On 2005-09-14 05:52, Godfrey DiGiorgi wrote:
> A RAW format image file is generally an enclosure file that contains  
> the following:
> - Camera metadata: all the parameter for JPEG conversion that would  
> have been used in-camera for JPEG format image files plus bits like  
> time/date/manufacturer private data/etc.

such as the camera's serial number

I guess this could include
- sensor faults (defective, calibration)
- image info (lense model, aperture, time, focal length, distance)

I guess it should be a simple task to build a poor lense with strong
vignetting or strong distortions - and compensate most of these flaws by
software later on? This would require some kind of reasonable
characterisation how to compensate those flaws. I guess part of it is the
reasons for proprietary raw formats.

http://openraw.org/faq/?id=14 does not show that many infos about 
What is metadata and what is it good for?

I don't know how much of this error correcting options are used now and will
be used in the future. But I wonder how big this portion of raw info is, the
more detailed it becomes.

> - typically a JPEG thumbnail and JPEG preview rendering at low  
> resolution (maybe more)
> - sensor data either uncompressed or losslessly compressed, a simple  
> 2D matrix
 
> > Is there some kind of 'golden raw' which may be used by the processing
> > software in order to compensate known errors? If it's not  
> > preprocessed into
> > the raw output, is it included within every raw file?
> 
> Some of the parameters are constants fixed by the manufacturer for a  
> particular camera and built into the RAW converter at compile/link  
> time (that's why RAW conversion software usually needs to be revised  
> to accommodate new camera models), the rest are supplied in the  
> camera metadata.
> 
> There are two RAW standards efforts going on to normalize and  
> regularize the representation of RAW camera data: Adobe's Digital  
> Negative specification and the OpenRAW effort. Both of these propose  
> a standardized enclosure format and a way of representing metadata/ 
> processing parameters such that a generic RAW conversion algorithm  
> can be devised independent of camera specific software development.  
> They are both still young efforts.

I wonder whether they are sufficient for all the manufacturer's needs. The
current abuse of EXIF "maker notes" is a sign that either the standard is
not suited very well, or manufacturers don't mind the standards.

- Martin

Reply via email to