Essential "raw" (Page 3): My understanding is that a raw file is one that is not processed by the camera. Just about every model camera has its own particular specs for raw.

jerryc41 Loc: Catskill Mts of NY

rehess wrote:

I can't find exactly what I want already, and I do... (show quote)

My understanding is that a raw file is one that is not processed by the camera. Just about every model camera has its own particular specs for raw.

rehess Loc: South Bend, Indiana, USA

jerryc41 wrote:

My understanding is that a raw file is one that is not processed by the camera. Just about every model camera has its own particular specs for raw.

But changing the size of a file, including "downscaling", is a form of "processing".

Notorious T.O.D. Loc: Harrisburg, North Carolina

Computers are all about the bits and bytes...they only deal in on or off bits. Any data is a stream of on and off bits and as soon as you start organizing it you are manipulating it. Much of how computers really work is so far removed from users and even developers compared to 30-40 years ago. Even 35 years ago there were not that many programmers that worked well at assembly level programming in my experience. Many tricks had to be used because processing power and memory were so limited, very limited.

PHRubin wrote:

My understanding; Bits are elemental parts of a digital file. Pixels are elemental parts of an image. RAW data is as it comes off a sensor in digital form, and isn't an image until a conversion is done. If, indeed, there are different precisions or noise reduction processes done to the RAW data, it is between truly raw and an image and can be called anything.

lamiaceae Loc: San Luis Obispo County, CA

rehess wrote:

I can't find exactly what I want already, and I do... (show quote)

I can set a few things in my Pentax cameras, such as noise reduction but don't. Not sure if N.R. only affects JPGs and/or RAW as I shoot Raw and leave most setting at null. Raw is RAW, truly Raw is likely rhetoric.

gvarner Loc: Central Oregon Coast

If a manufacturer offers various in-camera sizes for its output, always use the highest and ignore the others.the whole pie is better than part of it.

Novicus Loc: north and east

BebuLamar wrote:

All RAW files from the same camera have the same n... (show quote)

Did you mean " Photosites " ?

As I understand it , a Pixel consists of Four Photosites ...2 photosites see Green..1 red..1 Blue...so then the software makes a Pixel out of it , does that not mean that the " cooking " already starts there ?

rehess Loc: South Bend, Indiana, USA

Novicus wrote:

Did you mean " Photosites " ?

As I understand it , a Pixel consists of Four Photosites ...2 photosites see Green..1 red..1 Blue...so then the software makes a Pixel out of it , does that not mean that the " cooking " already starts there ?

I guess this is another place where definitions become important. As a computer scientist, I never considered moving data from one place to another to be "processing" - in fact, we don't count such steps when determining the "complexity" of an algorithm, + and - are level 1, while * and / are level 2.

Apparently the whole process of creating a "raw" file is more complex than I had been aware - at least some "raw" files are not as close to the hardware as I had thought.

MrBob Loc: lookout Mtn. NE Alabama

Too confusing and the thread has gotten totally convoluted.... everyone knows what bits and bytes are and as an assembly language programmer of old I used to think in Hexadecimal; I will use the largest RAW file available and convert that to the TIFF of my choosing. Too much interpolation going on now in a Bayer array... are we cooking at that level ?

blackest Loc: Ireland

Novicus wrote:

Did you mean " Photosites " ?

As I understand it , a Pixel consists of Four Photosites ...2 photosites see Green..1 red..1 Blue...so then the software makes a Pixel out of it , does that not mean that the " cooking " already starts there ?

Four photosites is 4 pixels say you have a green site either side of it are 2 blue sites and 2 red sites by interpolating the values either side and in the general area you can estimate pretty well what the missing values would be and it pretty much works out. thats for the bayer sensor for fuji they have a different method.

gvarner Loc: Central Oregon Coast

My Nikon D7000 does not allow varying sizes of its RAW file, only the JPEG's. I can't imagine the usefulness of any in-camera process that would reduce the size of a RAW file. By its very nature it has only one file size. Even when you edit a RAW you are not changeing it. You are creating a sidecar file with all the edits.

blackest Loc: Ireland

gvarner wrote:

My Nikon D7000 does not allow varying sizes of its RAW file, only the JPEG's. I can't imagine the usefulness of any in-camera process that would reduce the size of a RAW file. By its very nature it has only one file size. Even when you edit a RAW you are not changeing it. You are creating a sidecar file with all the edits.

maybe depends on circumstance, e.g say you were shooting at a high iso with reduced dynamic range you might be able to express the data in 7 bits rather than 14 this would pretty much reduce the file size by half. maybe you could read the values faster too, i don't know but maybe that would increase burst speed and duration. isn't that one of the reasons for reducing pixel count in order to increase burst speed??

I am just guessing, i don't design sensors but it must have been considered.

Peterff Loc: O'er The Hills and Far Away, in Themyscira.

gvarner wrote:

Even when you edit a RAW you are not changeing it. You are creating a sidecar file with all the edits.

That may be entirely dependent on the type of raw file and the software used. Canon .CR2 files and with DPP keep the edits in the .CR2 file, no side cars. The raw image data doesn't get changed, but the edits are bundled inside the .CR2 file, so the .CR2 file itself can be changed.

Notorious T.O.D. Loc: Harrisburg, North Carolina

I would say burst speed is more a function of the shutter and the buffer size. Smaller file size will allow more shots into the buffer before it fills and has to slow down. The buffer size will is important and how fast it fills depends on how fast the data can be moved to it, how much data and how fast the buffer can write to the card... My camera never fills the buffer IIRC if shooting jpegs. I think someone did a test where they just held the shutter button for 5 or 6 minutes...took like 5000 shots... Neat, but not practically needed...

blackest wrote:

maybe depends on circumstance, e.g say you were sh... (show quote)

Novicus Loc: north and east

blackest wrote:

Four photosites is 4 pixels say you have a green site either side of it are 2 blue sites and 2 red sites by interpolating the values either side and in the general area you can estimate pretty well what the missing values would be and it pretty much works out. thats for the bayer sensor for fuji they have a different method.

I was under the impression that Fuji use the bayer configuration , with smaller and Bigger photosites , perhaps you are refering to the FOVEON sensor from Sigma ?

blackest Loc: Ireland

Novicus wrote:

I was under the impression that Fuji use the bayer configuration , with smaller and Bigger photosites , perhaps you are refering to the FOVEON sensor from Sigma ?

You are quite right, i was thinking of the foveon sensor from Sigma.

As it happens fuji does have a patent on a 3 layer sensor

https://www.mirrorlessrumors.com/fuji-patent-discloses-a-new-three-layer-sensor-foveon-style/

As far as i can tell sigma make 1 dslr the 46 Mpix sd1 merrill.

https://www.sigmaphoto.com/cameras/dslr/sd1-merrill-digital-slr-camera-0

It's not cheap.