Histogram question

[Alpha_7]

I was trawling on flickr and noticed an interesting thread I thought the idea of marking the stops was a good one but what really got my attention was the third comment where the suggestion was that the histogram shown on screen on the back of a Nikon when shooitng RAW is based on the jpeg intrepretation. I don't understand the histogram enough myself to know if this is true or not, but if we are using the histogram as a guide to our exposure should it matter if we are shooting Jpeg or RAW ? I'd expect the same behaviour either way ?

[MATT]

I have no real answer for you, but it mentioned is showed the histogram of an embedded jpg in the RAW file.

I would not have thought that a RAW file had an embedded jpg?? Isnt hat why we shoot raw?


Cheers
MATT

[moz]

To the best of my knowledge all raw formats have an embedded jpeg, if only so the camera doesn't have to have the ability to read a raw file off the card then process it (the hardware is just not there), and to greatly speed up display of raw thumbnails. It makes sense to process the thumbnail to get the histogram (less dtata = smaller), and I'm sure I've seen other references to that approach missing small highlights and other problems. The alternative would be to process the whole image to jpeg, read out the histograph, then dump the jpeg. Sounds like lots of work for a small gain, and I suspect you'd get a further slowdown in RAW speed out of it. But maybe not, it depends where the bottlenecks are (remember that raw is compressed, and that compression has to be done somewhere... almost certainly the same CPU as does the jpeg transformation).

[Alpha_7]

Well I know Nikon use embedded Jpegs, and Moz what you said makes sense so I guess the Histogram is a guide, but due to the Jpeg compression it isn't 100% accurate, but for most shots a damn good indication of what the exposure was like.


Looks like the conversation at flickr continues too
http://www.flickr.com/groups/strobist/d ... 315731961/