How many stops a Canon raw file can be recorded:::

[andylo]

Ok, I am not even sure if I am asking question correctly here... but I give it a try anyway...

So we all know the RAW actually captures more details than JPEG, hence the RAW file size are bigger, and also why we pretty much recommend everyone show at RAW.

But how many stops of information a RAW (in Canon) can be captured? I guess the answer would be something like "n stop +/- from median point"??

And in contrast, how many stops the JPEG can capture?

Hope you understand what I am talking/asking about :p

[kiwi]

I'll suggest 7 (+/- 3 + orig) for a raw
I'll suggest 4 (+/- 1.5 + orif) for a jpeg

[etherial]

It is the same, the sensor still see's the same thing, the only difference is how the file is saved. Saving in RAW keeps all the data whereas JPEG chucks some out and leaves a courser resolution. The range from black to white is the same, RAW just has many more shades in between.

If they were any different, when you set your camera to RAW+JPEG setting to record both at the same time you would notice the difference.

[kiwi]

IU saw the question more along the lines of what can be recovered in post processing, but I can see it that way also

[andylo]

kiwi, I think you are correct, I am trying to see how many information I can recover from a raw file, so I will know how flexible I can be when I am actually taking the picture. I have seen some camera can capture more than 8 stops +/- and the photo has been token was so dark it merely see a shadow! But after post process - EVERYTHING has recovered!! And I was bloody amazed! So I want to see what sort of leverage I can get from a normal SLR.

[kiwi]

You can generall recover far more from shadows than highlights, but at the expense of noise. It's not an exact science. You are often better shooting using bracketed exposures and then combining in post of you are after wide dynamic range

[etherial]

If the camera records straight black or straight white, then that is lost. Software can attempt to interpolate what it thinks it was after the fact but it is guessing based on the areas around close to the extremes. (ie the shades of white just before white, and the shades of dark grey just above black) It is therefore conceivable that as a RAW file has more detail between black and white then the software has a better chance of guessing correctly with more accuracy.

But this doesn't mean you are capturing a greater dynamic range in the file (either JPEG or RAW), it just means you might have a little more scope for correction. Bracket exposure where you need to and capture the detail with the camera.