4k projector, and I'd love to get 4k media, so, while I might share your enthusiasm for 4k, I don't share your optimism about its release being "right around the corner".

In fact, I'll state my opinion now that no consumer 4k distribution channel will exist for any major studio by the end of 2013. Further, if any studio is distributing 4k, available to the masses, by the end of 2015, I'll say that it will a niche product only, with media sales lower than what D-theater sold, and with fewer big studio releases. D-Theatre did a little less than 50 according to this
I'll speculate that there might be a consortium created by the end of 2015, but not by the end of 2013, with a plan to sell a mainstream 4k player, but no actual sales. When I say a plan, I mean something concrete about format, codecs, distribution media or method, protection etc. Not just "were going to make a 4k optical disk".
I'll also speculate no mainstream distribution with, say 3 major studios on board, before the end of 2016. By this I mean no actual movies from a mainstream format make it into the hands of consumers before the end of 2016.

But hey, that's just me.

And don't ask me to bet. I'm not going to.




Edit:
Also, I don't think you'll see any announcements about moving to 4k broadcast OTA before the end of 2013, and no actual OTA broadcasts before the end of 2016, unless it is some small pilot project.

I still doubt we will actually be able to buy a 4K Sharp in 2012, but hopefully the Elite line will get 4K tech for their 2013 models. Especially since it looks like the Sharp 925 series is going to get most of the current Elite internal hardware.

http://www.twice.com/article/481186-...r_Big_LCDs.php