I believe the Colormunki Photo is the only one of the 3 you listed that can be used to profile a print. The others are good for emissive displays like monitors, but you need a spectrophotometer to profile printed media.
So one of the last steps in profiling a paper for B&W is to linearize its B&W response. That is, we want to make sure that when we print, for example, 30% gray onto the paper, it's 20 points lighter than 50% gray. That is, we need to make sure that the grey tones' brightnesses all have a linear relationship to each other. Another way to think about this is if you have a curve layer in Photoshop, we are making sure that the curve is a straight line.
To do this, you print a 21-patch gray scale with each patch representing 5% grey scale steps, and measure the brightness of each grey patch. QTR comes with a couple of scripting commands that take the results of such a measurement, and output a correction that is put in the LINEARIZE line. When I use Xrite's ColorPicker app, I measure the 21 patches, and output the results of the measurements as a CSV (Comma Separated Values) file. The QTR linearize app parses that file and outputs its correction curve. How I run the QTR linearize app on my measurements is by dragging the CSV file to the QTR linearize app.
The QTR app seems to be somewhat picky and people on the list have complained that when the profiles are installed, the corrections are sometimes rejected, which is where Richard's Excel spreadsheet comes in. I take the Lab L values produced by the QTR linearize app, and paste them into Richard's Excel spreadsheet, and it outputs linearization correction values that you paste into the GREY_CURVE field of the QIDF file.
So my linearization workflow is:
1. Measure with ColorMunki Photo and ColorPicker a 21-step grey scale with my profile of choice. I use Keith Cooper's standalone randomized wedge image that's tailored for the ColorMunki's scanning mode, but you can do it with the ColorMunki's spot mode on any step wedge, albeit much more slowly. I also use Print Tool with no color management for this step.
(When I determine ink crossover values, which is done before linearization, I use spot mode since the ink calibration charts are not amenable to the ColorMunki's scanning mode, which requires the tones to be in a certain order and of a certain shape. See ColorMunki black and white printing to get Keith's ColorMunki step wedge. He has it in a big test file as well as standalone as a skinny strip. I use the skinny strip so I can print multiple tests on one sheet of paper. I size it to 85% of actual size on a US letter sized page, and center it along the length of the paper so I avoid the microbanding areas, and the first one is placed 0.15 inches from the left edge. The next one is placed 2 inches, and the next 4 inches, and so on.)
2. Have ColorPicker output a .csv file of the measurements.
3. Drag the .csv file on top of the QTR linearize app.
4. Copy the output of the QTR linearize app, and paste it into a cell in Numbers. Numbers (and presumably Excel and OpenOffice) will put each of the columns and rows of the QTR output into its own cell.
5. Copy the Lab L column from the step 4 into the Lab L column of Richard's spreadsheet.
6. Copy the grey curve numbers from Richard's spreadsheet into a .qidf file. Copy that .qidf file into the QTR driver directory for your printer, and run the installation program.
Richard has a great explanation of the last 3 steps on his website if you want to use his spreadsheet. I've briefly compared the results of his spreadsheet to the QTR linearize app's results, and I find that his spreadsheet does a slightly better job and is more forgiving.