I've evaluated many paper manufacturers' generic profiles, and it's definitely NOT true (as a general rule) that they are better than profiles from source XYZ, just because in theory the manufacturers have access to better hardware. Many generic profiles are merely average; a surprising number are quite poor. Plus, generic profiles were produced who knows how long ago on an older batch of paper and a different printer and ink batch than what you're using. If you have a tool to profile your own batches of paper and ink on your own printer, then my Type A personality says do that. At least it lets you control some variables... you know the materials that are involved in the process, and you know what level of profiling regimen has been used.
I do agree with Andrew that the ColorMunki Photo produces surprisingly good results considering the easy to use "wizard" type of software with limited control, and the small number of patches used. When I evaluate Munki-produced profiles I can see some side effects from the small patch count; they vary depending on the paper in question, and generally I find the effects aren't that significant. While more advanced equipment could do a better job in the hands of somebody who knows what they're doing, I've also find numerous instances where a Munki paper profile is much preferable to the paper manufacturer's own generic profile. So far I've found only a few instances where prints were visually better using the manufacturer profiles than Munki profiles. These cases usually involve media that really deviates from the norm, such as Moab Slickrock.
Also remember that the rendering intent you select can matter. I've seen some manufacturer profiles that look okay under Relative Colorimetric. But there are certain images for which Perceptual would be more appropriate, and I've seen the same generic profile fall apart using Perceptual instead of RC.
As always, the proof of the pudding is in the eating. If you're curious which is better for a certain paper, make some test prints with the manufacturer's generic profile. Then create a custom Munki profile and make the exact same test prints using it. Look at both sets under a variety of your lighting changes, and see which you like better. Profiling is mostly science, but appreciation of a print mostly is not, though for me appreciation can be informed by an understanding of what's going on under the hood.
As Andrew also said, how you calibrate your monitor is a big part of the monitor-to-print matching equation. Plus you're saying that your lighting changes a lot; clearly a print can only look as good as the ambient let will allow it to look, and very different ambient lighting hitting your workstation area & monitor may influence you differently as you develop images. You need to have some consistency and quality in your working and viewing light if you really want to evaluate print consistency and quality.
I always calibrate the monitor a specific value (90 cd/m2 in my case, using a NEC Spectraview monitor). I'm not a fan of having monitor brightness altered on the fly in response to changes in ambient lighting conditions. I'm too Type A for that; I much prefer to control the ambient lighting in my workstation area, because once again consistency is important when trying to critically evaluate things from one image or print to the next. The time I bring consideration of ambient lighting into the equation is when I develop a master print file and I know something about the eventual viewing conditions of that specific print...
