Raw conversion ..

Alan Browne wrote in
:

Bayer Aspririn wrote:
On Wed, 30 Sep 2009 16:26:03 -0400, Alan Browne
wrote:

Data is data. Once you read it in you can do as you like to it and
then put it back somewhere (re-moisaic it) with those changes (or
save to a new file [prferably]).

"Re-mosaicing" is impossible. A good example of the problems and
errors of extrapolation taken to the extreme.

I'm not talking about taking the original set mosaic (call it "A")
converting it to an image (call it A') and then back to A again where
it is a perfect match.

re-mosiac? based on what type of maths? One may call it "processed" but re-
mosaic? I fail to see how that should be done on a final result of any set
of Bayer measuring cells and what the gain of this could be except for
inserting lots of mathematical rounding errors if it were possible at all?
... and why on earth anyone woul dwant to call it "re-mosaic"? to me that
makes no sense but plse elaborate?

Bayer Aspririn wrote:

For the sake
of utter simplicity (because demosaicing the values from a Bayer sensor
depends on the values of many more than just four adjoining photosites) if
you have the values of 9, 13, 5, and 8 for the RGGB photosites in one 2x2
photosite section, and then demosaic it into an RGB value of 11, 9, 10
(chosen algorithm dependent),

Bad assumption. You demosaic each RGGB-2x2 into *four*
RGB-triplets (while also looking at the neighbours).

-Wolfgang

John A. wrote:
On Tue, 29 Sep 2009 19:34:53 -0400, Alan Browne
wrote:

Eric Stevens wrote:
On Tue, 29 Sep 2009 07:30:51 -0700, John McWilliams
wrote:

Eric Stevens wrote:
On Mon, 28 Sep 2009 18:00:46 -0700, John McWilliams
wrote:

Matt Clara wrote:
"Wolfgang Weisselberg" wrote in message
...
Savageduck wrote:

In most cases you should end up working in TIF.
Why?
Because it's lossless, capable of handling 16 bit files, preserves
layers and other effects offered by photoshop, et al, allows a couple
forms of lossless compression, none of which a noob needs to know in
order to know which format will best preserve his or her electronic
image file.
Yes, TIFF is tops if you need layers, but keeping the RAW file is the
most complete archival source of a digital image. And it's way smaller
than a TIFF.
But a RAW format does not have and cannot preserve layers.
Which is one reason I wrote what I did. There's not even an inference
that RAW can have layers.
TIFFs have somehow got into the conversation. I was just making an
essential difference clear for those who might think its a toss-up as
to whether to use TIFF of RAW.
The whole notion of what output format to use is meaningless to the OP's
question. (Not that that was terribly meaningful).

You can also conserve things like layers in psd format (from photoshop).

Personally I convert to DNG for master storage. And of course since the
notion of layers is pretty useless for a raw image that doesn't matter
either (and now someone will say that DNG encapsulates a TIF ... and it
still doesn't matter 'cause all I care about is the raw image inside...)

I keep mine as PEF since I read a while back it has better compression
than DNG. I must admit I don't recall having tried it myself to see,
but then again I may have early on and moved on.

I think it makes the most sense to keep archives in a raw
non-demosaiced format rather than TIF. You might find a program later
with a better demosaicing algorithm. Also I'm given to understand that
stock licensors want RAW files.

I hadn't heard that - if they come from me, they'll have to be DNG. Not
that I do that sort of thing.

imbsysop wrote:
Alan Browne wrote in
:

Bayer Aspririn wrote:
On Wed, 30 Sep 2009 16:26:03 -0400, Alan Browne
wrote:

Data is data. Once you read it in you can do as you like to it and
then put it back somewhere (re-moisaic it) with those changes (or
save to a new file [prferably]).
"Re-mosaicing" is impossible. A good example of the problems and
errors of extrapolation taken to the extreme.
I'm not talking about taking the original set mosaic (call it "A")
converting it to an image (call it A') and then back to A again where
it is a perfect match.

re-mosiac? based on what type of maths? One may call it "processed" but re-
mosaic? I fail to see how that should be done on a final result of any set
of Bayer measuring cells and what the gain of this could be except for
inserting lots of mathematical rounding errors if it were possible at all?
.. and why on earth anyone woul dwant to call it "re-mosaic"? to me that
makes no sense but plse elaborate?

Since the resulting, albeit imperfect conversion results in a format the
represents a mosaic, that is as good a term as any. Perhaps call it a
pseudo mosaic.

For your other question see the prior part of the thread.

Alan Browne wrote in
:

snip
re-mosiac? based on what type of maths? One may call it "processed"
but re- mosaic? I fail to see how that should be done on a final
result of any set of Bayer measuring cells and what the gain of this
could be except for inserting lots of mathematical rounding errors if
it were possible at all? .. and why on earth anyone woul dwant to
call it "re-mosaic"? to me that makes no sense but plse elaborate?

Since the resulting, albeit imperfect conversion results in a format
the represents a mosaic, that is as good a term as any. Perhaps call
it a pseudo mosaic.


Well the best comparison I've heard was someone who said that "re-mosaic"
was like starting from vegetable soup trying to reconstitute the individual
vegetables so next time at your grocer you may try to ask for "vegetable
soup" when you want to buy vegetables :-D it may help! .. it may be
semantics but at least a definition should define what it stands for .. see
the eternal PPI/DPI mess ... :-)

imbsysop wrote:
Alan Browne wrote in
:

snip
re-mosiac? based on what type of maths? One may call it "processed"
but re- mosaic? I fail to see how that should be done on a final
result of any set of Bayer measuring cells and what the gain of this
could be except for inserting lots of mathematical rounding errors if
it were possible at all? .. and why on earth anyone woul dwant to
call it "re-mosaic"? to me that makes no sense but plse elaborate?
Since the resulting, albeit imperfect conversion results in a format
the represents a mosaic, that is as good a term as any. Perhaps call
it a pseudo mosaic.


Well the best comparison I've heard was someone who said that "re-mosaic"
was like starting from vegetable soup trying to reconstitute the individual
vegetables so next time at your grocer you may try to ask for "vegetable
soup" when you want to buy vegetables :-D it may help!

Raising a comparison that is ridiculous is poor argument.


A better analogy would be the inversion of non-square matrices
(pseudoinverse). A perfect inversion is impossible, but approximations
more than accurate enough for practical application (such as in aircraft
navigation software to name but one).