New Ideas for in-camera HDR

Rod Supports, the Dictator Interface, Viewfinder Tablet, everything related to hardware goes in here.
39 postsPage 2 of 41, 2, 3, 4

Re: New Ideas for in-camera HDR

Postby Daniel » Wed Sep 07, 2011 7:40 pm

The Foveon Sensor is excellent! i have just contacted them to see what possibilities there are to implement their sensor in a motion picture camera.

There would also be the possibility to make a 3CMOS sensor with full color imaging. that way you would not need to debayer, just combine all 3 of them.

If i understand you correctly there would be the possibility to set the 3 Sensors to 3 different exposures or beamsplitting them in 5/30/65. recombining them not only into a unbayered image including 3 different color qualities and brightness values?
Daniel offline
Posts: 20
Joined: Wed Sep 07, 2011 9:40 am

Re: New Ideas for in-camera HDR

Postby Biel Bestué » Thu Sep 08, 2011 6:18 pm

Daniel wrote:If i understand you correctly there would be the possibility to set the 3 Sensors to 3 different exposures or beamsplitting them in 5/30/65. recombining them not only into a unbayered image including 3 different color qualities and brightness values?


no don't get me wrong, it's an idea of mine not somehting Foveon says.
I guess you've seen how the x3 is layered isn't it? my idea would be to use the x3 simply for luma, so no Tining of the individual layers (maye they are not tinted, maybe they use a dichroic filter inbetween layers, that is a thing I would be glad to know)
so use the x3 for black and white, and use the bayer filter before the chip, so you get the colors, but then why use the x3? because since there are 3 layers and all of them record the same at the same time, why not make them record diferent amount of luma?

in the end this is how it would look like:

>> light >>>> bayer filter > first sensor > ND filter > second sensor > ND filter > third sensor >>>> smile >

:mrgreen:

the senors are BW but since they are filtered before, after debayering you would get a image with wider-than-widerest latitude EvAr! :lol: and so with a huge headache about the three streams coming from the sensor (6 if shooting 3d)


maybe this can be achieved without the ND filters by makiing the first layer the thickest (so it recieves the most light, so dark zone) the middle one thinner (medium lights), and the last one the thinnest (so recieves less light, making it the highlights sensor)



about the 3CMOS there is the problem with the lens, the prism touches the lens, wich is a big "no no" for expensive lens. and requieres especific lens to solve this wich also means expensive lenses. :|
>> Biel Bestué de Luna <<
Biel Bestué offline
User avatar
Posts: 1011
Joined: Sun Nov 07, 2010 3:53 pm
Location: Torroella de Montgrí - Girona - Catalunya

Re: New Ideas for in-camera HDR

Postby shijan » Thu Sep 08, 2011 7:09 pm

wow! it become more and more interesting!
shijan offline
Posts: 85
Joined: Thu Nov 04, 2010 4:05 am

Re: New Ideas for in-camera HDR

Postby Daniel » Fri Sep 09, 2011 6:57 am

@Biel Bestué

ok, now i see what you mean. but that way we would only have 3 BW-Image-streams, no? You would still have to use a beamsplitted 2nd Sensor to get the color information to combine it with? am i correct? You would need 3 different COLOR images to get the ultrawide DR?

i am experimenting right now to make a proof of concept with 2 E-Mount cameras using a beamsplitter BEFORE a Nikon Lens. Maybe i need a Lens to change the Focal distance so that i don't have problems with the distance.
Daniel offline
Posts: 20
Joined: Wed Sep 07, 2011 9:40 am

Re: New Ideas for in-camera HDR

Postby Daniel » Fri Sep 09, 2011 1:13 pm

oh by the way this is beeing produced already:
http://www.hdrlabs.com/news/index.php?i ... 7378279667

Image

Daniel offline
Posts: 20
Joined: Wed Sep 07, 2011 9:40 am

Re: New Ideas for in-camera HDR

Postby Daniel » Fri Sep 09, 2011 1:29 pm

oh man, i keep discovering wicked stuff on the web:
http://www.hdrv.org/IBL.php
Daniel offline
Posts: 20
Joined: Wed Sep 07, 2011 9:40 am

Re: New Ideas for in-camera HDR

Postby Biel Bestué » Fri Sep 09, 2011 2:26 pm

in a image there are two elements the first is implicit in the second on the commond RGB color space, but this is not the only way to represent an image
the idea is this: luma and chroma, luma is the amount of light while Chroma is the kind of color (not the amount of color)

there are no "color sensors" all sensors (imager from now on) are luma only (BW)

the chroma in the image is done via the debayering of a BW image,
this image is not a normal BW image as it's a filtered BW imager,
all the parts of the image are not "luma correct" of the real subject because the imager was filtered
this filter the imager has obove it has the bayer patern.

therefore any imager captures an image that only after being interpreted gives you the final colour image (that's event rue for 3ccd or 3cmos) this process, or interepretation, is called debayering, for 3ccd and 3cmos there is no debayering as images are not cptured the same way, there is no bayer patern on a3ccd or 3cmos imager

now that might not be that true for the faveon...

i don't know the exact way the Faveon works, as the whole info about faveo is full of PR, and the images are too simple to understand how the imager works.

are there 3 translucent imagers (if true, how do they work?) with dichroic filter inbetween (dichroic filters should change the color of the light that passes trhough with minimal luma loss)

or simply the three surfaces are just a way refract light to three imagers on the side?

i don't know how it works...

but if the imagers where translucent then why not skip the color process, (skip the dichroic filters inbetween) and put infront of the whole set of imagers the bayer patern, and then put ND filters inbetween the set of imagers to get diferent captures of the same thing and at the same time...

but then again all this is SCI FI without the capacity to rebutte all this.
>> Biel Bestué de Luna <<
Biel Bestué offline
User avatar
Posts: 1011
Joined: Sun Nov 07, 2010 3:53 pm
Location: Torroella de Montgrí - Girona - Catalunya

Re: New Ideas for in-camera HDR

Postby Biel Bestué » Fri Sep 09, 2011 4:58 pm

oh god I hate all those fake "hdr" tonemapped videos! they are not appealing to the eye, they are a gross malinterepretation of exposure! :lol: this images cannot combine well, look at the dark images in the AMP cam video, why is so badly exposed?! there no blacks in a underexposed image?! and their propaganda says "some tonemappers apear really natural while some others can creat unique visual effects"... yeah defects I would say! :mrgreen:

there is something that bothers me also about this AMP camera, they use the same kind of imager for the three images, wich means that if they share the same shutter and FPS means the whole under exposure and over exposure is handled with gain! :shock: no wonder they have problems with artifacts in color. also the type of artifacts they show seem the kind of artifacts of the debayering process.

also there seems to be this classic problem that also happens when manual bracketing is used where the most overexposed images display the natural bloming that happens in imagers on overxposed zones, i guess light bleed from photosite to photosite becoming a "glow" or a "bloom" if you add this blooming with the digital blooming of some tonemapping algorithms you end up with a very "soft" image, somewhat like a dreamy image. yeah quite a helpful "artistic" restraint isn't it?

people is saying this is the digital equivalent to what Ansel Adams did with his zone system, what an insult to Ansel Adams and the zone system, how can a fixed three exposure system can compare to the fine tunning of the exposure and printing process of Ansel Adams? please people rely more on the capabilty of your DPs and less with the capabilty of the technology your own camera delivers, as most of the images displayed in this PR video for AMPcam can be shot with correct light decisions with a normal camera, delivering similar and much more natural results than this HDR crap videos. :twisted:

yes HDR is the future of video, but not done like this nor displayed like this, the future will come with an imager that can capture a wider latitude and quantitzate the info at higher BITs, all other processes will have their ups and downs, at what BITs those HDR cams capture?
>> Biel Bestué de Luna <<
Biel Bestué offline
User avatar
Posts: 1011
Joined: Sun Nov 07, 2010 3:53 pm
Location: Torroella de Montgrí - Girona - Catalunya

Re: New Ideas for in-camera HDR

Postby Daniel » Fri Sep 09, 2011 10:56 pm

Thank you for your thoughout responses.
I am not at all fan of those tonemapped images either. they look weird on pictures but on moving pictures they look downright laughable. What i like about the possibility to capture video in HDR is that in a backlight situation i can mimic what the eye can do without having a complete white background or a completely dark foreground. I like the examples of use by RED HDRx footage: http://www.fxguide.com/featured/hdrx-the-impossible-shot/

it looks very natural to the eye, even though there is probably 15 or more stops in that picture. You don't need to expose darks and highlights as midtones, if your image has a "true" white and black, and a normal S-Curveit looks very natural to the eye.
For example: the waterballoon-shot at 4:25 in the demo-video looks completely natural to my eye. if there weren't the original images on the right, you wouldn't even recognise it as a HDR-composed shot.

with HDR i can allways dial the contrast back up and "loose" highlights or shadows if i want, but the opposite isn't true, so that means more grading possibilities in "not-so-ideal-scenes".

That doesn't replace a good DP or a Lighting-Crew. It makes their Job easier and faster. Imagine all the moment's where you wished your foreground would still be gradable because you had to expose for the highlights. And sometimes artificial lighting simply doesn't cut it, or even changes the feeling of the scene that would have looked better with just natural light.
I'm all for max possibilities to play with. Especially if they get us closer to our incredible experience of sight (what a great gift!). What Ansel Adams does in his work is very much what our eye do naturally.

ANYWAY. Back to HDR technicality.
I understand your thoughts about the foveon imager. Great idea. have each Photosite in the Bayer get 3 Exposures. From what i understand of their interactive demo you have much higher sharpness and detail with the foveon imager than with a bayer pattern. But that way you could do it with 1 imager alone.

interesting...interesting.
Daniel offline
Posts: 20
Joined: Wed Sep 07, 2011 9:40 am

Re: New Ideas for in-camera HDR

Postby Guest » Fri Sep 09, 2011 11:22 pm

Hi,

Forgive me for only having skimmed this thread but do the foveon imagers somehow do away the requirements to compensate for motion blur differences between exposures?
Guest offline

PreviousNext
39 postsPage 2 of 41, 2, 3, 4

Who is online

Users browsing this forum: No registered users and 1 guest

cron