Record uncompressed 14-bit (4:4:4) on your HVX200

So much baseless speculation over only one post... this guy hasn't even proven that the e-mail actually came from Juan. Of course I won't deny the possibility, but I'd very much like to see some e-mail headers, or a screenshot of his inbox before I even consider the likelihood of his claims. URL redirection + silence ≠ proof of anything. There could a multitude of reasons for both; including, but not limited to: site maintenance and being busy.

Now, like everyone else, I've been waiting to hear something about this for a while now, so while I understand the anxiousness I don't understand why we're counting chickens before they hatch.
 
Why are we counting our chickens before they hatch? Because I'm betting that these chickens will NEVER hatch. I don't expect we'll ever get a formal announcement from Juan at this point. It is dying with a whimper. . . like an old soldier. . . just fading away.

But, oh, what a loss I think this might have been.

Stephen
 
Stephen Pruitt said:
Why are we counting our chickens before they hatch? Because I'm betting that these chickens will NEVER hatch. I don't expect we'll ever get a formal announcement from Juan at this point. It is dying with a whimper. . . like an old soldier. . . just fading away.

But, oh, what a loss I think this might have been.

Stephen
Click to expand...
So we're counting chickens based on a bet?
 
But ... why? Why dance on its grave or trumpet it? Why not just acknowledge that we don't know what's happening, and that we're waiting for more info? Why are people betting that it'll die? Makes no sense...
 
well.. I wasnt going to join in.. but oh.. what the heck

Here goes:
If I was going to pick some BIG company it wouldnt be Panny or Sony.. it would be.. you ready ..it would be...

APPLE!

This technology would be right up their alley. Especially if you need a MAC to use it, plus throw in the whole FCS studio, ProRes, Color, custom Macbook Pros.. man.
It would put Apple in a most interesting position right out of the gate especially if its something that could be put to use on more than one brand of camera. They would OWN the low budget Indie Film market~
Also, think about it, Apple is known for its "quietness" before they release something big.

Ok, ok, sorry about that. Im not big on conspiracy theories and all... but that would make for a nice conspiracy wouldnt it ;)
 
focus enhancements

hydra -ethernet -portable HD

all integrated without a laptop

actually it would probably have to run off ram for those 90+ megabyte per second transfer rates.


so you would record say 5 minutes of 2k video to ram buffer and then stop recording and transfer to a hard disk. + a sculptor 2k like gui.
 
Last edited:
killfilm said:
Barry_Green said:
Hydra promised 4:4:4 at standard-def, and we don't know the exact color sampling at HD or 2K res. It offered uncompressed 2K at variable frame rates, and uncompressed 1080P at variable frame rates up to 60fps.
Click to expand...
Juan did announce 4:4:4 at 2k earlier on.
Click to expand...
Barry's right: at 2K, the color is significantly less than true 4:4:4.

I e-mailed Juan about this a couple of years ago with respect to Andromeda. His reply said the following: "At 1.5k resolution, Andromeda works pretty much like a 1-chip HD camera, such as the Kinetta or Arri D-20. However, on a one-chip camera which uses a mosaic sensor, it is technically impossible to sample a real 4:4:4 image because since there is only one chip, each pixel element of that chip captures only ONE color, either Red, Green or Blue but not more than one. Despite this fact, Arri and Kinetta for example, advertise 4:4:4 chroma sampling in their specs. I personally don't feel we should advertise it in this manner for the reason above. Nailing down the Chroma sampling of a mosaic sensor is tricky, because chroma sampling is listed in a Y:U:V space where most mosaic sensors use RGB color filters and thus the output is in RGB space values. Once the missing color values have been filled in (de-mosaicing) then the image is full color (4:4:4), but it wasn't ~sampled~ that way."

In a follow-up e-mail, he said: "HD modes are only available when ALL the color data is recorded from the 3 CCD's, which is what 4:4:4 recording mode refers to. This is because all the data from the CCD's is needed to build a good HD mosaic. Even though the resulting demosaiced image is full color uncompressed, and other companies would list this as 4:4:4, I will tell you right now that neither the HD coming out of our system nor theirs really IS 4:4:4. The only way to obtain full color sampling is to have three sensors, or a single Foveon sensor, to the best of my knowledge. I can also tell you now, the same Andromeda concept applies to the HVX."
 
...and all the little Andomeda's were tied helplessly to the operating tables somewhere in Area 51. If I remember right I think it was in hanger 18.
 
karapetkov said:
Can't you extract it from the Redcode Raw files, using RedCine?!

Please clarify, or give a link to a thread that treats this.
Click to expand...
Tzedekh's post clarifies a lot.

Simple fact is: the Red sensor is capable of 4k in black and white, but not in color. No single-chip system (other than the Foveon) could sample full resolution in color. Sensor pixels are monochrome, not color! You have to combine multiple pixels together to build a color pixel.

Typically a Beyer pattern sensor will deliver about 70% of its monochrome resolution once it's been demosaic'd into a color signal. Red doesn't deliver a final 4K, it delivers somewhere around 2.7K - 3K of resolved image resolution. Its sensor is 4K, its frame size is 4K, but its resolved image detail is maybe close to 3K.

If any system wants to offer full 4:4:4 color at full sensor resolution, they'd *have* to have three sensors. Or they'd have to have dramatically more black 'n' white pixels, about 40% more, so that the resulting demosaic'd and scaled-down image would have full color.

So -- if Red offered a scaled 2K mode, or you downrezzed to 2K from the RedCine application, you'd probably have full 4:4:4 color at that point and full 2K resolution. But you're not going to have 4:4:4 in 4K mode -- it'd require three sensors to get that. Or maybe you could get close enough with a 6-megapixel sensor.

This is not a "knock" against red, it's just simple sampling math. SI-2K has the same issue, they're a single sensor and therefore they can't have 4:4:4 at their 2K resolution. Dalsa's the same way. It's just not possible.

Foveon could get there someday, but Foveon is really sort of three sensors all wedged together so that makes sense.
 
Thank you for the explanations.

I have a kind of an absurd idea concerning the philosophy of the Foveon sensor.

Now, if the three-layer architecture of the Foveon's pixels is [theoretically] the closest to the way film captures color, wouldn't a similar kind of architecture be applicable to eventual future displays\projectors:

One visible pixel whose appearance and color reproduction is actually produced by three [red, green and blue] pixels, underlying each other. A display with three underlying layers of pixels, not just one.

I know that it would be a bit impossible physically, cause the upper [frontal] layers of pixels would mask the bottom [back] ones. But if these layers [at least the upper two - green and blue], can have levels of transparency, it would become possible.

Suppose that a single [three] pixel of the Foveon sensor captured mainly red signal, but it also caught some green and\or blue. It may be little, but not none.

This information is then fed to the three-layer display where we have our three-level pixel displaying it 1:1.

Say the red sub-pixel would be the most active, we'll have a little green on the green sub-pixel, and the blue one would be almost completely transparent.

Wouldn't that be closer to film acquisition and mode of display.

:violin:
OK, having read what I just wrote, I admit that it sounds very sci-fi, but, doesn't it make sense?

:Drogar-Shock(DBG):.
 
Barry_Green said:
If any system wants to offer full 4:4:4 color at full sensor resolution, they'd *have* to have three sensors.
Click to expand...
Are there any systems like this at all? Why don't we see 4:4:4 in more 3-sensor set-ups?

Or they'd have to have dramatically more black 'n' white pixels, about 40% more, so that the resulting demosaic'd and scaled-down image would have full color.
Click to expand...
But aren't all demosaic operations merely interpolating color values? How would this represent true color? Not contesting your point; just curious.

So -- if Red offered a scaled 2K mode, or you downrezzed to 2K from the RedCine application, you'd probably have full 4:4:4 color at that point and full 2K resolution. But you're not going to have 4:4:4 in 4K mode -- it'd require three sensors to get that. Or maybe you could get close enough with a 6-megapixel sensor.
Click to expand...
So does the Red actually offer 4:4:4 in sub-4K resolutions without downrezzing, or are they all similarly-affected?

Foveon could get there someday, but Foveon is really sort of three sensors all wedged together so that makes sense.
Click to expand...
How long do you suppose it'll be before we start seeing these things in camcorders?
 
karapetkov said:
Now, if the three-layer architecture of the Foveon's pixels is [theoretically] the closest to the way film captures color, wouldn't a similar kind of architecture be applicable to eventual future displays\projectors:

One visible pixel whose appearance and color reproduction is actually produced by three [red, green and blue] pixels, underlying each other.

I know that it would be a bit impossible physically, cause the upper [frontal] layers of pixels would mask the bottom [back] ones.

OK, having read what I just wrote, I admit that it sounds very sci-fi, but, doesn't it make sense?
Click to expand...
It's not sci-fi at all, and makes perfect sense. It would be the electronic equivalent of transparent (positive) film.

BUT the layers of pixels would need to have complementary colours, not primaries, exactly the same as the dyes in a processed film emulsion - so magenta, cyan and yellow, not red, green, blue. In practice, I believe it is much more difficult to do it this way for colourimetry reasons, and separate (smaller) red, green, blue pixels give better results.

And it's apparently similar with Foveon. Nothing wrong with the idea in theory - but in practice the theoretical advantages of larger three colour pixels are offset by losses in the deeper semiconductor layers. Practically, it's better to have more, but smaller, single colour pixels than fewer, larger, Foveon ones. If Foveon technology got more efficient, that balance would obviously change.
 
Barry_Green said:
Simple fact is: the Red sensor is capable of 4k in black and white, but not in color. No single-chip system (other than the Foveon) could sample full resolution in color.
Click to expand...

So why is nobody taking a Foven Sensor camera (like my beloved Sigma) weld some wires at the sensor output and connect the whole shbang to a laptop with andromeda like software?

I guess it is not that easy. I think the mecanical part (grabing the image right after the sensor) could be done by any skilled TV techican, the problem is the software.

Frank
 
You must log in or register to reply here.
Back
Top