Concatenate numpy array in new dimension

Question

I am trying to create a 2D matrix, that each datapoint is a 2D array. / 10 \ / |a1_0|a2_0|a3_0|....| |a1_1|..............| 1000 |a1_2|..............| |...................| \ |...................|

where a1_0, a1_1, a2_0 are all (1025, 16) array.

To my understanding, that means I have to create a matrix with shape = (10, 1000, 1025, 16). (please let me know if I'm wrong)

And all I have are a1_0, a1_1, a2_0...

So my question is how to concatenate these arrays with most efficient ways if I have to concatenate 2 numpy array into a new dimension?

I've tried np.vstack() and np.concatenate() but seem not right.

I am using a function to transfer a piece of audio into a spectrogram (which is a picture), the function will return a variable called spec, which is a (1025, 16) numpy array

Therefore, at first I was trying to do:

for a0 ~ a9, concatenate:
   for a1_1 ~ a1_999, concatenate:
       spec = gen_spectrogram (audio)

I am using a function to transfer a piece of audio into a spectrogram (which is a picture), the function will return a variable spec, which is a (1025, 16) numpy array — sealpuppy
– sealpuppy, Commented Nov 13, 2018 at 9:53
Initialize array out with shape : (10, 1000, 1025, 16) and then iteratively assign : out[i,j] = ... — Divakar
– Divakar, Commented Nov 13, 2018 at 9:56
Sorry I don't understand with the out thing, is it a numpy method? — sealpuppy
– sealpuppy, Commented Nov 13, 2018 at 10:10
@Divakar Are you suggesting that I can create an empty array with shape (10, 1000, 1025, 16) and then iteratively assign last 2 numbers? — sealpuppy
– sealpuppy, Commented Nov 13, 2018 at 10:14

randomwalker · Accepted Answer · 2018-11-13 10:36:13Z

2

First stack ai_0, ..., ai_999:

ci = np.stack([ai_0,...,ai_999])

Then stack c0, ..., c9

conc = np.stack([c0, ..., c9])

The result should be of shape (10, 1000, 1025, 16)

answered Nov 13, 2018 at 10:36

randomwalker

1,7031 gold badge12 silver badges14 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Concatenate numpy array in new dimension

1 Answer 1

Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Related