The idea: a hybrid foveona and buyers

Pages: 1 2

28.10.2015 17:46:00
Clearly shown in the figure. The idea is to make
1. values ​​in the matrix color change is smaller in absolute value = & gt; not increase so much noise (in comparison with foveonom) at the transition from sampling to RAW sRGB color
2. 3. simplify manufacture
having eight instead of three color components, can improve the accuracy of color
 

28.10.2015 18:15:00
For you to do "guy" pictures with an interval of 1. 5 year (one at ultrazoom, the other on a fix) and pull the article 20 years ago, I did not expect. . .

28.10.2015 18:40:00
quote:
:
on you to do "guy" pictures with an interval of 1. 5 year (one at ultrazoom, both at fixed) and pull the article 20 years ago, I did not expect. . .
they do not paired . Just one story. There was no disk imaging it is paired. This I have played up. foveon Night, and evening, and everything turns out better detail than Baer. And the night is especially beautiful.


SDIM1430. jpg


SDIM2455. jpg

28.10.2015 18:44:00

I think 20 years ago in some professional camcorders Panasonic has a prism that divides the light beam into three. And there were 3 matrix, red, blue and green (for filters). So everything has already been invented.

28.10.2015 18:45:00
quote:
:
quote:
:

This topic your "masterpieces" - ottfop stop spam. To do this, you have its own theme.

quote:
:

I think 20 years ago in some professional camcorders Panasonic has a prism that divides the light beam into three. And there were 3 matrix, red, blue and green (for filters). So everything has already been invented. Well see
flaw need 3 matrix. And there are three shutter curtain? + Shaking. + Price. + Size. + Adjustment.
+ presence limits the prism optics and makes three-chip systems are incompatible with conventional lenses (they are in these cells is usually non-removable). + separations
not something I'm talking about a single matrix, and used with conventional interchangeable lenses.

29.10.2015 9:46:00
quote:
:

Patent Samsung saw?
http: // image-sensors-world. blogspot. co. uk / 2015/10 / sa ... color-sensor. html

US20150286059 seen. This kaku is impossible to produce!
US20150286060 more or less realistic for manufacturing, but still difficult, and it is working for different positions and size of the exit pupil is a different (color rendering will vary depending on the aperture value!!).

29.10.2015 10:00:00
quote:
:
I think 20 years ago in some professional camcorders Panasonic has a prism that divides the light beam into three. And there were 3 matrix, red, blue and green (for filters). So everything has already been invented.

fact trёhmatrichnye Panasonic camcorders and available, and there are no filters, just a prism decomposes light. But there is only the matrix 1/3 "with a small resolution for a camera with a larger matrix size to grow strongly. And a few years ago Panasonic patented matrix of microprisms, three under each photocell, but has not yet been implemented.


you do not understand principle of the matrix Foveona, a thin layer on top of a missed almost the whole world.


29.10.2015 10:25:00
quote:
:


you do not understand the principle of operation of the matrix Foveona, a thin layer on top of a missed almost the whole world.
No. In Foveona thickness of the layers is very different, because

and thinner layer will effectively absorb short-wavelength photons. So it is necessary to later in the conversion to sRGB subtracted less correlated signal.

29.10.2015 10:31:00
quote:
:
clearly shown in the figure.
color separation different absorption depth in silicon, unlike foveona layers 2, not 3, with the boundary between the upper and lower layer of variably changing mosaic of 2x2 (as the color filters at Baer).
The idea is to make
1. values ​​in the matrix color change is smaller in absolute value = & gt; not so much the noise increases (compared with foveonom) samples during the transition from RAW to colors sRGB
2. simplify the manufacturing
3. with 8 color components instead of 3, you can improve color accuracy
Silicon is much worse than the color of shares in a range of customized deliberately "colored glass" Bayer. Again, the color will float depending on local heating of the matrix, ie. K. A geometric thickness pixels changes with changes in temperature and noise. How to deal with this plan?

Simply fill in the matrix MP 200-300 Bayram and everything. This is possible now. The total area will not change, but the resolution will increase.

29.10.2015 10:38:00

clearly shown in the figure.

other words, the output will (B + G, R) (B, G + R). E. We obtain the same reduction in color resolution Baer and as a bonus get Foveona disadvantages?

29.10.2015 10:39:00
quote:
:
Simply fill in the matrix MP 200-300 Bayram and everything. This is possible now. The total area will not change, but the resolution will increase.

So technology has long worked under such a density, the main thing that did not declare the full resolution, and made in the chamber 4 times resize for "pixel" color
Assumption Cathedral 12 MP:

quote:
:
And the thinner layer will effectively absorb short-wavelength photons. So it is necessary to later in the conversion to sRGB subtracted less correlated signal.

Fauverney have the thickest layer of the upper, it absorbs the waves (and less long, but still too much). If you make the top layer is thinner, not quite tsvetorazdeleniya.

29.10.2015 11:13:00
quote:
:
quote:
:
and thinner layer will effectively absorb short-wavelength photons. So it is necessary to later in the conversion to sRGB subtracted less correlated signal.
Fauverney have the thickest layer of the upper, it absorbs the waves (and less long, but still too much). If you make the top layer is thinner, not quite tsvetorazdeleniya.
https: // upload. wikimedia. org / wikipedia / commons / thumb / a / ab / Absorption-X3. png / 500px-Absorption-X3. png
and thickness of layers in foveone chosen not to make the same sensitivity, and most of the other criteria.

quote:
: Again, the color will float according to the local heating of the matrix, ie. A. Geometric thickness of the pixels varies with temperature and noise. How to deal with this plan?
How should change the temperature to a 3 micron layer thickness increased at least 0. 1 micron? Well something can and will change. Fighting Software.

quote:

Simply fill in the matrix MP 200-300 Bayram and everything. This is possible now. The total area will not change, but the resolution will increase.
Not so this is just 1. 5 Now all crop still on the matrices with front-illuminated. And a decrease in the pixel is not to get rid of the lack of that filters consume about 50% of light.

quote:
:

clearly shown in the figure.

other words, the output will (B + G, R) (B, G + R). E. We obtain the same reduction in color resolution Baer and as a bonus get Foveona disadvantages?
No, the output received eight linearly independent color components. We call them
B1 Y1
B2 Y2
C1 R1
C2 R2

to get sRGB. G = Y2 + C1 - B1 - R2
unlike foveona we subtract only the very small signals and increase the noise does not occur.
compared with buyers: we have an honest luma without moire (obtained by folding the top and bottom layers) better b / w sensitivity
compared with fovenom: easier to manufacture, have the color at high ISO
color resolution less brightness, well, to hell with him.

29.10.2015 15:14:00

unlike foveona we subtract only the very small signals and increase the noise does not occur.

Here is incorrect. The more components in the formula - the more noise in the final signal, ie. To. Will form part of the noise. In addition, it turns out 1 green 4 point? And despite the fact that green - the color of the most critical?

29.10.2015 16:04:00
noise increase depends not only on the number of terms, and also in front of them by the coefficients. If they add up - so only reduces noise from it. We have to deduct foveona adjacent signals that are heavily zakorrelirovany. One of the authors foveona therefore wrote that the layers, would be more desirable to do (how many, I do not remember, it seems 6). But many layers difficult to produce - they even manufacture 16x3 MT had problems, however, they decided they did.

29.10.2015 18:23:00
quote:
:

quote:
: Again, the color will float according to the local heating of the matrix, ie. A. Geometric thickness of the pixels varies with temperature and noise. How to deal with this plan?
How should change the temperature to a 3 micron layer thickness increased at least 0. 1 micron? Well something can and will change. Fighting Software.

quote:

Simply fill in the matrix MP 200-300 Bayram and everything. This is possible now. The total area will not change, but the resolution will increase.
Not so this is just 1. 5 Now all crop still on the matrices with front-illuminated. And a decrease in the pixel is not to get rid of the lack of that filters consume about 50% of light.
.
At 15 nM for a variation of 10 degrees C. Moreover, as the layers of different thickness and float on the temperature will in general. While the entire visible spectrum of eye fits into the wide range of 250nM. The shift in the spectrum of the wavelength of 15 nm, as it were already clearly different color. And it will look like a big block noise.

50% is just 1 stop. You want to say that the noise is less than Bayer foveon 1 stop?

29.10.2015 22:00:00
citation:
: At 15 nM for a variation of 10 degrees C. Moreover, as the layers of different thickness and float on the temperature will in general. While the entire visible spectrum of eye fits into the wide range of 250nM. The shift in the spectrum of the wavelength of 15 nm, as it were already clearly different color. And it will look like a big block color noise.
and what have nanometers wavelength in nanometers thickness of silicon? ? ?

quote:

50% is just 1 stop. You want to say that the noise is less than Bayer foveon 1 stop?
I do not know exactly how many. If you say sprinkled on Canon ISO 800 instead of 100 and put the hold on the converter, you get something better and something foveona. . . And then we got into the habit to compare crop camera 2012 with a fresh FF.
I once read from Tutubalina study it dies shot on foveon, he expressed the idea that the gain in these cells as such, no. Part (some unclear) noise comes not because of the color conversion, and the noise of the electronics.

But to have one foot ISO, even in black / white, we would like.

29.10.2015 23:02:00

well loaded . . Yes, everything is already in foveone what else is needed? I would only count as MP increased to 30 on each layer.

29.10.2015 23:08:00
Well, not everything. . . I've personally not happy, I do not know, there is an orange or lemon, is it not said who saw the scene live
number of MPs to 30, and the layers - up to six, and then think about how to make such a miracle .

30.10.2015 3:07:00
quote:
:

quote:

50% is just 1 stop. You want to say that the noise is less than Bayer foveon 1 stop?
I do not know exactly how many. If you say sprinkled on Canon ISO 800 instead of 100 and put the hold on the converter, you get something better and something foveona. . . And then we got into the habit to compare crop camera 2012 with a fresh FF.
I once read from Tutubalina study it dies shot on foveon, he expressed the idea that the gain in these cells as such, no. Part (some unclear) noise comes not because of the color conversion, and the noise of the electronics.

But to have one foot ISO, even in black / white, we would like.
Foveon more noise Bayer same size of 4-5 stops, so anyone and not interesting in terms of versatility and color.

quote:
:
Well, not everything. . . I've personally not happy, I do not know, there is an orange or lemon, is it not said who saw the scene live
number of MPs to 30, and the layers - up to six, and then think about how to make such a miracle .
15 megapixel version (to measure) was already 5 layers. In connection with eerie blooming layers required to isolate the intermediate, which further reduces the already stunted color separation.

30.10.2015 7:29:00

noise increase depends not only on the number of terms, and also in front of them by the coefficients. If they add up - so only reduces noise from it.

No noise reduced only if the components are formed of the same signal, but not different. E., The noise is always increasing, the only question is how much correlated folding useful signals. In other words, in the case of noise Baer = 1, and in the case of this hybrid noises are n & gt; 1. And if the signals are highly correlated, it sounds a little bit and reduced, but for the color and then still have to use the big factors. And if the signals correlate poorly - it just increases the noise.

If you say sprinkled on Canon ISO 800 instead of 100 and put the hold on the converter, you get something better and something foveona. . . And then we got into the habit to compare crop camera 2012 with a fresh FF.

Canon this matrix long before 2012 developed While it may still to ISO 800

number of MPs to 30, and the layers - up to six

and ISO 10 at f / 0. 5 And not to forget it, even a bright sunny day will be like a coloring book.

30.10.2015 8:02:00
quote:
:
If you say sprinkled on Canon ISO 800 instead of 100 and put the hold on the converter, you get something better and something foveona. . .
Yes? I experimented with laytrumom, Nikon D5100, 14-bit RAV. Located mirror for easy viewing.

30.10.2015 15:55:00
All this nonsense.
Engineering monochrome color reproduction matrix and filter change, or if no mechanics - matrix with its filters and / or a desired color response.
For art - modeling the human eye. Many monochrome sensitive rods and cones of colored (foveon - error).

30.10.2015 16:01:00
quote:
:
quote:
:
If you say sprinkled on Canon ISO 800 instead of 100 and put the hold on the converter, you get something better and something foveona. . .
Yes? I experimented with laytrumom, Nikon D5100, 14-bit RAV. Located mirror for easy viewing.
so ETOGES sonevskoy matrix with Nikon and Canon do not

30.10.2015 16:04:00

I once read from Tutubalina study it dies shot on foveon, he expressed the idea that the gain in these cells as such, no.

in Merrill - no (remove during a single exposure, and various ISO, and see for yourself that just "thought")

The Quattro - is.
 

30.10.2015 17:17:00
quote:
:
all this nonsense.
Engineering monochrome color reproduction matrix and filter change, or if no mechanics - matrix with its filters and / or a desired color response.
For art - modeling the human eye. Many monochrome sensitive rods and cones of colored (foveon - error).
The Bayer is not happy?

30.10.2015 18:48:00
quote:
:
quote:
:
quote:
:
If you say to the sprinkled Canon instead of 800 ISO supply 100 and hold on in the converter, you get something and not better foveona something. . .
Yes? I experimented with laytrumom, Nikon D5100, 14-bit RAV. Located mirror for easy viewing.
so ETOGES sonevskoy matrix with Nikon and Canon do not

Given that the experiment povedёn me during the development of Nikon and Laytruma 2. 5 years ago. Now I would have done better (prettier), but it's useless, since the aim was to ascertain the presence of such an effect.

30.10.2015 22:56:00

Simply fill in the matrix MP 200-300 Bayram and everything.

very true thought. That "all" - and neither you false colors or moire of all types and higher resolution.
Actually it's do, but one lost sheep, and she had turned polubayer

31.10.2015 0:10:00
quote:
:
Well, not everything. . . I've personally not happy, I do not know, there is an orange or lemon, is it not said who saw the scene live

This curve BB and everything. We need to put right. I have something still in place.


SDIM4628. jpg

31.10.2015 1:07:00
quote:
:

unlike foveona we subtract only the very small signals and increase the noise does not occur.
compared to Bayer: have honest luma without moire (produced by folding the top and bottom layers) better b / w sensitivity
compared with fovenom: easier to manufacture, have the color at high ISO
color resolution less brightness, well, and to hell with him.

luminance moire foveone is on, there is no color. To win at Bayer color moire, perform only one action, including color squelch to the desired radius exactly the same way as is done in the foveone (blur color for interpolation of color noise).

quote:
:
quote:
:
Well, not everything. . . I've personally not happy, I do not know, there is an orange or lemon, is it not said who saw the scene live
This curve BB and everything. We need to put right. I have something still in place.
You might, but in your pictures - tin.

31.10.2015 1:15:00

For art - modeling the human eye. Many monochrome sensitive rods and cones of colored
When enough light for the cones, Rod signal is tightly locked.

31.10.2015 9:05:00
quote:
:
luminance moire foveone is on, there is no color.
luminance moire does not cause any problems. Moreover, the luminance moire may occur in most of the scene.

quote:
:
quote:
:
Well, not everything. . . I've personally not happy, I do not know, there is an orange or lemon, is it not said who saw the scene live
This curve BB and everything. We need to put right. I have something still in place.


SDIM4628. jpg
pictures here you post, and this topic should not be. MotoAlex made a pair of still life pictures on dp2m and d750. Nikon to get less noise, but the Mandarin turned like a lemon that MotoAlex predpodnes as lack foveona, until he discovered that it is not really a lemon and tangerine. E. Foveon worked better color accuracy. I thought you zlodarstvovat ))) but no. strange.

31.10.2015 11:05:00
quote:
:
quote:
:
luminance moire foveone is on, there is no color.
luminance moire does not cause any problems. Moreover, the luminance moire may occur in most of the scene.
It would be interesting to look, never seen (I'm not Lanselot), except if it is a snapshot of the image with moire.

31.10.2015 11:29:00
you have a small kitchen strainer on there?

31.10.2015 11:41:00
quote:
:
you fine sieve in the kitchen have?
possible, but not sure, because I do not use them.

31.10.2015 11:44:00

you fine sieve in the kitchen have?
from wire strainer, and his image - moire

http: // vgrin. host56. com / photo / articles / foveon_inside ... ndex. htm # aliasing

31.10.2015 11:52:00
quote:
:
I thought you zlodarstvovat ))) but no. strange.

And I do not envious and malevolent . A kind and funny . I'm just the fact that the color will be displayed exactly at foveone, but subject to set the exact bb. This is especially true when shooting under artificial flowers. Avtobb from Sigma terrible, I do not deny stating . Therefore, the right to take in such circumstances should be able to. On Baer easier with this, but as I said do not get ideal. Also, as I noted at the Baer laid algorithm for automatic color saturation. Therefore, the shadow of the Baer usually colored. This is not because they saw colored, but because they painted Baer.

31.10.2015 11:57:00
quote:
:

you fine sieve in the kitchen have?
from wire strainer, and his image - moire

http: // vgrin. host56. com / photo / articles / foveon_inside ... ndex. htm # aliasing
Well, it's his image, but his strainer in nature moire not.

quote:
:
quote:
:
I thought you zlodarstvovat ))) but no. strange.
And I do not envious and malevolent . A kind and funny . I'm just the fact that the color will be displayed exactly at foveone, but subject to set the exact bb. This is especially true when shooting under artificial flowers. Avtobb from Sigma terrible, I do not deny stating . Therefore, the right to take in such circumstances should be able to. On Baer easier with this, but as I said do not get ideal. Also, as I noted at the Baer laid algorithm for automatic color saturation. Therefore, the shadow of the Baer usually colored. This is not because they saw colored, but because they painted Baer.
Gyy, Bayer had all colored pixels, so the original color image. If you do not twist the saturation, then it will be like in nature. And about the shadows. . . . the color of the lighting at Bayer does not float as colorful "slides" in front of each pixel of the parameters do not change, as opposed to mathematically estimate about color foveona, who like green = gray and so on. n., depending on the light level.

31.10.2015 12:20:00
quote:
:
from wire strainer, and his image - moire
Well, this is a philosophical question. I think you understand what I mean.

31.10.2015 12:27:00
quote:
:
quote:
:
from wire strainer, and his image - moire
Well, this is a philosophical question. I think you understand what I mean.

not understand. The eye is not a periodic structure, so we moire patterns in nature and do not see. He sees only the Lanselot.

31.10.2015 12:58:00
I looked to the left, where I have a window with a curtain, and saw the moire on the folds. Actually. CHYADNT?

31.10.2015 14:16:00
quote:
:
I looked to the left, where I have a window with a curtain, and saw the moire on the folds. Actually. CHYADNT?
FGM do not treat. Maybe go to the chronic stage. In nature there is no moire, it appears only on the projection.

31.10.2015 14:19:00
quote:
:
at Bayer color all pixels, so the original color image
They are not colored and painted. Moreover, the narrow number of markers.

Offhand. . . 550D (this from a friend and I had yesterday was a party, so come to mind) - what is it?



Prickly autumn


River Tour


Eilat in the sky.

Why, when you see the pictures from the "colored" baera- they are false and have distorted from reality, simplified, but naslaschёnnye color?

But black and white foveon.


SDIM3355. jpg


SDIM8419. jpg


SDIM5423. jpg


SDIM6605. jpg


31.10.2015 14:32:00
quote:
:
quote:
:
at Bayer color all pixels, so the original color image
They are not colored and painted. Moreover, the narrow number of markers.

Offhand. . . 550D (this from a friend and I had yesterday was a party, so come to mind) - what is it?

Why, when you see the pictures from the "colored" baera- they are false and have distorted from reality, simplified, but naslaschёnnye color?

quote:

quote:
:

rest, my dear.

pig is known to find the dirt.

The more that you specifically it somewhere seeks out and drag here, + special camera is purchased, which is only able to remove dirt. And sharply by Sharp is not disconnected.

31.10.2015 18:25:00
quote:
:

The more that you specifically it seeks out somewhere here and drag
Eprst: I look to Yandex photo sample brand camera. The real pictures of real users. And nothing I did not specifically choose wrong! ! ! ! ! ! ! I'm looking for the best! So they are not present. Well get THEMSELVES and show!

I'd have bought Bayern SLR, if you saw is real pictures with her really high quality. Instead of the hat, which is packed with the entire Internet. Only foveona I saw, and see, and have not seen a Baer never anything worthwhile.


31.10.2015 19:38:00

In nature there is no moire

there. Fabric such.

it occurs only on the projection.

It - aliasing. Or overlapping spectra, if in Russian.

31.10.2015 20:19:00
quote:
:
quote:
:
all this nonsense.
Engineering monochrome color reproduction matrix and filter change, or if no mechanics - matrix with its filters and / or a desired color response.
For art - modeling the human eye. Many monochrome sensitive rods and cones of colored (foveon - error).
The Bayer is not happy? It
that suits!
Bayer simple and allows different models. For example, I had Fudge with different pixel size. It was supertsvet!
Scenes from here foveona cause annoyance. I love relsiki and wires, and what to do foveon ((

31.10.2015 20:31:00
quote:
:

Bayer simple and allows different models. For example, I had Fudge with different pixel size. It was supertsvet!
Scenes from here foveona cause annoyance. I relsiki and wires like, and what to do foveon ((
Well..... man.. Pleased that new. you, to say wrong...
Baer always color easier. It it interpolates, shorter than the lying, doing about the same. And foveon thanks to full color pixels make it the way it was in reality.

Here rails foveon, by the way..



SDIM8913. jpg

polnorazmer

https: // img-fotki. yandex. ru / get / 5704/11304747. 43 / 0_8b679_9a56c711_orig
 

31.10.2015 20:37:00
quote:
:
quote:
:
at Bayer color all pixels, so the original color image
They are not colored and painted. Moreover, the narrow number of markers.

But black and white foveon.


SDIM3355. jpg

But Bayer. The sky is blue, grass is green. (ISO 1600, because I was experimenting with a new camera in the evening and the morning shift forgotten long ago it was.)
Pages: 1 2

The idea: a hybrid foveona and buyers

info@www.about-digital-photo.com