Create a PR and then we can see what this would look like.

XTC and TRR files do not store the dt but include each individual time (and we just calculate dt from the first two frames) so I am not sure that the same approach as for DCD will work.

You could try a custom on-the-fly transformation https://userguide.mdanalysis.org/stable/trajectories/transformations.html where you set the Timestep.time to frame * time: ts.time = ts.frame * dt. For example:

import MDAnalysis as mda; import MDAnalysisTests.datafiles as data

u = mda.Universe(data.TPR, data.XTC)

print("Original dt", u.trajectory.dt)
print("Original times", [ts.time for ts in u.trajectory])

gives

Original dt 100.00000762939453
Original times [0.0, 100.00000762939453, 200.00001525878906, 300.0, 400.0000305175781, 500.0000305175781, 600.0, 700.0000610351562, 800.0000610351562, 900.0000610351562]

Now we are adding the transformation

def set_time(dt):
    """Set time to frame * dt"""
    def wrapped(ts):
        ts.dt = dt
        ts.time = ts.frame * dt
        return ts
    return wrapped

u.trajectory.add_transformations(set_time(33.33))

print("New dt", u.trajectory.dt)
print("New times", [ts.time for ts in u.trajectory])

and we get

New dt 33.33
New times [0.0, 33.33, 66.66, 99.99, 133.32, 166.64999999999998, 199.98, 233.31, 266.64, 299.96999999999997]

(You can then analyze the trajectory or write it out to another XTC but with correct time stamps.)

I think that #4908 should implement what I suggest. In general it is quite similar to dcd writer and reader, though the logics is a bit different. I added tests, to cover new functionality and everything looks good to me

@pbuslaev sorry that it took a while to look at your PR. I just reviewed #4908.