tzalman wrote in post #4200476
O.k., here is the requested demonstration:
Three photos of a step wedge shot at f/4 and 1/4, one at 200 ISO without HTP, the second at 200 with HTP and the third at 100 ISO without HTP, were opened in UFRaw. This Open Source free converter has the nice feature of showing a histogram of the RAW data - the only converter that I know of that does this since DPP's purported RAW histogram is in fact derived from processed data after picture style and white balance (and apparently HTP) have been applied. The first screen shot is of UFRaw. The upper histogram is input (RAW) and the lower one is RGB output. The first shot is fully exposed to the right. The second and third shots are identically one stop underexposed. Miriam, post 75, could have gotten the same results from LR merely by exposing one stop less.
Now lets see what DPP does. The RAW tab histogram is on the left, the output histogram is on the right. For the first shot the output is clipped (and so it appeared on the camera's LCD) although, as we have seen, it really isn't. Notice that for the second shot (the one with HTP) not only is the shape of the RAW tab histogram different, the margins have changed also. In terms of output the highlights of photo no. 2 fall between those of the other two while the shadows are like no. 1.
O.k., here is the requested demonstration:
Three photos of a step wedge shot at f/4 and 1/4, one at 200 ISO without HTP, the second at 200 with HTP and the third at 100 ISO without HTP, were opened in UFRaw. This Open Source free converter has the nice feature of showing a histogram of the RAW data - the only converter that I know of that does this since DPP's purported RAW histogram is in fact derived from processed data after picture style and white balance (and apparently HTP) have been applied. The first screen shot is of UFRaw. The upper histogram is input (RAW) and the lower one is RGB output. The first shot is fully exposed to the right. The second and third shots are identically one stop underexposed. Miriam, post 75, could have gotten the same results from LR merely by exposing one stop less.
Now lets see what DPP does. The RAW tab histogram is on the left, the output histogram is on the right. For the first shot the output is clipped (and so it appeared on the camera's LCD) although, as we have seen, it really isn't. Notice that for the second shot (the one with HTP) not only is the shape of the RAW tab histogram different, the margins have changed also. In terms of output the highlights of photo no. 2 fall between those of the other two while the shadows are like no. 1.
So it is your understanding that Canon is simply instructing the camera to automatically modify the in-camera exposure compensation (with HTP on) to enhance the amount of highlight detail, knowing that the dark areas are clipping to black one-stop later, then going back and overexposing the dark areas via a custom exposure compensation curve in DPP whilst leaving the highlights underexposed?
So basically the camera still has close to 1 stop more dynamic range (in the lower region), and Canon is removing all the thought-process steps and simply asking "Would you like to have detailed shadows or highlights?".
It sounds like a useful quick-fix to circumvent going through and doing it manually.
