In R2016a or earlier, in order to use == to compare two arrays, the two arrays must be exactly the same size, or else one of them must be a scalar. For example [1 2 3 4] == 3 is okay, and [1 2 3 4] == [5 6 7 8] is okay, but [1 2 3 4] == [5; 6; 7; 8] is not okay because they do not have the same size (1 x 4 compared to 4 x 1) and neither one is a scalar.
In R2016b or later, in order to use == to compare two arrays, the two arrays can be exactly the same size, or else one of can be a scalar, or in every dimension in which the sizes do not agree, one of the two must be a singular dimension. For example [1 2 3 4] == [5 6; 7 8] is [1 4] compared to [2 2], and the 1 vs 2 for the first dimension is potentially okay, but the 2 vs 4 for the second dimension is not okay. But [1 2 3 4] == [5; 6; 7; 8] is [1 4] compared to [4 1], and 1 vs 4 for the first dimension is okay and 4 x 1 for the second dimension is okay. The result of [1 2 3 4] == [5; 6; 7; 8] would end up being a 4 x 4 array of logical values -- 1 compared to each of the four values, 2 compared to each of the four values, and so on.
