Why is the last row's frame size different from others when extracting LLD?

[Epholy]

Hello.
I'm trying to extract LLD from a period of sound, and here is my code:

import opensmile
smile = opensmile.Smile(
    feature_set=opensmile.FeatureSet.eGeMAPSv02,
    feature_level=opensmile.FeatureLevel.LowLevelDescriptors
)
signal = smile.process_file('./BASIC5000_0235.wav', start=1.518, end=1.596)

After running this code, I got 3×25 vector like this:

I wonder why there is only 3 rows, and why the last row starts from 1.538s but ends at 1.596s instead of 1.558s.