diff --git a/matlab/one_pfb.m b/matlab/one_pfb.m
index 80bd7dc07a9dd35d9d7674e6b0bc3785b7daa038..b24198eceb0e5d3b5df5413e7b1453bdadf411b2 100644
--- a/matlab/one_pfb.m
+++ b/matlab/one_pfb.m
@@ -113,8 +113,8 @@
 %        0 -1023   1023   2047  3071  4095  --> + --> -1023
 %              .      .      .     .     .
 %              .      .      .     .     .
-%             -2      2      .     .     .               -2
-%             -1      1      .     .     .               -1
+%             -2      2      .     .     .  --> + -->    -2
+%             -1      1      .     .     .  --> + -->    -1
 %     1023     0      0   1024  2048  3072  --> + -->     0
 %              d      h      h     h     h                d
 %
@@ -124,15 +124,30 @@
 %   and the older samples are multiplied by the subsequent coefficients,
 %   because it is a convolution. Therefore the FIR coefficients need to be
 %   flipped up/down per column, to allow doing the filter as a d .* h vector
-%   multiply in pfir.m. The FFT operates on blocks of data with same index and
-%   time range as the WG. The data output of the FIR filter fits this input
-%   range of the FFT. Therefore no data flipping is needed.
+%   multiply in pfir.m. The FIR filter sums the rows for the N_taps = 16 to
+%   yield the input for the FFT. The FFT operates on blocks of data with same
+%   index and time range as the WG. The data output of the FIR filter fits
+%   this input range of the FFT. Therefore no data flipping is needed.
+%
+%   In summary:
+%   * The WG data and FFT input data is processed in blocks with the newest
+%     sample at the bottom and the oldest sample at the top. In the FIR filter
+%     the newest block is at the left and the oldest at the rigth. Therefor
+%     the FIR filter coefficients also have to be ordered from bottom left to
+%     top right.
+%   * The column size is determined by N_fft of the FFT. The FFT is calculated
+%     each time a block is shifted in. For a critically sampled PFB the block
+%     size is N_blk = N_fft, so then it also looks like the blocks only shift
+%     from left to right. Using N_blk < N_fft would yield an oversampled PFB
+%     with oversampling factor R_os = N_fft/N_blk. Then the blocks shift from
+%     top to bottom in each column and from left to right.
 %
 %   In the tb_verify_pfb_response.vhd the input stimuli is a block of N_fft =
 %   1024 ones followed by N_taps-1 blocks with zeros. In time the oldest data
 %   will appear first in the simulator Wave Window, so therefore the
 %   fil_re_scope signal in the tb_verify_pfb_response.vhd will show the FIR
-%   coefficients h[] in order 1023:0, 2047:1024, ..., so flipped per block.
+%   coefficients h[0:16383] in order h[1023:0], h[2047:1024], ...,
+%   h[16383:15360], so flipped per block.
 %
 clear all;
 close all;