lmTrying Loc: WV Northern Panhandle

Thanks to several members, I have been reading several linked articles on RAW, jpeg, DNG, TIFF. When it comes to RAW, the articles are not consistent. One says that each pixel records one voltage reading. And the color recorded is governed by a filter on the pixel. That being said, a 24 MP camera would record 24 MB. Another article says each pixel records. A voltage for each color. That being said, a 24 MP camera would record way more than 24 MB, which would be more consistent with other things I have read.

Now, for my brain to understand all the other stuff down the digital pike, I really need to have a handle on 1). what one pixel actually produces and sends to the memory card. 2). do all pixels record and send the same, or different information?

Thanx in advance.

Oh yeah, I finally found and hooked up with a camera club yesterday evening.

Rongnongno Loc: FL

A pixel does not produce anything. A pixel is the result of a process created by a sensor. 'Sensors' as refereed in cameras are really sensor arrays. The number of sensors on a sensor array produce the recorded image pixel size (Mega Pixel). The pixel size influences the accuracy of the image recording.

Sensors using the Bayer system record four values: R, G, B and G. Depending on the sophistication of the camera hardware this create the color depth. The color depth refers to the number of possible colors and luminosity levels per pixel. In turn this influences the image size in bytes (of more accurately in MB).

The 'voltage method' of recording the image is changing. We now have 'invariant' sensors that will change the way we, as photographers, look at light.

A pixel does not send anything anywhere. If you are referring as to what a pixel is made of, read above.

As to each pixels recording the same thing? Think one second about this... If they did, the so-called image would be of a single color. Ask instead if the pixels use the same process to record an image and the answer would be 'Yes'. The output becomes different because each pixel is exposed to a different ray of light.

By the way all cameras start with the 'raw' capture, even those producing only JPG captures.

raw is a camera specific and not a standard, even within the line of camera offering from the same manufacturer, even if they use the same extension. This is why there is often a delay before all the software available can open a newer camera raw. A raw file usually contain a JPG thumbnail for quick viewing. Some programs ill offer a choice to see the raw data (after processing) or the JPG. The difference is that the raw created image appears of a lesser quality than the thumbnail.

JPG is a WEB standard used to show a digitized image, regardless of origin.

DNG is Adobe version of an universal format often confused with 'raw'. A DNG can encapsulate anything, including a JPG. Difference is that in theory the content then cannot be destroyed. DNG is not a standard.

TIFF is a standard. It is one of the most versatile format. One drawback: The size in MB is rather large.

In short: A pixel produces nothing. It is a colored dot that can B&W or a single color selected from trillions of combinations if the camera is capable of creating it and IF the format selected does not result the camera output. The associations of several pixels create shapes and when enough are around an image. How a pixel produced is BS unless you are an engineer.

Pablo8 Loc: Nottingham UK.

Have you ever just picked up a camera , and used it for taking pictures?

CHG_CANON Loc: the Windy City

Megapixels and file size (measured in bytes or megebytes MB) have virtually no relation to each other. The same photo, saved at different JPEG quality levels or different formats such as RAW or TIFF or DNG, etc, will yield wildly different file sizes in megabytes. In general more MP cameras will tend to create more MB-sized files, but there are too many variables involved to make an exact correlation between the sensor size and the resulting file size.

davidrb Loc: Half way there on the 45th Parallel

You have asked an unusual question. Most folks are here to try to learn how to operate a camera. You seem more inclined to try to build one. Perhaps sysops will open a new section to accomodate this subject but there hasn't been much demand til now. Maybe the people singing the extinction of dslrs should listen in on this thread.

d3200prime

Deleted

Steve Perry Loc: Sylvania, Ohio

This link will help it all make sense:

https://www.cambridgeincolour.com/tutorials/camera-sensors.htm

warrior Loc: Paso Robles CA

8 bits to a byte. Getting into binary

Garyminor Loc: Raleigh, NC

A pixel is a picture element! That is, a spot of color that makes up a picture. It has three values. One each for red, green, and blue. This is picture data.

A single data value (datum) from the sensor represents the intensity of a single color, either red, green, or blue. This is not a pixel! This is raw data.

Notice the use of the expression raw or picture data as opposed to a raw or picture file.
Raw data is contained in many different file formats such as CR2, NEF, DNG, etc.
Picture data is contained in other file formats such as JPEG, TIFF, etc.

There is a process that produces an array of pixels (a picture) from an array of sensor elements (raw data).
You might think that that data from four sensor elements (red, green, blue, green) would be used to create a single pixel (red, green, blue).
However, it's much more elaborate than that. Generally, a large number of sensor elements are used to produce a single pixel.
The neighboring pixels could reuse a lot of these sensor values so that there could be more, less, or the same number of pixels as there are sensor elements. It's generally assumed that the process creates the same number of pixels as there are sensor elements, although that is not required.

Do you need to know this to take good pictures? No, but I enjoy the mental exercise.

SalvageDiver Loc: Huntington Beach CA

Hi ImTrying,

To answer your question simply and directly, there is one color filter, either red, green OR blue placed over each photocell. There is not multiple color filters over a single photocell.

This is often a confusing subject for anyone but engineers. The article Steve Perry referred to is a good description. Here is another good primer from Adobe for your casual reading pleasure.

www.adobe.com/digitalimag/pdfs/understanding_digitalrawcapture.pdf

First, terminology. Sensors, photocells, photosenstive sites, sensor elements and pixels are often used interchangeably when discussing image sensors. In general, when one refers to a sensor pixel (short for picture element), they're just referring to a single photocell on an image sensor. A 24MP (megapixel) sensor has 24 million independent photocells (or light collecting elements, pixels etc..).

All a photocell does is collect photons and convert to an electrical signal. The photocell, by itself, has 'no' ability to distinguish color.

To determine color, color filters are used. Most cameras use a 'Bayer' color filter array (CFA). It's a mosaic of Red, Green and Blue color filters arranged is a very specific pattern. This is fully described in the various referenced articles. But, a 24MP image sensor will have a Bayer CFA that has 24 million individual color filter elements placed on top of it. It has one color filter for each sensor photocell.

in the raw image, each pixel is associated with only one color. The raw image before raw processing is just a grayscale image. To get the 3 channels of color information for each pixel, that we are all familiar with, requires complicated mathematical (demosiacing) algorithms to interpolate the other two colors from its nearest neighbors. This is the role of the Raw Processors, either in the camera or in the various post-processing applications.

To address what is sent to the memory card depend of whether you are storing raw or jpg files. Raw files store the basic grayscale (one color per pixel) image directly from the sensor to the file. Your PP application contain the raw processors necessary to convert that grayscale image into the 3-channel color image you see on your monitor or printer. For jpg's, all the work is done in-camera using the camera's built-in raw processor. There are advantages and disadvantages to either approach, but that's for a much different (and often passionate) discussion.

Mike

lmTrying Loc: WV Northern Panhandle

Rongnongno:

Thank you for your enlightening and gentle response. I understand my mistakes in terminology, at 1 in the morning the brain gets a little foggy in the same way terminology gets miss-used in conversation. Thank you for your kind corrections. A teaching fellow of mine always told his students that understanding the terminology was a big part of the understanding of any subject.

The "Bayer system" is a new term to me, but I understand each single sensor records 4 "voltage" (using this term loosely) values, that will assign somewhere between 256 to 16,384 gradients of color and brightness. If I understand it correctly, if several neighboring sensors are reading the same values, then those values get "compressed" similar to a jpeg. And this would explain why RAW files are not the same MegaByte size. I assume there are other systems besides Bayer and that is why there are 12 bit and 14 bit bytes in different raw formats.

I did say it poorly that each pixel records the same thing. I did mean, does each sensor use the same process of recording and sending information.

From my readings, I was assuming that all digital cameras capture in raw and then record to the memory card as directed by the program the manufacturer deems appropriate for that particular camera. I am also assuming that the differences in raw format are due to improvements in "programing". (again, using this term loosely.) And I am assuming that Adobe created DNG to give them consistency in their own programs, not provided by the different camera manufacturers. In other words, it's easier to convert all the different raw formats into one dng format, then work from there.

And finally, a pixel is a product of the information provided by the sensor array. A pixel can vary in size if the surrounding pixels are the same. Many pixels are combined to make a "dot", as in a Dots Per Inch from a printer. The exact science of how all this is accomplished, is nothing I have to understand.

Again, thank you. How am I doing. Did I get this correct?

lmTrying Loc: WV Northern Panhandle

Yes. Part of the problem is that I have thousands of digital pictures. Many taken in raw, and the rest in, I assume, jpeg. Having started in film, taking the picture has been relatively easy. Now I am preparing for the second half of the digital process, Post Processing. This second half of the process may become complicated by the fact that I may be doing a lot of traveling in an RV. Therefore I am considering using Corel's AfterShot and PaintShop Pro. I have dabbled with Picassa, Canon DPP, Vista, PhotoShop Express and a few others. I want to focus, reorganize and start producing. Workflow is another term I have to get a handle on. But I think my starting point for understanding all this is to understand the concepts of my first question. I have a few photos that I have printed. But I have been looking at the work posted on this forum for the last year and now I want to try my hand at making a few real wall hangers. To this point, I have been alone on a deserted island with only UHH as a reference and guide and a few magazine style books I have found in BAM. I have finally hooked up with a local camera club, so maybe now I can move on up.

lmTrying Loc: WV Northern Panhandle

Thank you for your input. Actually I understand the things you said. The reference I made by saying, "That being said," was a sort of short hand way of tagging approximate numbers to different concepts. It's a real challenge to know how much, or how little information to put into a post, question or answer, to keep everything flowing, not confusing, and not boring. Thanks

lmTrying Loc: WV Northern Panhandle

I have grown up for 67 years being told that I ask questions that most other people don't even think of. For the most part, I know how to operate my camera. I am by no means perfect, nor even close. Professionals need not fear me as a competitor. I do get lucky once in a long while. What I am inclined to do is get the information off the memory card and produce some images that I might like to hang on my wall. How to do this starts with understanding something about raw, dng, tiff, and jpeg formats. I would rather build model car kits than cameras. As far as the extinction of dslr cameras, even if they are replaced with mirrorless as some predict, mirrorless still have sensor arrays. And as far as smaller and lighter goes, In the early 80's I read up on every affordable slr camera being manufactured. I narrowed it down to 4. The skinny little Minolta scared me to pick it up and I couldn't wait to set it back down on the counter. That's when I became a Canon shooter. I have recently considered a Canon mirrorless, but just looking at them reminds me of the Minolta. Don't start a new section in the Hogg just for me.

lmTrying Loc: WV Northern Panhandle

Thanks Steve for the link. However, it will take quite a while to digest all the information on this site. As to the specific link on the Bayer array, I have seen this rgbg filter before, but not explained as deeply as here. The bottom line of this explanation, as I understand it, is that each singular sensor consists of 4 wells. So a 4MP cameras sensor array has 16,000,000 wells. Am I right?