jerryc41 Loc: Catskill Mts of NY

It's semantics. I'm always careful to use the term "raw file" rather than raw image because I know people immediately jump on my "wrong" terminology. You can see what your camera took when set to raw mode, right? Well, it must be a picture, right? Raw, JPEG, TIFF, ARW - it's all just 1s and 0s.

jerryc41 Loc: Catskill Mts of NY

mborn Loc: Massachusetts

photoman022 Loc: Manchester CT USA

jeep_daddy Loc: Prescott AZ

Ones and zeros my friend - Ones and Zeros.

JohnFrim Loc: Somewhere in the Great White North.

There is no reality; there is only perception.

Why is a printed photo considered an image? In reality, it is a piece of paper with an arrangement of coloured dots. Is the image a "photograph" or just a "snapshot"? The answer to that lies in how the image is perceived.

Consider this: we have agreed that a particular arrangement of black dots on a piece of white paper represents a letter of the alphabet. We have further agreed that a collection of those dots represents a word, and that many words on many pages are a book. But the "story" that we enjoy is our perception of the words and their meanings; the reality is that we have paper and dots of black ink.

Think back to the early days of creating "images" with line printers -- they were creative assemblies of alphabetic characters with different densities arranged to create shading and contours, and when viewed from a distance they created the perception of an image. So in reality, we had a piece of paper with black ink dots, but we interpreted the arrangement as an image.

ALL computer files are collections of 1s and 0s that are read, interpreted by software, and presented on a screen. A "text file" is a collection of data bits that, when interpreted by a "text application", gets displayed/printed as an arrangement of dots on a white background that we as humans recognize as words. Similarly, an "image file" gets interpreted by an "image application" and the resultant presentation on the screen looks like a picture to the human brain. In fact, if you "read" the image file with a "text application" you get "a screenful of text" that is generally meaningless because the interpreted stream of 1s and 0s do not result in recognizable patterns of dots on the screen.

So regarding raw vs JPEG, technically they are both computer files and NOT images. But in the case of JPEG, the digital data is stored in an agreed-upon standard pattern that "image applications" read and display in a way that we as humans will interpret as an image. The issue with raw files is that the data is not stored in a standardized format. If the computer/photo industry could agree on a ".RAW" format we would see that extension on filenames and the argument of raw not being an image file would go away.

Of course, the "true raw data" coming from the sensor may still have to be "processed and packaged" into that standard .RAW format, which would probably just add to the workload of the camera. It is much easier and faster to simply store the digital data with the least amount of processing in the camera and leave the heavy lifting of image creation/adjusting/manipulating to the desktop computers. Software for processing raw files still has to know the proprietary structure of the raw data file in order to recreate the image.

Reality: a bunch of dots on a screen based on 1s and 0s in a computer file
Perception: a letter; a word; a story; an image; a snapshot; an amazing photograph

f8lee Loc: New Mexico

Sorry - you are incorrect - read what I already posted earlier.

JohnFrim Loc: Somewhere in the Great White North.

And what part of what I wrote conflicts with what you wrote, which BTW I thought was very good? I simply tried to add another level to the discussion.

f8lee Loc: New Mexico

You wrote in part:
"ALL computer files are collections of 1s and 0s that are read, interpreted by software, and presented on a screen. A "text file" is a collection of data bits that, when interpreted by a "text application", gets displayed/printed as an arrangement of dots on a white background that we as humans recognize as words. Similarly, an "image file" gets interpreted by an "image application" and the resultant presentation on the screen looks like a picture to the human brain. In fact, if you "read" the image file with a "text application" you get "a screenful of text" that is generally meaningless because the interpreted stream of 1s and 0s do not result in recognizable patterns of dots on the screen." is an example of your lack of understanding."

A text file and a JPEG file are similar in that both entail the use of a lookup table to produce what you see on screen. The ASCII coding that almost all text based computer files is based on nowadays requires that when the CPU "sees" the string 1000001 it will return an upper case A. Period. That's it. There is no interpretation other than the lookup table, and every computer everywhere that sees that 1000001 will generate the same upper case A.

Likewise, a JPEG or TIF or other image file denotes a specific RGB value for each dot in the image - again, with a corresponding lookup table that create the specific color associated with that set of values. There may be differences among computers in displaying that specific color, but that has only to do with the imperfections in the hardware - different monitors or ambient light etc. can impact the color you perceive.

A raw file (again, no caps required) is the stream of data that comes off the imaging chip. There is no defined lookup table for this data that specifies that color a particular image dot will be when the R, G and B photo sites in the area have whatever readings they take. So different demosaicing programs can actually legitimately create different variations of a given raw file - the changes may be subtle, but many pros swear to how C1 does a better job than Adobe for instance. Contrast this with the fact that if you take that text file and load it onto ANY computer with a program that reads text files, and 1000001 will always generate the exact same output - an upper case "A".

John_F Loc: Minneapolis, MN

I would defer on the "A JPEG is..." remark. A joeg is a discrete cosine transform of the original image data, presumably RAW. To display the image the transform must be reverse transformed (math addicts please tell the right name).

Longshadow Loc: Audubon, PA, United States

JPEG file......

JohnFrim Loc: Somewhere in the Great White North.

I really don't see a conflict in what we are both saying. I used the word "interpret" and you said "lookup tables". Same thing, really. My point was that a string of binary bits may be a particular letter to a text application, or a set of coloured dots to an image application. If read by the "wrong" application you will not get the intended result.

Bill_de Loc: US

Yup Gene, it looks like a lot of us do.

--

f8lee Loc: New Mexico

Alas, that's the rub...they are not the same thing (except in the sloppiest manner of English language).

To "interpret" implies a decision is made (in this case, as to the specific color to be rendered at a given picture element) - a lookup table requires no interpretation, just a mechanical knee-jerk reflex that when 10000001 appears show the letter "A" or when the RGB values 148,0,211 appear display dark violet. One equals the other - dark violet is 148,0,211. This is not the case in raw demoasaicing, where two programs (perhaps Lightroom and Canon's own program for their raw files) can derive different RGB values when presented with the same set of values for the array of R G and B photosites surrounding that pixel. In other words, one program on the computer may give a different result than another program on that same computer using the same original raw file. This is not the case for RGB values in a TIF or JPEG etc. file - all programs will "find" the same color for a given set of values (because there is no interpretation, it is just looked up on that table). Again, different monitors, graphic cards and other variables might make your screen look different than mine, but in all cases the RGB values are unchanged.

This is also the reason that you might find quite a difference between raw files processed years ago and then again with recent versions of the same software (I surely did when I was playing around and reprocessed NEF files from 8 years ago with a current version of DxO, which came out much better than DxO of 8 years ago did). The raw file is not an image until it is demosaiced. While there is an embedded JPEG (where the in-camera computer has done the processing I described above) the raw file itself is just a string of 0's and 1's that need to be coalesced before an image is made.

JohnFrim Loc: Somewhere in the Great White North.

C'mon, many of us are retired and the clock is ticking. We have to make the most of what time we have left... and that's why there is UHH.