Canon EOS 30D support

[DJJohnson]

Thanks for posting those links. What you don't realize is that you posted more proof of what I had just posted here about you. I was thinking of refering them to your discussions there too to show them you don't understand RAW data, but I didn't think they'd care to have to listen to more of your nonsense.

What part of hoogo's "No" reply to your question didn't you comprehend?

You don't comprehend what I told you, you can't comprehend what hoogo has told you, you don't even believe Gerhard when he tells you that you are in error. How many more people are going to have to tell you that you just don't understand how to properly use RAW files from a digital-camera?

Please go educate yourself.

[gdoten]

What part of hoogo's "No" reply to your question didn't you comprehend?.

Hoogo gave his opinion about this question: "Wouldn't you agree that if you have to make adjustments to the image that the program is not properly interpretting the camera's raw data?" No problem there. I happen to disagree with him, but that's fine.

The real question of interest here to shut you up once and for all in the PSP group is this one: "Are you saying that PL has code in it for reading 30D-produced .cr2 files?" It's a simple question; no metaphysics or existentialism involved to answer.

Why are you purposely making this so difficult?

-glenn-

[Hoogo]

Does PL know how to read the Canon 30D-specific data that is present in a 30D .cr2 file or not? Why is this question so difficult to answer?

Why are the given answers not satisfying? I don't get that very special feature about this question that has not been answered yet. Or are you talking about Exifs?

[Hoogo]

...Instead of using the RGBG array to define one colored pixel...Is, or would, something like this be possible with PL32?...

The interpolation does not just fit 4 pixels into one. A 4mpix-camera has real 4mpix, 1mpix red, 1mpix blue, 2mpix green. There are quite some algorithms taht try to keep as much resolution while calculating the colors. You may try http://www-ise.stanford.edu/~tingchen/main.htm, but it seems to be broken at the moment. All in all it is said that the resolution in brightness drops by factor 1.5, not 4.
In a beta-version there has been a "no interpolation"-interpolation that just showed the bayer-pattern and that invited for some crazy experiments, but if you try to use it for b/w you will find that it turns a blue sky into a hires-chessboard. You can still try it with dcraw or ufraw, both had this option a while ago, they surely will still have.

[Hoogo]

My guess is that a Corel programmer dropped him on his head as a child and he's never gotten over it.

I like this one

[DJJohnson]

My guess is that a Corel programmer dropped him on his head as a child and he's never gotten over it.

I like this one

Judging by how clumsy and careless that the Corel programmers are with they way they destroyed PSP, I can see how this thought would come into his mind.



p.s. Thanks for the further info on using the RGBG data in that novel way. After reading that discussion long ago I never checked to see if anyone had tried it. I find it odd though that data from the individual sensors couldn't be used to recreate the scene properly. Maybe with further analyis of what colors are supposed to be where, that those checkerboard patterns could be aleviated by raising the signal levels from the sensors that were too dark, from being resolved through a contrasting color-mask.

[gdoten]

Does PL know how to read the Canon 30D-specific data that is present in a 30D .cr2 file or not? Why is this question so difficult to answer?

Why are the given answers not satisfying? I don't get that very special feature about this question that has not been answered yet. Or are you talking about Exifs?

The Canon 30D, when generating a raw file, writes some 30D-specific metadata about the image. Does PL utilize that 30D-specific, only found in a 30D raw file, metadata when displaying a 30D image?

The importance of this question is that it indicates whether the PL developers were sucessful in fully reverse engineering the 30D metadata that is only found in a 30D-generated raw file. If PL does not utilize this metadata it would explain why it does not display may 30D camera images correctly.

If it does read and apply this 30D-specific metadata then I have no idea what the problem is. I'm certainly not an expert when it comes to this stuff. It's just a hobby for me.

-glenn-

[gdoten]

Does PL know how to read the Canon 30D-specific data that is present in a 30D .cr2 file or not? Why is this question so difficult to answer?

Why are the given answers not satisfying? I don't get that very special feature about this question that has not been answered yet. Or are you talking about Exifs?

The Canon 30D, when generating a raw file, writes some 30D-specific metadata about the image. Does PL utilize that 30D-specific, only found in a 30D raw file, metadata when displaying a 30D image?

The importance of this question is that it indicates whether the PL developers were sucessful in fully reverse engineering the 30D metadata that is only found in a 30D-generated raw file. If PL does not utilize this metadata it would explain why it does not display may 30D camera images correctly.

If it does read and apply this 30D-specific metadata then I have no idea what the problem is. I'm certainly not an expert when it comes to this stuff. It's just a hobby for me.

-glenn-

Wait, wait, wait! Hold the phone! While driving into work this morning I think I've figured out what the disconnect is in understanding my question. You're question about Exif data made me think of this.

When I say the 30D-specific metadata I don't mean the metadata like Exif data or whether the camera was set in the "boy, it's a sunny day" mode, etc. (Although any decent image editing program would be able to read and properly act on that information.)

The 30D-specific .cr2 raw metadata I am referring to is the proprietary data structures that Canon uses to write out the camera's sensor information with. There is metadata, in an unpublished format, that is written out along with the raw sensor values. This is the information that image editing software developers have to reverse engineer in order to be able to display the image correctly.

Does that make the question more understandable and reasonable? Without being able to decode and then know how to utilize this sensor metadata then a program cannot properly display the image as shot by the photographer and faithfully captured by the camera. I've ben assuming all along that this is why the Canon EOS 30D is not on the PL website page showing the list of supported cameras--because it doesn't know how to read the sensor metadata.

Hope this helps.

-glenn-

[Gerhard Huber]

When I say the 30D-specific metadata I don't mean the metadata like Exif data or whether the camera was set in the "boy, it's a sunny day" mode, etc. (Although any decent image editing program would be able to read and properly act on that information.)

You are not correct informed about digital camera raw images.
In such files, for example cr2 files, there is the information stored that comes from the cameras sensor. Beside this there is often one or more JPEG previews. Some cameras store IPTC data also - here you can find things like the author of the image. Then there are camera specific data. Almost cameras store them in the EXIF data block. Beside the usual data there is a manufacturer block inside the EXIF data. That's all.
The only data of interest for reading cr2 files is the EXIF block. In the manufactorer notes there are things like your camera settings, white point values, and so on. So far so good. But to decode the raw image data, there is no special way stored in the data. Every (really every) raw decoding software (including the camera itselve when it saves in JPEG) uses special algorythms to get a "good" result. Since automatic algorythms seldon produce perfect results, every raw decoding software has additional functions to controll things like gamma, brightness, contrast, whitepoint, ....

Gerhard

[Hoogo]

I found the new adress for the bayer-interpolation-page: http://scien.stanford.edu/class/psych22 ... n/main.htm
And by the way: Raw-files of the 30D open just fine. I already forgot what the duscussion was about.