Some raw-related things we often read that are wrong

Iliah Borg

Forum Pro
Messages
29,605
Solutions
26
Reaction score
26,453
Location
AK, US
Below is a short list of things that spring to mind. Please add your favorites.
  • The readability of raw files is not guaranteed in future
  • Raw files are dark and green
  • Raw is not an image
  • JPEGs are 8-bit RGB, and ready for display
  • JPEG histogram is a good substitution for raw histogram when it comes to determining the optimal exposure and ETTR
  • Image looks dark when it is underexposed
  • Sensor gamut
  • Sensor colour reproduction
  • Exposure triangle (shutter speed / aperture / ISO)
  • High ISO noise
  • ISO changes sensor sensitivity
  • Metering system is calibrated for 18% grey
  • Cameramakers cheat with ISO
  • CCD is better than CMOS
 
Oh - yeah - like that about the Exposure triangle.

:-)

Though, Iliah - does it specifically relate to RAW?
 
Oh - yeah - like that about the Exposure triangle.

:-)

Though, Iliah - does it specifically relate to RAW?
Tangentially, I think, but still for a raw shooter exposure triangle school of thought is IMHO more unwanted.

It is often suggested that brightness of an image == exposure. JPEGs are always affected by ISO setting. For raw files, at some ISO settings ISO is not applied at all, instead the files contain an instruction to apply brightening in conversion. In some cameras ISO does not affect raw data at any ISO settings. In some cameras intermediate ISO settings are a combination of the higher ISO stop and downscaling the raw data. Some cameras are "ISO-less", etc.

IMHO Making ISO a part of the exposure for a raw shooter is hardly even remotely acceptable.
 
  • Raw files are dark and green
Care to elaborate on this? If I understand it correctly, the BFA is more sensitive to the green channel so that under daylight (and similar) light that channel has a higher recorded luminance than the red and blue channels.
 
  • Raw files are dark and green
Care to elaborate on this? If I understand it correctly, the BFA is more sensitive to the green channel so that under daylight (and similar) light that channel has a higher recorded luminance than the red and blue channels.
To view an image file correctly, we need to respect the explicit and implicit (default) metadata that is associated with the file.

Suppose we have a TIFF in a Lab space. By default, implicitly, we consider its white point to be D50. If the file is recorded for D65, or any other white point that is different from D50, and we ignore this condition while converting it to be displayed, the colour turns out wrong. Same with raw files, if we ignore the white point (given as white balance), the colour on the resulting image looks green or in some other way tinted. If we ignore the black level in a raw file, or set it incorrectly, we will have magentish look, with low contrast due to "magenta veiling". If we ignore the saturation point, we are getting magenta highlights, and so on.

"Dark" is also an issue with implicit metadata being ignored. Raw file is linear, attempting to display it without supplying the correct gamma = 1 results in very compressed shadows.

Here is a short article that covers some of the aspects: https://www.fastrawviewer.com/viewing-raw-is-not-impossible

--
http://www.libraw.org/
 
Last edited:
Oh - yeah - like that about the Exposure triangle.

:-)

Though, Iliah - does it specifically relate to RAW?
Tangentially, I think, but still for a raw shooter exposure triangle school of thought is IMHO more unwanted.

It is often suggested that brightness of an image == exposure. JPEGs are always affected by ISO setting. For raw files, at some ISO settings ISO is not applied at all, instead the files contain an instruction to apply brightening in conversion. In some cameras ISO does not affect raw data at any ISO settings. In some cameras intermediate ISO settings are a combination of the higher ISO stop and downscaling the raw data. Some cameras are "ISO-less", etc.

IMHO Making ISO a part of the exposure for a raw shooter is hardly even remotely acceptable.
 
Below is a short list of things that spring to mind. Please add your favorites.
  • The readability of raw files is not guaranteed in future
  • Raw files are dark and green
  • Raw is not an image
  • JPEGs are 8-bit RGB, and ready for display
  • JPEG histogram is a good substitution for raw histogram when it comes to determining the optimal exposure and ETTR
  • Image looks dark when it is underexposed
  • Sensor gamut
  • Sensor colour reproduction
  • Exposure triangle (shutter speed / aperture / ISO)
  • High ISO noise
  • ISO changes sensor sensitivity
  • Metering system is calibrated for 18% grey
  • Cameramakers cheat with ISO
  • CCD is better than CMOS
 
You have to shoot RAW or you aren't a real photographer.

You take so much pride in shooting RAW that you add it to your signature line.
 
  • Raw files are dark and green
Care to elaborate on this? If I understand it correctly, the BFA is more sensitive to the green channel so that under daylight (and similar) light that channel has a higher recorded luminance than the red and blue channels.
To view an image file correctly, we need to respect the explicit and implicit (default) metadata that is associated with the file.

Suppose we have a TIFF in a Lab space. By default, implicitly, we consider its white point to be D50. If the file is recorded for D65, or any other white point that is different from D50, and we ignore this condition while converting it to be displayed, the colour turns out wrong. Same with raw files, if we ignore the white point (given as white balance), the colour on the resulting image looks green or in some other way tinted. If we ignore the black level in a raw file, or set it incorrectly, we will have magentish look, with low contrast due to "magenta veiling". If we ignore the saturation point, we are getting magenta highlights, and so on.

"Dark" is also an issue with implicit metadata being ignored. Raw file is linear, attempting to display it without supplying the correct gamma = 1 results in very compressed shadows.

Here is a short article that covers some of the aspects: https://www.fastrawviewer.com/viewing-raw-is-not-impossible
Thank you for that. I will definitely being reading up on this and may come back with more questions.
 
  • Raw files are dark and green
Care to elaborate on this? If I understand it correctly, the BFA is more sensitive to the green channel so that under daylight (and similar) light that channel has a higher recorded luminance than the red and blue channels.
To view an image file correctly, we need to respect the explicit and implicit (default) metadata that is associated with the file.

Suppose we have a TIFF in a Lab space. By default, implicitly, we consider its white point to be D50. If the file is recorded for D65, or any other white point that is different from D50, and we ignore this condition while converting it to be displayed, the colour turns out wrong. Same with raw files, if we ignore the white point (given as white balance), the colour on the resulting image looks green or in some other way tinted. If we ignore the black level in a raw file, or set it incorrectly, we will have magentish look, with low contrast due to "magenta veiling". If we ignore the saturation point, we are getting magenta highlights, and so on.

"Dark" is also an issue with implicit metadata being ignored. Raw file is linear, attempting to display it without supplying the correct gamma = 1 results in very compressed shadows.

Here is a short article that covers some of the aspects: https://www.fastrawviewer.com/viewing-raw-is-not-impossible
Thank you for that. I will definitely being reading up on this and may come back with more questions.
It's a good topic to discuss, yes.
 

Keyboard shortcuts

Back
Top