Home · Register · Search · View Winners · Software · Hosting · Software · Join Upload & Sell

Moderated by: Fred Miranda
Username   Password

  New fredmiranda.com Mobile Site
  New Feature: SMS Notification alert
  New Feature: Buy & Sell Watchlist
  

FM Forums | Canon Forum | Join Upload & Sell

1      
2
       end
  

Archive 2013 · Even more shocking news. True RAW histogram added to 5D3!
  
 
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.2 #1 · Even more shocking news. True RAW histogram added to 5D3!


EB-1 wrote:
What is wrong with the histogram in RAW mode now? Canon histograms have been fine for years other than the lack of right border.

EBH


They are not bad but they are based upon the processed JPG image and don't really tell you about whether the RAW file is blown or not just whether something is blown in the particular (and limited) color space of jpg and tone mapping used for the JPG. If you are really trying to squeeze out all the DR you can it's more accurate to go with a true RAW histogram.



May 14, 2013 at 02:30 AM
RichFisher
Offline
• •
Upload & Sell: On
p.2 #2 · Even more shocking news. True RAW histogram added to 5D3!


Dramatically improved highlights.

Wonder what a raw histogram looks like. The JPG histogram, which we normal see, will be very different from the raw histogram.



May 14, 2013 at 03:32 AM
kevinsullivan
Offline
• • •
Upload & Sell: On
p.2 #3 · Even more shocking news. True RAW histogram added to 5D3!


They are not bad but they are based upon the processed JPG image and don't really tell you about whether the RAW file is blown or not

A pertinent question for non-ML shooters, then, is, "What JPEG settings should one use to get the camera to display a histogram that is closest to the RAW histogram?" And, second, "How far off is it?"

I'm guessing that the "Neutral" picture style is the right choice, as it's intended to produce the greatest latitude for postprocessing (of JPEG files). It's low in contrast, saturation, etc. A good "experiment" for someone to perform might be to test the hypothesis that, "The neutral pictures style gives an in-camera histogram that is quite close to the actual raw histogram." One might for example compare the in-camera histogram against the results produced by dcraw & rawhistogram.exe, for a range of images.

Kevin



May 14, 2013 at 03:09 PM
gdanmitchell
Online
• • • • •
Upload & Sell: On
p.2 #4 · Even more shocking news. True RAW histogram added to 5D3!


If you use the older 5D2 and live view, in addition to relying on the histogram you can can find out a lot about highlights by zooming in on the very bright areas and inspecting them. I can usually tell if, despite what the histogram suggests, I might be oversaturatiing or blowing out highlights.

Dan



May 14, 2013 at 03:21 PM
super35
Offline
• •
Upload & Sell: Off
p.2 #5 · Even more shocking news. True RAW histogram added to 5D3!


kevinsullivan wrote:
A pertinent question for non-ML shooters, then, is, "What JPEG settings should one use to get the camera to display a histogram that is closest to the RAW histogram?"
Kevin


Kevin, have a look at the new "Video Camera X Series Look" picture style from Canon. You can download from the Canon Japan site here.

Just looking at in DPP it has more latitude and lower saturation than any other styles, including neutral and faithful and works on any Canon DSLR with picture styles going all the way back to the original 5D.

My recommended settings are 0,-4,0,0

Sharpness 0
Contrast -4
Saturation 0
Color Tone 0



May 14, 2013 at 03:47 PM
kevinsullivan
Offline
• • •
Upload & Sell: On
p.2 #6 · Even more shocking news. True RAW histogram added to 5D3!


Thanks, super35.


May 14, 2013 at 04:48 PM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.2 #7 · Even more shocking news. True RAW histogram added to 5D3!


kevinsullivan wrote:
A pertinent question for non-ML shooters, then, is, "What JPEG settings should one use to get the camera to display a histogram that is closest to the RAW histogram?" And, second, "How far off is it?"

I'm guessing that the "Neutral" picture style is the right choice, as it's intended to produce the greatest latitude for postprocessing (of JPEG files). It's low in contrast, saturation, etc. A good "experiment" for someone to perform might be to test the hypothesis that, "The neutral pictures style gives an in-camera histogram that is quite close to the actual raw histogram." One might for example
...Show more

There wss a custom version that someone made and instruction for replicating it. It looks pretty weird though and looking at images on the back it's hard to make sense of them at all so you have to ultra trust the histogram.

And as mentioned the new cinema x might be a good choice for now. And intermediate compromise, not nearly as good as RAW histogram but one that still lets you look at the image and kinda get a clue unlike with what I mentioned just above.




May 14, 2013 at 06:45 PM
RichFisher
Offline
• •
Upload & Sell: On
p.2 #8 · Even more shocking news. True RAW histogram added to 5D3!


super35 wrote:
Kevin, have a look at the new "Video Camera X Series Look" picture style from Canon. You can download from the Canon Japan site here.

Just looking at in DPP it has more latitude and lower saturation than any other styles, including neutral and faithful and works on any Canon DSLR with picture styles going all the way back to the original 5D.

My recommended settings are 0,-4,0,0

Sharpness 0
Contrast -4
Saturation 0
Color Tone 0


The advantage of keeping contrast set higher when shooting raw, is that you will get an early warning indication of clipping.



May 14, 2013 at 07:19 PM
kevinsullivan
Offline
• • •
Upload & Sell: On
p.2 #9 · Even more shocking news. True RAW histogram added to 5D3!


The advantage of keeping contrast set higher when shooting raw, is that you will get an early warning indication of clipping.

Put another way, you get told there's clipping even when there really isn't any. I think the point of this thread is that in some cases what some people want is an accurate in-camera indication as to whether sensor saturation really did occur.



May 14, 2013 at 07:22 PM
 

Search in Used Dept. 



andyjaggy82
Offline
• • •
Upload & Sell: Off
p.2 #10 · Even more shocking news. True RAW histogram added to 5D3!


I had no idea that the histogram was based of the jpeg and not the actual raw information. I just learned something new.


May 14, 2013 at 07:38 PM
Stoffer
Offline
• • •
Upload & Sell: Off
p.2 #11 · Even more shocking news. True RAW histogram added to 5D3!


skibum5 wrote:
If you are really trying to squeeze out all the DR you can it's more accurate to go with a true RAW histogram.


You know what I have been pondering about lately? A metering mode that exposures to the right based on the raw-info. Sure it is Do-it-on-you-own-risk metering, but boy wouldn't it be awesome to have when you know you just want to save those highlights by a hair and do the rest in post for superior DR? Are you reading me, skibum5?



May 14, 2013 at 07:42 PM
kevinsullivan
Offline
• • •
Upload & Sell: On
p.2 #12 · Even more shocking news. True RAW histogram added to 5D3!


A metering mode that exposures to the right based on the raw-info

Yup, I, too, have been thinking about this for a long time. The problem, I suspect, is that the metering sensors and the recording sensors are not the same, and that to make it work the camera would have to take multiple exposures, as it hunts for an STTR exposure. Doable (certainly doable if one has software control over camera ops), but this is maybe not exactly what people are looking for.



May 14, 2013 at 07:56 PM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.2 #13 · Even more shocking news. True RAW histogram added to 5D3!


kevinsullivan wrote:
Put another way, you get told there's clipping even when there really isn't any. I think the point of this thread is that in some cases what some people want is an accurate in-camera indication as to whether sensor saturation really did occur.


It can be very confusing for certain intense colors like on some flowers where just the color space used by jpg alone clips so early. A deep red rose will clip like mad in sRGB when it's only moderately clipping in AdobeRGB and not clipping at all in ProPhotoRGB or RAW. The clipping to color space can be such a large extra factor it had be hard to judge how to expose certain intense flowers. If you save it all by the jpg histogram you expose the rest of the image crazy dark but if you let it clip too too much it might even clip in raw.



May 14, 2013 at 07:59 PM
kevinsullivan
Offline
• • •
Upload & Sell: On
p.2 #14 · Even more shocking news. True RAW histogram added to 5D3!


A deep red rose will clip like mad in sRGB when it's only moderately clipping in AdobeRGB and not clipping at all in ProPhotoRGB or RAW.

Indeed. Even though I don't really use JPEG outputs, I still do care to have by camera set to AdobeRGB for just this reason: to avoid undue "false clipping" in my RGB histogram when shooting saturated scenes.



May 14, 2013 at 08:05 PM
Stoffer
Offline
• • •
Upload & Sell: Off
p.2 #15 · Even more shocking news. True RAW histogram added to 5D3!


kevinsullivan wrote:
Indeed. Even though I don't really use JPEG outputs, I still do care to have by camera set to AdobeRGB for just this reason: to avoid undue "false clipping" in my RGB histogram when shooting saturated scenes.


Jep, a good tip right there.



May 14, 2013 at 08:07 PM
raydee
Offline

Upload & Sell: Off
p.2 #16 · Even more shocking news. True RAW histogram added to 5D3!


The real problem with the OEM implementation besides being based on the "cooked" low res preview jpeg is that it "averages" the blown highlight warning between the channels using only the luminance signal. Thus it might not display the blinking warning when there is a strong biased or monochromatic highlight (bright blueish sky, yellow red sunset etc) with only one channel blown.


May 15, 2013 at 12:41 PM
MintMar
Offline
• •
Upload & Sell: Off
p.2 #17 · Even more shocking news. True RAW histogram added to 5D3!


kevinsullivan wrote:
Yup, I, too, have been thinking about this for a long time. The problem, I suspect, is that the metering sensors and the recording sensors are not the same, and that to make it work the camera would have to take multiple exposures, as it hunts for an STTR exposure. Doable (certainly doable if one has software control over camera ops), but this is maybe not exactly what people are looking for.


With the AE sensors in the hump this seems impossible to do IMO, but flip that mirror up, I think it would be technically possible in live view, maybe with some small lag.



May 15, 2013 at 01:19 PM
Deborah Kolt
Offline
• •
Upload & Sell: On
p.2 #18 · Even more shocking news. True RAW histogram added to 5D3!


kevinsullivan wrote:
A pertinent question for non-ML shooters, then, is, "What JPEG settings should one use to get the camera to display a histogram that is closest to the RAW histogram?" And, second, "How far off is it?"

I'm guessing that the "Neutral" picture style is the right choice, as it's intended to produce the greatest latitude for postprocessing (of JPEG files). It's low in contrast, saturation, etc. A good "experiment" for someone to perform might be to test the hypothesis that, "The neutral pictures style gives an in-camera histogram that is quite close to the actual raw histogram." One might for example
...Show more

Switching to Neutral was what CPS recommended when I was unhappy about the preview on the 1Dx, and I find it considerably closer to actual output exposure. Apparently they tweaked Standard again to be punchier like lower cost cameras to make it easier for users upgrading for the first time. I'm still unhappy with white balance rendition on the 1Dx previews, but at least now I can accurately judge exposure.



May 17, 2013 at 04:39 AM
1      
2
       end




FM Forums | Canon Forum | Join Upload & Sell

1      
2
       end
    
 

You are not logged in. Login or Register

Username   Password    Retrive password