jenny wrote:
Ooooo cat jumped up on me and lost the thread! But let's go back to your reply claiming data is lost. When? Now, really, nothing is lost if shooting JPEG. You would lose nothing but what you gained in switching to RAW if you went back to the smaller file...which you probably will if you print. Where do you get more---of what? --for a RAW file?
You are down to describing your own camera as proof of...what? I get about 3X size file in RAW as a straight JPEG. Says nothing about either of us though....
You invited me to reply to your input in this thread, or actually was it to challenge me (?) and claimed to be neutral about the subject. So where did you get more of what to make a RAW file? Do they grow like bacteria or bean sprouts? Sorry, but I just don't feel this is going to turn into a productive discussion. Most people are on one side or the other of this fence and your lack of neutrality is showing. Back to square one, to each his/her own. :roll:
Ooooo cat jumped up on me and lost the thread! But... (
show quote)
Jenny, Ok, as you stated let's go back to square one. In addition, I'll provide this tidbit of background info on me. For the last 15 years, years or so, I've been involved with computer programming. Though my experience goes back some years prior to that, I spent the majority of the time prior designing aerospace equipment. I bring up the programming part as a basis for credibility on computer files. No, I'm not an expert, but I am comfortable discussing and explaining file structures.
My cat just jumped on my lap, which will provide a bit of difficulty in typing quickly. I have to do that over her back.
Going back to RAW files. Each of those captured by my camera are approximately 14 megabytes. In fact, I just took two photos of the same subject, tripod as to not change any bit of the subject.
The RAW file is 13,256 kilobytes or 13.2 megabytes. The jpg file is 3,484 kilobytes or 3.4 megabytes. That is a significant amount of difference in the amount of information stored in these files. The jpg capture was set for large file size, quality optimized. Essentially, this is the set to provide the largest image with the highest quality. The camera, by the way, is a Nikon D700.
I would suggest reading the articles in wikipedia.org
http://en.wikipedia.org/wiki/Jpghttp://en.wikipedia.org/wiki/Raw_image_formatHere is part of the reason the jpg file is smaller. The algorithm which is used to take the RAW data and render a jpg image from it, compresses the data.
The JPEG compression standard, which was originally specified in 1994, specifies a coding process which has several specified steps:
1. Image Preparation (Separating the Components)
The image is separated into components, which are adapted to fit the level of quality chosen.
A JPEG file consists of distinct components, which are processed independently from each other. Both teh contents and resolution of the components could differ depending on application and level of quality chosen.
Now we have some variables to specify, or declare. The most common format is YUV using the two colour components U and V and the brightness component Y. Depending on the desired image quality the resolution of the color components could be reduced, for example by the factor two or four; i.e. beside the full brightness information two or four pixels will be combined together to one value of color information.
This approach results from the selective sensitivity of the human eye for the perception of differences in color and brightness. This results from the anatomical structure of the human eye, specifically the fovea. The amount of rods and cones present. If I'm not mistaken, the rods are each sensitive to one color. The brain combines the amount of information transferred and combines it to render what we see. The digital camera sensor, along with the Bayer filter pretty much do the same thing.
The brightness is substantially of higher importance for the total perception of the image. If the brightness component of an image is removed the remaining colors form a meaningless abstract. Therefore it is recommended to provide a proportionally larger part of the data volume for the encoding of the brightness. These procedures are called 4:2:2 or 4:1:1 subsampling.
2. Transformation (DCT: Discrete Cosine Transform)
Each individual component is transferred into a format, that allows to make conclusions about the structure of the contents. This offers the option to distinguish between basic and more complex contents. It also involves a good deal of math, that I'd rather not go into here.
The requirements to an ideal format for image compression can be outlined as follows:
The image information should be differentiated according to their meaning for the qualitative impression.
The internal structure should offer the opportunity to decide, to what extent the omission of information deteriorates the image quality. This is crucial for a presentable jpg image.
Only relevant information should require data capacity. If a section does not contain information like monochromatic areas, this should be directly reflected within encoded data.
The pixelwise representation used for conventional data formats do not fulfil these requirements. For that reason the original data has to be converted by the help of the so-called Discrete Cosine Transform (DCT). After conversion the data show the following characteristics:
Simple structures are given low values; complex, more detailed structures, high values.
Values resulting from the DCT reflect the geometrical structure of the contents. Details, as horizontal or vertical patterns, not contained in the image will be given the value zero.
Monochromatic areas are given a single value. All other values become zero.
Each component is normally given 8 x 8 pixels. Each component will then be transformed.
3. Quantization (Weighting of the Contents)
The transformed data are weighted according to their meaning for the image contents.
The result of the DCT is a set of values also describing the image contents. In contrast to the original format they do not affect the image quality in the same manner. This is important. The reason is the subjectively different perception of details.
With the help of quantization more relevant values are described with a higher accuracy. Values influencing the subjective perception to a smaller extend will be devalued and represented with lower range of values.
Quantization is not reversible and substantially affects the image quality.
The quantization is processed with special tables which will be stored within the corresponding JPEG file. They are essential for decoding.
4. Entropy Coding (Huffman Coding)
Elimination of redundant information.
Data, which result from the preceding steps, are provided with a different quantity of redundancy, which does not supply any contribution for the quality of the image representation. This redundancy is reduced e.g. with the help of a Huffman coding.
The basis for this coding is a code tree according to Huffman, which assigns short code words to symbols frequently used and long code words to symbols rarely used.
Beside Huffman coding another procedure is specified, which does not have a practical meaning because of the patent situation (arithmetic coding).
Additionally a special form of Run Length Encoding is applied.
This describes the manipulation of the data into a file. Pretty much reverse the steps and that produces the data necessary to produce the image we either see on a screen or print on a printer. All of these mathematical machinations and compression transforms result in a smaller file. The purpose of this was to allow the transfer of data in the shortest amount of time possible. So, this compression results in a, still identifiable and recognizable, image, but with a much smaller "footprint" and thus a faster transmission time, to say nothing of the smaller amount of storage space required. This was done at the sacrifice of losing a substantial amount of data. However, the image is still of good quality and, when applied to image scan present a remarkable representation of the original scene.
So, we can all make our choices for format. I'll stick with RAW, as I like working with the maximum amount of data pertinent to the image.
Sources:
Encyclopedia of Graphics File Formats - James D. Murray and William vanRyper
Multidisciplinary Approaches to Visual Representations and Interpretations - edited by Grant Malcolm
Data Structures In C - Noel Kalicharan
The JPEG still picture compression standard G.K. Wallace (Digital Equipment Corp., Maynard, MA, USA)
The number of quanta necessary for the perception of light of the human eye - HA Van Der Velden
Data compression apparatus and method using matching string searching and Huffman encoding - C Mayers, DL Whiting
http://www.tedmontgomery.com/the_eye/