Okay, these are not bad results. I would do nearly the same.
But for example, if you play a bit with non-matching centers, you could get for example:
I guess I'll just quickly do some then.
U2 M U M U M2 U2 M U' M U2 M' U M2 U' M2 U' M' U M' U2 M U M2 U
U M' U2 M' U2 M U M' U' M' U2 M' U' M2 U' (15)
M U M' U' M U2 M' U2 M' U M' (11)
M U - misoriented centers, but now UB-DB is oriented
M' U' M (solve EO - note that I put UR into its place on purpose (UB)
U2 M' U2 M' U - solve UL-UR
M' - mperm
U M2 U M2 U2 M' U' M2 U M U M' U2 M' U2 M2 U2 M2 U2 M' U M2 U2 M' U'
M U M' U2 M' U2 M U M2 U M2 U2 (12)
That's pretty good.
M U2 M' U2 M U' M U2 M2 U M U M U M2 U M2 U M' U M' U' M U2 M2
U2 M' U2 M U M' U' M' U2 M U' M2 U M' U2 M (16)
Couldn't find better. Alternative:
M2 U2 M' U2 M' U' M' U M U2 M' U' M' U2 M' U (16)
M2 U2 M' U2 M' U' M' - EO, but instead of UL-UR, we will solve UF-UB (treated as UL-UR) in the next step
U M U2 M' U' - solve UF-UB
M' U2 M' U - solve the remaining 4 edges, AUF
Alternative2:
U2 M' U2 M U M' U' M' U2 M U' M2 U M' U2 M (16)
U' M2 U M U2 M2 U2 M2 U2 M2 U M2 U' M' U' M U M2 U M2 U' M' U2 M2 U'
M2 U' M U M U' M' U2 M U M2 U' M2 (13)
Good.
U2 M2 U M U M' U2 M U2 M' U2 M2 U M2 U M U' M' U2 M' U2 M' U' M U'
M U2 M' U M U M' U M2 U2 M U' M2 U M2 (15)
M' U2 M U M' U M U M2 U2 M (11)
with mis-oriented centers:
M' U2 M - rearrange the EO in a way that 4b will be very easy. Used this table:
http://www.speedsolving.com/forum/showthread.php?38176-Roux-Misoriented-Centers-Table k10 after a U' setup (U' + UM'U2M) Maybe it is worth to learn the easy cases, such as column j,k,l,m.
U M' U M U' - solve UL/UR
U2 M2 U2 M - mperm
I think your move count is pretty okay if you use the standard method. You can improve your times with smoother transition, lookahead, fingertricks, increasing TPS. If you want to decrease your move count, learn mis-oriented centers, and tricks for 4a+4b (like the table above).
Your sub-5 with LSE and ~10 CMLL+LSE shows that your CMLL seems a bit slow. Are those 2-look CMLLs?