16bit RGB with Image module

[Jason B]

Hi all,

I'm still new to all of this, but I'm trying to do something here that
*seems* like it should be pretty simple. All I want to do is take an array
of pixel data from a file (no header data or anything, just pixel data) in
RGB565 format and save it off to a bitmap file - or display it, in the case
below:

import sys, Image

if len(sys.argv) == 2:
print "\nReading: "+sys.argv[1]
image_file = open(sys.argv[1], "rb")
pixel_data = image_file.read()

im = Image.fromstring("RGB", (326, 325), pixel_data)
im.show()

When run, I get:

ValueError: not enough image data

Which I'm pretty sure is because it's expecting a 24bit image. Of course if
I tune down the width and height or change the format to B&W ("L") then it
*does* display an image, the B&W one even having recognizable features, just
not the right one.

I've read through the documentation a thousand times trying to understand
the raw decoder and plugins, etc. but I still can't figure this out...

Any help is greatly appreciated!

Thanks,
J

 

[Jason B]

Well I kept screwing around and funny thing, this works:

import sys, Image

if len(sys.argv) == 2:
print "\nReading: "+sys.argv[1]
image_file = open(sys.argv[1], "rb")
pixel_data = image_file.read()

im = Image.fromstring("RGB", (326, 325), pixel_data, "raw", "BGR;16")
im.show()

Although I have no idea *why* it works, other than the fact that I'm now
using the correct number of bits per pixel.

Anyone have thoughts on this?

Thanks!
J

 

[Will McGugan]

Well I kept screwing around and funny thing, this works:

import sys, Image

if len(sys.argv) == 2:
print "\nReading: "+sys.argv[1]
image_file = open(sys.argv[1], "rb")
pixel_data = image_file.read()

im = Image.fromstring("RGB", (326, 325), pixel_data, "raw", "BGR;16")
im.show()

Although I have no idea *why* it works, other than the fact that I'm now
using the correct number of bits per pixel.

Anyone have thoughts on this?

Well RGB will will use 24 bits per pixel, not 16, so the first error you
got makes sense. I guess using the raw decoder with "BGR;16" makes it
use 16 bit. Although, I couldn't find any reference in the docs!

I'm sure Fredrik Lundh could shed some light on this...

Will McGugan