Unfortunately it's more than two decades since I used CorelDraw. So all I can do is tell you how it works in InDesign, and hope it translates ok.
In InDesign we set up our document in the CMYK profile that the print shop advises for the job (make sure they DO advise one, don't just guess). It's vitally important that the document is CMYK, not RGB, so that blacks are accurate. We do our design - text and linework and graphs and whatever else - in that CMYK document.
For photos, we edit them in Photoshop and leave them in RGB mode, but we soft-proof them as you mentioned above, and make any adjustments that are necessary. We DON'T convert them to CMYK.
We place the RGB photos into the CMYK document.
Then, as you said, the PDF conversion makes the photos CMYK during the output process.
This is one of the reasons we do it as described above. If your final PDF output is an RGB e-book, the photos stay exactly as they are, and the other items get converted to RGB.
This means that no matter what you're outputting, none of the elements of your design get converted more than once. Multiple colour mode conversions are very dangerous. There is no "round trip" in colour conversion. If a photo is in RGB, then gets converted to CMYK then back to RGB again, the values don't end up back where they were. They'll be different. This must be avoided.
Gosh I hope this makes sense.