Ugly Hedgehog - Photography Forum
Home Active Topics Newest Pictures Search Login Register
Main Photography Discussion
RAW and JPEG question
Page <<first <prev 10 of 10
Jun 16, 2019 11:50:09   #
BebuLamar
 
John_F wrote:
A 16 MP image yielding a 34 MB file means the pixels are 16 bit depth. The excess 2 MB file size would be file overhead - but a bit way high tho. At a 16 bit pixel depth the 3 color channels (RGB) could be 5 bits. When such a file comes into Photoshop, PS will create 8 bit color channels. PS can be adjusted to make greater bit depth color channels, but production color printers might not be able to handle such color detail except for certain specialty equipment.


Not 5 bit per color. In the RAW file each pixel only has 1 value and not 3 values (RGB). This is the greatest difference between RAW and any other image files formats.

Reply
Jun 16, 2019 12:09:43   #
John_F Loc: Minneapolis, MN
 
To my way of thinking a raw image file should represent the actual charge state of every charge center on the sensor. In digital terms this will be a long string of “0”s and “1”s. The job of the camera’s firmware is to organize this string into rows and columns with color representative. With my Sony a6300 a (6000 by 4000) 24 MP iraw image is 25+ MB as read by the computer. So a 14-bit pixel depth image ends up being 8-bit pixel depth in conputer. What is in the computer is what any image editing app will operate on and Photoshop will make a 3 byte color depth for each pixel. It is clear some algorithm does such.

My constant harp is that the entire problem is due to there being no standard raw file format. That is what the Joint Photographics Group did to make the JPEG image format. They had to accommodate the needs of all sectors of the pictures/illustrations industry.

Reply
Jun 16, 2019 12:09:59   #
DirtFarmer Loc: Escaped from the NYC area, back to MA
 
BebuLamar wrote:
That is what I said.


"...But by grouping a number of dots to create one pixel they are able to produce a continuous tone image with lower resolution"

They don't produce a continuous tone image. They produce the illusion of a continuous tone image.

Semantics.

Reply
 
 
Jun 16, 2019 12:13:15   #
DirtFarmer Loc: Escaped from the NYC area, back to MA
 
BebuLamar wrote:
Not 5 bit per color. In the RAW file each pixel only has 1 value and not 3 values (RGB). This is the greatest difference between RAW and any other image files formats.


Correct. We're talking about raw data here. Each pixel is one color from the Bayer filter. Once it's demosaiced, each pixel is 3 colors. At that point the pixel will need 48 bits.

Reply
Jun 16, 2019 13:07:43   #
Streets Loc: Euless, TX.
 
DirtFarmer wrote:
Correct. We're talking about raw data here. Each pixel is one color from the Bayer filter. Once it's demosaiced, each pixel is 3 colors. At that point the pixel will need 48 bits.


Demosaiced? Damn! This is gettin' interesting. What's next; "derawed" to describe jpeg?

Reply
Jun 16, 2019 13:13:15   #
BebuLamar
 
DirtFarmer wrote:
Correct. We're talking about raw data here. Each pixel is one color from the Bayer filter. Once it's demosaiced, each pixel is 3 colors. At that point the pixel will need 48 bits.


That what it does. When you open the RAW file in ACR and click on open image and the image opened in Photoshop. The amount of memory needed to hold the image is 3 times that of the RAW file.

Reply
Jun 16, 2019 14:00:47   #
PHRubin Loc: Nashville TN USA
 
BigDaddy wrote:
Not my way of thinking, it is how my 24 mb Nikon D5200 NEF file sizes come out. As you can see, they are very close to what I said. These are bytes btw, not bits.

Checking further, my old Panasonic fz 18 is an 8 mp camera and it's raw file sizes are around 13mb. Why these variations I don't know. I've read RAW files are compressed by some manufacturers, and this could be what is going on, I sure don't know what all these folks do with raw files, they are all proprietary and different. My Panasonic files are closer to what BebuLamar is getting with his camera, but not my Nikon.

What are your raw file sizes, and from what camera?
Not my way of thinking, it is how my 24 mb Nikon D... (show quote)


I really didn't know, so I just took a shot. The camera is an EOS 80D which has 24 Mp (6000X4000). The RAW file size for this one shot was 29MB.

I'm sure the internal firmware compresses the data somewhat. Each pixel must put out much more than one bit.

Reply
 
 
Jun 16, 2019 14:12:01   #
BebuLamar
 
PHRubin wrote:
I really didn't know, so I just took a shot. The camera is an EOS 80D which has 24 Mp (6000X4000). The RAW file size for this one shot was 29MB.

I'm sure the internal firmware compresses the data somewhat. Each pixel must put out much more than one bit.


Yes I think it was compressed. My 16MP camera in lossless compressed 14 bit mode the files are about 17 to 20 MB and they are varied. Uncompressed 14 bit mode it's 34MB.
But I think you meant one byte not one bit.

Reply
Jun 16, 2019 18:40:39   #
davyboy Loc: Anoka Mn.
 
I think you’re all nuts!!!!!

Reply
Jun 16, 2019 18:40:42   #
davyboy Loc: Anoka Mn.
 
I think you’re all nuts!!!!!

Reply
Jun 16, 2019 19:14:19   #
BebuLamar
 
davyboy wrote:
I think you’re all nuts!!!!!


Thank you!

Reply
 
 
Jun 16, 2019 19:46:06   #
DirtFarmer Loc: Escaped from the NYC area, back to MA
 
davyboy wrote:
I think you’re all nuts!!!!!


To paraphrase George Orwell:
We're all equal, but some of us are more equal.

Substitute "nuts" for "equal"


(Download)

Reply
Jun 17, 2019 11:30:56   #
John_F Loc: Minneapolis, MN
 
A pixel is just a name for a cluster of bits. Those bits can be devoted to whatever the firmware program causes them to represent. My point was that 16 bits can be divided into 3 groups of 5 bits plus 1 leftover bit. The three fives could be devoted for the Red, Green, and Blue color channels resulting into a lousy 32 color shades per channel. Eight-bit color (256 shades per channel) is generally standard for color printers. Any pixel depth that is less than 24 bits shortchanges color fidelity. See why compression is necessary.


BebuLamar wrote:
Not 5 bit per color. In the RAW file each pixel only has 1 value and not 3 values (RGB). This is the greatest difference between RAW and any other image files formats.

Reply
Jun 17, 2019 11:47:11   #
BebuLamar
 
John_F wrote:
A pixel is just a name for a cluster of bits. Those bits can be devoted to whatever the firmware program causes them to represent. My point was that 16 bits can be divided into 3 groups of 5 bits plus 1 leftover bit. The three fives could be devoted for the Red, Green, and Blue color channels resulting into a lousy 32 color shades per channel. Eight-bit color (256 shades per channel) is generally standard for color printers. Any pixel depth that is less than 24 bits shortchanges color fidelity. See why compression is necessary.
A pixel is just a name for a cluster of bits. Thos... (show quote)


But as I pointed out a RAW file has only 1 value for each pixel. This value can be 8, 12, 14 or even 16 bit but only 1 value and not 3 values like other type of image files.

Reply
Page <<first <prev 10 of 10
If you want to reply, then register here. Registration is free and your account is created instantly, so you can post right away.
Main Photography Discussion
UglyHedgehog.com - Forum
Copyright 2011-2024 Ugly Hedgehog, Inc.