The time has come

I this section you can introduce yourself, post invitations and more...

Moderator: Stany Buyle

Post Reply
Iliah Borg
AFFILIATED MEMBER
Posts: 20
Joined: Tue 16 Feb 2010 17:53

The time has come

Post by Iliah Borg »

Stany asked me to join here to answer your questions about UniWB and LibRAW, and maybe some other technical questions, like of cabbages - and kings. Your questions, please.
User avatar
Stany Buyle
moderator
Posts: 2893
Joined: Sun 30 Nov 2008 15:05
Active Member?: Yes
Post pictures?: Yes
Contact:

Welcome and some questions...

Post by Stany Buyle »

Good morning Iliah,

First of all welcome to our forum.

Some questions:
  • As far as I think to understand you're involved in the development of a raw "reader" called libraw. Is libraw also a raw converter or how do I have to understand this? Can an ordinary user like myself do something with libraw?
  • What is the difference between a front illuminated structure and a back illuminated sensor structure in relation to the photographic results?
  • Is there a psysical limit for sensivity and overall IQ of sensors(CCD and CMOS) as we know them now or did Nikon D3s reach the maximum of efficiency about light gathering and processing?
  • Bayer filter and RGB color model are what ordinary users like myself are common with in relation to the initial step after photons hitting the photosites in order to "see" coloured images , is there any upcoming changement about this? Does the Bayer filter forfill all requirements for near future digital imaging evolution? I was also following a little the Foveon story from Sigma, is that system better?
  • We had a very interesting thread on this forum started by Lil Judd about UniWB... What is UniWB? How can you implement it in my PP or my camera? What are the advantages?
  • About your raw converter RawMagic, what are the major differences with other converters? I just downloaded the trial today but the compatibility isn't clear to me... It says "PC only" ... 64bit as well? I get very strange view(at least for me) when opening a raw image.
That was it for today...lol

Thanks in advance for your reply.
kindest regards,

Stany
I like better one good shot in a day than 10 bad ones in a second...
http://www.fotografie.cafe
Iliah Borg
AFFILIATED MEMBER
Posts: 20
Joined: Tue 16 Feb 2010 17:53

LibRAW

Post by Iliah Borg »

LibRAW is a tool for raw converter developers and those who want to have access to raw data to perform various measurements over it. LibRAW was created to simplify the use of dcraw code and to improve over it. Two examples of the use of LibRAW are a free multi-platform raw converter digiKam http://www.digikam.org/drupal/about and a tool to analyze digital image quality, Imatest http://www.imatest.com/home. LibRAW also includes a pre-compiled 4channels utility that allows any user to extract separate (R, G1, G2, B) channels from raw files in order to, say, bring them to Photoshop and figure out the calibration of in-camera exposure meter or see the effect on-lens colour correction filters have over the per-channel exposure. Pre-compiled LibRAW examples lack GUI, they are command-line utilities. We made a OS X GUI wrapper around 4channels, http://pochtar.com/LibRAW_4channels.zip
User avatar
Stany Buyle
moderator
Posts: 2893
Joined: Sun 30 Nov 2008 15:05
Active Member?: Yes
Post pictures?: Yes
Contact:

Re: LibRAW

Post by Stany Buyle »

Iliah Borg wrote:LibRAW is ...
Goodmorning Iliah. Thanks for explaining LibRaw even though I will have to re-read it multiple times more to fully(correction, partly) understand.
For myself, two very interesting things in your explanation are the link to Imatest where I think to see multiple tools uded by Chasseur d'Images test lab(my favorite reviews), Dpreview and other testers. Unfortunately all quite expensive...

The second thing that was catching my attention are 2 greens in Raw files(R, G1, G2, B), I thought it was just red, green and blue... Can you clear that a little more up, please?
In one of your latest reactions on Dpreview I was also wondering about your explanation about green being more sensitive.
I[color=#000000]liah Borg [url=http://forums.dpreview.com/forums/read.asp?forum=1041&message=34548962]on Dpreview[/url] wrote:The reason is not that greens are 2x the number of reds and blues; but that each green is roughly 1.5x more sensitive to daylight than reds and blues.[/color]

Is there any relation between the two greens you mention in the extraction from the raw file or do I understand completely wrong?

Thanks in advance.
kindest regards,

Stany
I like better one good shot in a day than 10 bad ones in a second...
http://www.fotografie.cafe
James
moderator
Posts: 255
Joined: Fri 28 Nov 2008 22:43

Re: The time has come

Post by James »

Hey Stany,

I think I can answer a lot of this but am happy to be corrected if I make any mistakes!

The difference between front and back illuminated structures is where the electrical parts of the chip sit, on a front illuminated sensor all of the electronics are on the same side of the chip as the light sensitive part, on a back illuminated structure it's the other way round, this means that you can have a larger surface area (so each photosite can be larger on the same size chip) so it should be more sensitive to light, which means it can give you a cleaner result at higher sensitivities.

As for limits to sensitivities I would be skeptical of anyone who said we had reached them - when I was at university in the late 90s the original pentium processers were seen to be approaching the limit of what could be achieved in terms of clock speed and complexity due to various techincal reasons (mainly heat dissipation and etching the silicon), 12 years later the same level of gains are still being made as always, it's lack of imagination that is nearly always the limit. (With the processors they realised that they could switch to xray instead of visible or UV light in one of the steps :) - whatever it was it was changing the part of the spectrum that was used that solved the issue then). The next limit was the interference that would be caused by having electrons move so closely together - that's been surmounted too.

I think the foveon system is massively better per photosite! The DP1 is effectively using a 7 or 8 year old sensor (it could be older, that's when I became aware of it) and for certain applications it was still outresolving my D300 despite only having 1/3 the number of pixels - that's pretty incredible! I think that bayer going will be a major breakthrough when it happens - I am not sure it needs to yet though. Also the chip has limitations too (you have to pass the light through each layer so you are still losing some light for example) but as a per photosite example it seems a lot better to me (a few users have the DP1 here so ask for their impressions too!)

And for your question about greens - most (if not all) bayer matrixes have 2x the number of green sites to blue or red but they can be laid out differently, an altearnative is to have a second type of green in the matrix. This would mean that you should get more detail on green objects than other colours and is supposed to be pretty similar to the way that the eye sees (which I guess is why it works so well). Reading Iliah's comment you posted suggests to me that he is saying daylight has more information in the green part of the spectrum (or the green filters are wider) and so per site you are getting more information (in daylight) in the green sites than either the red or the blue.

Hope that helps :)
User avatar
Stany Buyle
moderator
Posts: 2893
Joined: Sun 30 Nov 2008 15:05
Active Member?: Yes
Post pictures?: Yes
Contact:

Re: The time has come

Post by Stany Buyle »

James wrote:Hey Stany,
I think I can answer a lot of this but am happy to be corrected if I make any mistakes!
The difference between ...
Hope that helps
:)
Thanks for posting this reaction James, very intersting!
kindest regards,

Stany
I like better one good shot in a day than 10 bad ones in a second...
http://www.fotografie.cafe
pam.meier
moderator
Posts: 421
Joined: Thu 28 Aug 2008 18:54

Re: The time has come

Post by pam.meier »

Welcome on our forum, Iliah. I´m bloomoose on dpreview and you have greatly helped me with my A900.
I can´t wait to learn about your findings on cabbages ;)
Iliah Borg wrote:Stany asked me to join here to answer your questions about UniWB and LibRAW, and maybe some other technical questions, like of cabbages - and kings. Your questions, please.
BrianSmith
AFFILIATED MEMBER
Posts: 76
Joined: Thu 4 Dec 2008 21:37
Active Member?: Yes
Post pictures?: Yes
Location: Isle of Skye, Scotland
Contact:

Re: The time has come

Post by BrianSmith »

James wrote:Hey Stany,

I think I can answer a lot of this but am happy to be corrected if I make any mistakes!

The difference between front and back illuminated structures is where the electrical parts of the chip sit, on a front illuminated sensor all of the electronics are on the same side of the chip as the light sensitive part, on a back illuminated structure it's the other way round, this means that you can have a larger surface area (so each photosite can be larger on the same size chip) so it should be more sensitive to light, which means it can give you a cleaner result at higher sensitivities.
First, welcome Iliah. We have occasionally exchanged e-mails over Raw Magick, I am xrdbear on dpr.

James, do you know how the back illuminated structure is created? Is the surface layer ( with photosites ) created on top of the electrical connections and processing electronics or is the backside shaved off by plasma etching or somesuch to reveal the photosites? Either way it must result in an incredibly thin and fragile chip to work with. I used to work with Power MOSFET wafers and these were fragile enough.
James
moderator
Posts: 255
Joined: Fri 28 Nov 2008 22:43

Re: The time has come

Post by James »

I would assume that it's done by etching but that really is a wild assumption based on nothing other than how I would try to do it :)

I guess it's something that's switched recently as Samsung and Toshiba both announced similar chips in the last few months - your guess is probably a lot better than mine Brian :)
Post Reply