There seems to be very well defined grain in the color crop.
What I see in the color scan is a lot of digital noise. This makes sense since C41 negatives are low gamma so there will be a lot of analog and/or digital gain being applied to yield a normal (digital) contrast range. This translates into noise. This isn't helped of course by the fact that all three color channels are needed to create three distinct images, and there's no possibility for bunching the channels together into a single signal to cut down noise, which with B&W is possible. You can tell it's digital noise because there's no physical geometry about it except for distinct pixel-to-pixel variation.
As to your B&W scans: like
@_T_ says, it's a little hard to tell what exactly we're looking at, and for instance the HP5+ scan is not labeled as either 5kdpi or 10kdpi. In any case, there are a couple of things you need to keep in mind when scanning at very high resolutions and the limitations of the output:
* The odds that your scanner resolves its claimed 10,000dpi are next to zero. To the best of my knowledge, your scanner is also sold under the Reflecta brand, and allegedly, this particular scanner can resolve around 4300dpi:
https://www.filmscanner.info/en/ReflectaRPS10M.html This is really, really good!
* 100% crops of film scans pretty much always end up looking insanely 'mushy' (apart from possible digital noise as mentioned above). Don't expect something that looks like the output of a digital camera. Focus problems can be part of this, as are inherent limitations to the resolving power of the lens used and all manner of optical deficiencies, and yes, focus problems. But even with exact focus, you end up with an unsharp image viewed at 100%.
* Scanning grain is a pretty complex phenomenon - grain itself actually is. There's a really nice article that's already a few years old, but that illustrates quite nicely what the difference between 'grain' (as we or scanners see it) and actual silver particles, what kind of sizes/magnitudes we're talking about, how granularity is measured, what kind of real-world resolutions this translates into etc., and most importantly, what grain means when scanning film:
https://vashivisuals.com/wp-content/uploads/2017/07/2007-04-vitale-filmgrain_resolution.pdf
One of the key takeaways of the Vitale article above is that what we know as grain, is in fact a kind of interference pattern that emerges if you look through a 3-dimensional emulsion with dye clouds or silver particles suspended in it, which appear to partly overlap when viewed perpendicular to the emulsion (and then everything changes if the viewing direction changes, and/or the backlight angle changes!). The pattern will be different if the viewing angle changes - and the interference pattern will furthermore interact with the resolving power of an observing apparatus, like a scanner. To make matters even worse, the dye clouds or silver particles turn out not to be discrete 'binary' artefacts that have a distinct density, but rather complex geometries with varying optical density if you further enlarge them (far beyond the resolution of any scanner). This all boils down to that there's only a weak relationship between physical silver or dye clusters and the pixel pattern that we experience as grain. Given the several factors that interact, it can be challenging to interpret a digital scan in terms of grain. Simply put; most of the time it's just not very clear what you're looking at in the first place.
Another issue is of course that of resolution. Your scanner will resolve in the real world something like 5um 'dots' in a best case scenario. The 'grain' in film is on the scale of around that mark (5um) up to maybe 50um if you take a large bandwidth. So in a typical case, the grain will be somewhere around the 'size' of 1 or 2 pixels of your scanner. This means that your scanner, including its tendency to add noise to the signal, its optical shortcomings, the inherent problems with visualizing grain etc. etc. will simply never be able to image 'distinct' grain. A lower-resolution scanner may in some cases even
appear to be doing a better job if there happens to be a crisp interference pattern between the scanner and the compound effect of silver particles (partly) in the optical path.
And I haven't even touched on the issues of film flatness, limitations of optical/camera systems (alright, not so relevant when just looking at grain), film processing (which does influence granularity in several ways)...
So this is a very long way of saying that for many reasons, it's pretty much impossible to tell what in the heck we're looking at
If you want to figure out how well your particular scanner works, I see no other solution than to do the tedious work of scanning resolution targets and then interpreting those. But in the end, I doubt that doing so will make you a better photographer, and in itself, it will also not make your scans any better. Keep in mind you're using a consumer-grade piece of equipment of a couple of hundred eurodollarpounds to image down to a 5um resolution. If you think about it, it's a bit of a miracle this sort of works in the first place
