Bug in converting matlab code to python

Question

I have

mask = 255 ('000000FF')
r = uint8(bitand(bitshift(rgb, -16), mask))

This code is written in matlab. Now is this code correct in Python:

r = (rgb>>16 & mask).astype('uint8')

For rgb = 1884829820, I am getting 88 as the answer for matlab while I am getting 256 for Python. I don't understand why?

Note that SO isn't a free code translation service so you should post your own attempt at coding this in Python, and explain the problem you're having. Are you doing this in plain Python or in Numpy? — PM 2Ring
– PM 2Ring, Commented Apr 30, 2017 at 10:16
Search for bit operations in python and and an attempt first. If you have further problems it's of course fine to make a question. — Grimmy
– Grimmy, Commented Apr 30, 2017 at 10:18
Thanks. Your update should include a small sample array so we can see the exact format & datatype of your data. BTW, are you sure mask is 256? A mask of 255 (0xff) would make more sense. — PM 2Ring
– PM 2Ring, Commented Apr 30, 2017 at 10:26
@talos1904 In the update, try to find a more descriptive title for your question as well :) — Grimmy
– Grimmy, Commented Apr 30, 2017 at 10:28

Grimmy · Accepted Answer · 2017-04-30 11:28:29Z

2

mask = 0xFF # = 255 also works
rgb = 1884829820
r = rgb >> 16 & mask
print(r)

Outputs:

0xFF is 255, not 256. That is your problem.

256 will create the mask 0b100000000

answered Apr 30, 2017 at 11:28

Grimmy

4,17725 silver badges25 bronze badges

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

1 Comment

talos1904 Over a year ago

Yes! I understood the code as well as the problem. Thank you! :) It's time to savor.

PM 2Ring · Accepted Answer · 2017-04-30 11:37:31Z

I assume this code is supposed to extract the red value from a RGB color value stored as a 24 bit unsigned integer, with red in te high bits and blue in the low bits. So mask needs to be 255 (0xff in hexadecimal) to capture the 8 bits of data from the red channel after we shift them into place.

Here's a small Numpy demo.

import numpy as np

np.set_printoptions(formatter={'int': hex})

colors = np.array([0x123456, 0x789abc, 0xdef012, 1884829820], dtype=np.uint32)
print(colors, colors.dtype)

mask = 255
reds = ((colors >>16) & mask).astype(np.uint8)
print(reds, reds.dtype)

output

[0x123456 0x789abc 0xdef012 0x7058387c] uint32
[0x12 0x78 0xde 0x58] uint8

I've used hex output to make it clear that we're actually getting the red channel bits. Note that 0x58 == 88 in decimal.

Collectives™ on Stack Overflow

Bug in converting matlab code to python

2 Answers 2

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related