I have been using the latest QImage for years.  I have been printing on an Epson 7900.  It died and I am now printing on an Epson P9000.  In the process of getting it calibrated I think I have found a flaw in my workflow that needs improvement for the best prints possible.  Here is my current workflow:
I calibrate my monitor with Colormunki
Then I develop icc profiles for each paper with Colormunki
I then put those two profiles in QImage Ultimate for printing (turning off profiles in Epson driver of course)

Now here is my image workflow:
Shoot images using Nikon D850 in raw with 14bit depth
Transfer raw files to hard drive and open/process in Lightroom
Once I get my "keepers" I export them from Lightroom using a low-res image with sRGB for web posting and export to jpeg with Adobe RGB full resolution JPEG
Print from those JPEGS as needed using QImage Ultimate.

Here is where I think I have a problem:  BY SAVING AS JPEG WITH ADOBE RGB AND PRINTING FROM THOSE JPEG FILES INSTEAD OF FROM LIGHTROOM DIRECTLY (using Qimage in Lightroom) have I potentially thrown out pixels because they were saved in Adobe RGB?  Because Lightroom is using Prophoto color space would all pixels printable by my P9000 be captured by changing my workflow?  Maybe this is cutting hairs/unnecessary? It's certainly easier to print directly from QImage rather than Lightroom.

This color management seems a bit complex and I want to do all I can to print the best prints possible. 

Thanks for any help/thoughts.

Greig

Brian,
Thanks for the input.  I have been trying to save/print in Prophoto and it is a mess trying to keep everything straight. 
Where I get sidetracked is when I plot aRGB in Perfx Gamut Viewer and then overlay my Colormunki printer profile, the yellows and dark blues are outside aRGB color space.  I think your point is that they are likely never seen in a photograph and are likely not captured in the first place.
I think I will keep my "old" workflow as you suggest!
Thanks again!

It's a complex subject but just to try to break down the basics...

In order for a color that is outside aRGB gamut to make it to print, a lot has to happen:

(1) Your subject's color has to be outside the aRGB gamut: let's say a super saturated yellow flower in bright sunlight.
(2) The lighting and exposure generally has to be such that the subject (yellow flower) is the brightest object in the frame.
(3) Your camera's sensor has to have enough gamut to capture that color in the first place (shooting in raw of course).
(4) Your raw developing software has to have internals (a camera ICC profile) that can support that out-of-aRGB color on developing.
(5) Your printer has to be able to reproduce that color given the ink, paper type, and printer profile you are using.

Now, for an image that makes it through all those checkpoints, technically ProPhoto would produce better color for those saturated yellows.  Whether or not you notice in print would depend on how much of the yellow flower is out of gamut, whether you use perceptual or RC intent, and other factors.  These are rare cases but that's not to say they don't happen.  I've never personally seen a case where I actually needed ProPhoto to reproduce bright colors and I tend to take a lot of flower shots with very saturated colors.  But part of that is because I let Qimage develop my raws.  If you develop your raws in Qimage there is no need to select a color space like aRGB or ProPhoto because Qimage has camera profiles for most of the popular cameras: it uses the camera profile rather than converting to a color space.  That means that color stays in the optimal color space (the color space of the camera) and in 16 bits/channel all the way to when the data is delivered to the driver for printing.  And even at print time, it does a 16 bit/channel dither so you get effective 16 bit/channel dithering even on 8 bit printer drivers.

If you develop your raws in other software and have to save them out to print in Qimage, I'd just use aRGB because it'll be sufficient for 99.9% of photos.  Then you could use ProPhoto on that .1% only if you notice a problem with bright colors in print.  Again, never happened for me but I suppose given the infinite variety of conditions going through steps 1-5 above, some could run into that more than others.  Some people may be working with specialized logo colors, for example, that may be outside the aRGB gamut in which case ProPhoto makes sense.  But for nearly all "in the wild" photographic content, ProPhoto is overkill.

Regards,
Mike

But you ARE throwing away important color information by using jpgs which are 8 bit rather than tiffs or RAW or even PSDs which support 16 bit files.  See https://www.youtube.com/watch?v=LqE8FBiDLwE, beginning about 59:58

I respect the fact that you know more than I do re bit depth. However I have seen issues, ie. banding, when editting anjpg vs z16 bit tiff image. This has convinced me to donall editting as 16 bit files.

Makes perfect sense

The graphic looked good and I thought he was going to explain it right but what he said was:

Perceptual: "Compresses colors that are outside to be inside, but keeps the colors that are inside basically the same."
RC: "Moves them all relative.  So the colors that are outside move in, but the colors that are in also get moved closer together."

The above verbal description is basically backwards.  If you swap his two verbal descriptions and remove the "Moves them all relative" statement, it'd be basically correct.  The point being, perceptual scales all colors, even those that are in-gamut which can have an overall dulling (desaturation) effect even on colors that could have been reproduced accurately (are in-gamut).  Unlike perceptual, relative colorimetric renders in-gamut colors correctly and doesn't move them and only moves out-of-gamut colors inward... generally to the gamut hull, that is, the closest color that CAN be reproduced.  This is why RC is often recommended: because many images won't contain out-of-gamut colors or at least few out-of-gamut colors and with RC, you know the colors that can be reproduced will be reproduced accurately rather than being "dulled" via scaling.

Mike

The two offer different tradeoffs when it comes to printing images that contain out-of-gamut colors.  RC is often considered a more "advanced" approach because when you do run into an image that has out-of-gamut colors in the main subject, you have to be able to recognize what is causing the loss of detail or other issues you might be seeing.  At that point, you can decide how to handle that situation and whether to use perceptual for that image, adjust the exposure, etc.  The reason perceptual is the default is because it produces consistent results and results in the fewest "what is causing this" complaints.

It's a bit like telling someone to "just shoot raw" without first learning to properly expose the scene, how to pick the proper aperture, what shutter speeds are acceptable for various focal lengths and lighting, etc.  So we start with perceptual and work our way up to the "more accurate" RC once we've gotten our printing workflow straightened out, profiles created, and so on.

Regards,
Mike