Transform numeric vector to specific length n R

I have numeric vectors of different lengths and want to "standardize" their length to, say, 100 to make them comparable and transform the raw signal to a smoother version.

I do need control over the length and 'smoothing' parameters so I used the get_dct_transform function from the syuzhet package.

However, this is not super efficient with larger raw vectors:

set.seed(123)
sample_space = seq(-2, 2, .01)
a = sample(sample_space, 1000, replace = T)
b = sample(sample_space, 10000, replace = T)
c = sample(sample_space, 100000, replace = T)

system.time(syuzhet::get_dct_transform(raw_values = a, low_pass_size = 5))
#  user  system elapsed 
# 0.020   0.000   0.021 

system.time(syuzhet::get_dct_transform(raw_values = b, low_pass_size = 5))
#  user  system elapsed 
# 1.562   0.370   1.937 

system.time(syuzhet::get_dct_transform(raw_values = c, low_pass_size = 5))
#    user  system elapsed 
# 152.802  27.118 180.223 

Are there faster ways of doing these kinds of transformations (DCT or FT) in R?

Another SO answer suggested the fft function but I'd need direct control over the length and filter parameter. Same issue for the emuR::dct.

I'm not married to the DCT - other methods to transform a vector of length v to an 'smoothed' vector of length n would also be okay.