Reading mode: Light Dark

RAW format / extended dynamic range?

Go to last post
M

maritime

Member
Messages
13
Reaction score
0
Location
US
See p.15 of the 300D-review:

"RAW latitude (digital exposure compensation)

Shooting RAW on the EOS 300D provides you with approximately one stop (1 EV) of additional latitude above the clipping point (pure white - 255,255,255) of the original exposure. In the example below you can see that a large area of the image was over exposed, applying a -0.6 EV digital exposure compensation to the RAW image retrieves some of this detail. "

What does it mean? Is it possible to exploit (if needed) this in any way to extend the dynamic range in the final jpg-image?
 
maritime said:
See p.15 of the 300D-review:

"RAW latitude (digital exposure compensation)
Shooting RAW on the EOS 300D provides you with approximately one
stop (1 EV) of additional latitude above the clipping point (pure
white - 255,255,255) of the original exposure. In the example below
you can see that a large area of the image was over exposed,
applying a -0.6 EV digital exposure compensation to the RAW image
retrieves some of this detail. "

What does it mean? Is it possible to exploit (if needed) this in
any way to extend the dynamic range in the final jpg-image?
Click to expand...
The JPEG image is restricted to 8-bits per colour, whereas the RAW image actually uses 12.

This means that where a pixel is very close to pure white (for example) the JPEG may not be able to record the subtle difference between the real colour and pure white so it has to call it pure white. The RAW image can record more of the subtle tones between what a JPEG call pure white and what really is pure white. Thta is what can be puled out from the RAW image.

Unfortunately, the JPEG has already discarded that data (because it can't record it in an 8-bit format) and what has been discarded cannot be recreated.

--
KRs
Chris

My meagre efforts are at http://www.dslr.co.uk
 
Chris Maddock said:
maritime said:
See p.15 of the 300D-review:

"RAW latitude (digital exposure compensation)
Shooting RAW on the EOS 300D provides you with approximately one
stop (1 EV) of additional latitude above the clipping point (pure
white - 255,255,255) of the original exposure. In the example below
you can see that a large area of the image was over exposed,
applying a -0.6 EV digital exposure compensation to the RAW image
retrieves some of this detail. "

What does it mean? Is it possible to exploit (if needed) this in
any way to extend the dynamic range in the final jpg-image?
Click to expand...
The JPEG image is restricted to 8-bits per colour, whereas the RAW
image actually uses 12.
This means that where a pixel is very close to pure white (for
example) the JPEG may not be able to record the subtle difference
between the real colour and pure white so it has to call it pure
white. The RAW image can record more of the subtle tones between
what a JPEG call pure white and what really is pure white. Thta is
what can be puled out from the RAW image.
Unfortunately, the JPEG has already discarded that data (because it
can't record it in an 8-bit format) and what has been discarded
cannot be recreated.

--
KRs
Chris

My meagre efforts are at http://www.dslr.co.uk
Click to expand...
Thank you for responding.

If you then shoot and save in the RAW-format - it actually keeps a lot more information in the 12 bits - i agree. But - is it possible to extract 2 jpeg images (with the Canon bundled RAW-conversion software and the digital exposure compensation technique explained in the review) first from the "bottom/black" of the 12 bits and second from the "top/highlight" of the 12 bits - and then later combine those jpeg's in PS? (like bracketing technique - where you actually shoot 2 or 3 separate pictures)?

-maritime-
 
maritime said:
Chris Maddock said:
maritime said:
See p.15 of the 300D-review:

"RAW latitude (digital exposure compensation)
Shooting RAW on the EOS 300D provides you with approximately one
stop (1 EV) of additional latitude above the clipping point (pure
white - 255,255,255) of the original exposure. In the example below
you can see that a large area of the image was over exposed,
applying a -0.6 EV digital exposure compensation to the RAW image
retrieves some of this detail. "

What does it mean? Is it possible to exploit (if needed) this in
any way to extend the dynamic range in the final jpg-image?
Click to expand...
The JPEG image is restricted to 8-bits per colour, whereas the RAW
image actually uses 12.
This means that where a pixel is very close to pure white (for
example) the JPEG may not be able to record the subtle difference
between the real colour and pure white so it has to call it pure
white. The RAW image can record more of the subtle tones between
what a JPEG call pure white and what really is pure white. Thta is
what can be puled out from the RAW image.
Unfortunately, the JPEG has already discarded that data (because it
can't record it in an 8-bit format) and what has been discarded
cannot be recreated.

--
KRs
Chris

My meagre efforts are at http://www.dslr.co.uk
Click to expand...
Thank you for responding.

If you then shoot and save in the RAW-format - it actually keeps a
lot more information in the 12 bits - i agree. But - is it possible
to extract 2 jpeg images (with the Canon bundled RAW-conversion
software and the digital exposure compensation technique explained
in the review) first from the "bottom/black" of the 12 bits and
second from the "top/highlight" of the 12 bits - and then later
combine those jpeg's in PS? (like bracketing technique - where you
actually shoot 2 or 3 separate pictures)?

-maritime-
Click to expand...
See:

http://www.robgalbraith.com/ubbthreads/showflat.php?Cat=&Board=UBB8&Number=161611&page=0&view=collapsed&sb=5&o=&fpart=1
Good luck!
--
Regards,
DaveMart
Please see profile for equipment
 
maritime said:
Thank you for responding.

If you then shoot and save in the RAW-format - it actually keeps a
lot more information in the 12 bits - i agree. But - is it possible
to extract 2 jpeg images (with the Canon bundled RAW-conversion
software and the digital exposure compensation technique explained
in the review) first from the "bottom/black" of the 12 bits and
second from the "top/highlight" of the 12 bits - and then later
combine those jpeg's in PS? (like bracketing technique - where you
actually shoot 2 or 3 separate pictures)?
Click to expand...
It should be, yes. I see you've already been given a link that should point you in the right direction, but I haven't gone to that extent myself. All I've done is to extract one image with positive exposure compensation and one with negative compensation then merge the two to simulate a gradual neutral density filter. That works well - however I don't use Canon's software so don't know how easy/good it is for that. I use Breezebrowser and/or Capture One.

--
KRs
Chris

My meagre efforts are at http://www.dslr.co.uk
 
Chris Maddock said:
The JPEG image is restricted to 8-bits per colour, whereas the RAW
image actually uses 12.
This means that where a pixel is very close to pure white (for
example) the JPEG may not be able to record the subtle difference
between the real colour and pure white so it has to call it pure
white.
Click to expand...
This actually doesn't have much to do with it. 8 bits gives you 256 steps per color channel. This means that any quantization error - the error from having to represent a value by one of several discrete steps - is, at most, 1/512 of the tonal range, which is 0.2%. An extra stop is a factor of 2. You can't find an extra stop by correcting a 0.2% error.

Also, remember that when you convert from RAW to JPEG on your computer, you also end up with an 8-bit/channel image. Thus, if you are pulling out an extra stop during the conversion, it must be representable by 8 bits in the end. This means that the camera could have done the same thing if it had wanted to.

The extra stop probably comes from one of two places. First, the camera can simply be struggling with how much of the highlights should be clipped in order to preserve an overall pleasing exposure, and may make the wrong choice in many cases. When you do the conversion yourself, you can make an aesthetic judgment that the camera cannot.

The second possible reason has to do with white balance. When an image is white balanced, the gains are adjusted for each color channel in order to make "white" white. At the same time, the camera tries to maintain the same overall brightness level. This means that even if all channels are fully captured without clipping in the RAW image, one or more channels can clip as the white balance is performed. The camera could avoid this by allowing the exposure to shift downward during white balance in order to make headroom for the shifting color channels. This, however, might darken the image unacceptably for some, and thus the camera errs on the side of good overall exposure. If you do your own conversion, you can avoid this.

Note that the RAW converter called Powershovel provides an option to control channel clipping during white balancing in order to boost dynamic range.

David
 
DRG said:
This actually doesn't have much to do with it. 8 bits gives you 256
steps per color channel. This means that any quantization error -
the error from having to represent a value by one of several
discrete steps - is, at most, 1/512 of the tonal range, which is
0.2%. An extra stop is a factor of 2. You can't find an extra stop
by correcting a 0.2% error.
Click to expand...
You're assuming that RAW is simply quantizing the same color space differently. This is unlikely.

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/
 
So, can't the algorithm compress the 12 bits of gradation into 8 bits instead of clipping the shadow and highlight detail?
Chris Maddock said:
maritime said:
See p.15 of the 300D-review:

"RAW latitude (digital exposure compensation)
Shooting RAW on the EOS 300D provides you with approximately one
stop (1 EV) of additional latitude above the clipping point (pure
white - 255,255,255) of the original exposure. In the example below
you can see that a large area of the image was over exposed,
applying a -0.6 EV digital exposure compensation to the RAW image
retrieves some of this detail. "

What does it mean? Is it possible to exploit (if needed) this in
any way to extend the dynamic range in the final jpg-image?
Click to expand...
The JPEG image is restricted to 8-bits per colour, whereas the RAW
image actually uses 12.
This means that where a pixel is very close to pure white (for
example) the JPEG may not be able to record the subtle difference
between the real colour and pure white so it has to call it pure
white. The RAW image can record more of the subtle tones between
what a JPEG call pure white and what really is pure white. Thta is
what can be puled out from the RAW image.
Unfortunately, the JPEG has already discarded that data (because it
can't record it in an 8-bit format) and what has been discarded
cannot be recreated.

--
KRs
Chris

My meagre efforts are at http://www.dslr.co.uk
Click to expand...
 
Thank you all for valuable links and answers.

My preliminary conclusion (after done some reading and following some links) is that RAW-exctraction cannot replace the Auto exposure bracketing function in 300D (manual p.76) for extended dynamic range - in difficult highlight/bright skies conditions. But sometimes it can - to a certain extent (+1 stop exposure level - eq. doubling the DR).

I will follow up after some study'ing - a lot of interesting links here.

(while waiting for my 300d to come..)

-maritime-
 
Do not save intermediate step images using Jpeg. Instead use a lossless format such as PSD or Tiff. If in the end you want a Jpeg hold off and make that your last step. Every-time you re-save a Jpeg you lose a little information (quality). Think of a photo copy of a photo copy, of a photo copy and you'll get the idea.
maritime said:
If you then shoot and save in the RAW-format - it actually keeps a
lot more information in the 12 bits - i agree. But - is it possible
to extract 2 jpeg images (with the Canon bundled RAW-conversion
software and the digital exposure compensation technique explained
in the review) first from the "bottom/black" of the 12 bits and
second from the "top/highlight" of the 12 bits - and then later
combine those jpeg's in PS? (like bracketing technique - where you
actually shoot 2 or 3 separate pictures)?
Click to expand...
 
Ron Parr said:
I wonder if it would have more potential with shots from SLRs since
the RAW file has more bit depth?
Click to expand...
Extremely possible. It's been my question exactly. I wouldn't mind going out with the G3 and a dSLR and shooting the same scene at similar exposure settings, both in RAW, and seeing what I'd gotten and where the improvement might be, if any.

--

Ulysses
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top