Included below are 3 screenshots of exactly the same image that was developed using DPP4. The reason for screenshots is to include the view seen while viewing the (same) image in different software products used for developing raw files.
First is the image in DPP4. Notice that the clipping indicator is turned on but it might be hard to find any clipping in the highlights.
Next is the same image file when viewed in Rawtherapee. Also with the clipping indicator turned on. Notice that most of the sky is clipped.
Next is very same image file when viewed in Darktable. Clipping appears quite similar to that shown in Rawtherapee.
The question might be which view contains valid clipping indication. Might it make sense to vote for the case that is repeated (i.e., 2 out of 3 in this case). Assuming DPP4 is wrong, what good is the clipping indicator?
In DPP you can change the threshold for where you want clipping to be indicated. The default values are 0 and 255, so they will not indicate clipping unless you change them. Other software products may have a setting baked in and unadjustable.
Yes I know but I have them set the same. Highlights are set to 253 in all cases.
Something I think could be having affect is the difference between whether all 3 channels need to be out of gamut or just one. I believe with Darktable and Rawtherapee it is only one. I don't know when it comes to DPP4.
After a little more research I've come to realize that another relevant factor is that, at least in the documentation for other post-processing software, clipping pertains to the output ICC profile to be used for a resulting image file. Of course, output profile cannot be controlled in DPP4 and in fact NO such profile can be found among the files that make up the software. Reference this topic for some elaboration if interested.
However, the files developed by DPP4 can/do contain an ICC profile [named "sRBGv1.31(Canon)"]. Therefore, I was able to extract that profile from the DPP4 developed image file and specify that as the Output ICC Profile to be used in the other software. When that is done little or NO change to the above displayed results appears. However, said software supplies a variety of Output Profiles that can be used. When I choose from among the wider gamut supplied profiles there is significant reduction in the amount of clipping. However, NOT so much as to render what remains insignificant.
I suppose there may be folks who develop raw files and don't really care what they look like on devices other than the one used to develop them. Or maybe they only care about other display devices, in which case the differences may be considered insignificant. However, when wanting to develop images suitable for printing on various kinds of paper/material it does matter. This is especially true when those media are expensive enough that printing test cases is very undesirable. In that, we want to get results that are as reliable as possible via soft-proofing.
It is pretty hard to trust the DPP4 soft proof results when such dramatic discrepancies in clipping exist.