I've attached a patched JpegImagePlugin.py file (sorry for the 14.5kb
attachment to a mailing list). It should fix the inversion problem both
upon load and upon save of CMYK JPEG files.
Modifications can be found starting on the following lines:
line 35: add ImageChops to the import list, for inverting images
line 273: comment out previous patch for Adobe Photoshop 2.5 CMYK files...
I sure hope nobody uses 2.5 anymore, and this patch breaks newer versions.
line 288: invert CMYK images after load
line 416: invert CMYK images before save
I'm sure there are more elegant an efficient ways of fixing this problem,
but this works. I hope it helps you, as I've had the same issues myself.
Fredrik - if you're ok with the crude methods of fixing this problem, can
you include it in future builds? Thanks.
Kevin.
----- Original Message -----
From: <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>; "Cesare Leonardi" <[EMAIL PROTECTED]>;
<image-sig@python.org>
Sent: Saturday, April 15, 2006 2:01 PM
Subject: Re: [Image-SIG] Problem with JPEG and CMYK color space
In the JPEGImagePlugin.py module, there's a hack to handle Photoshop 2.5
CMYK JPEG images (around line 273)... that's causing half the problem upon
loading JPEG's. By commenting that out, you get a proper image, although
it's still inverted.
I can't seem to find the right place to re-invert the image upon load, but
I'll see what I can find. Anyone with suggestions would be welcome.
There's probably a better way to do this though, so the data is read
correctly in the first place (in CMYK, 0,0,0,0 = white, 255,255,255,255 =
full black... this is probably the problem).
Kevin.
----- Original Message -----
From: <[EMAIL PROTECTED]>
To: "Cesare Leonardi" <[EMAIL PROTECTED]>; <image-sig@python.org>
Sent: Saturday, April 15, 2006 1:32 PM
Subject: Re: [Image-SIG] Problem with JPEG and CMYK color space
Actually, the problem is that PIL also has problems reading CMYK JPEG
files
apparently. I just proved that by loading your CMYK test files and doing
a
"show", then loading a CMYK TIFF file and doing a show (note that to use
show, it converts to RGB with the internal conversion, so the colors
won't
be perfect). The TIFF loaded fine, the JPEG didn't.
The patch I've submitted fixes the problem with saving incorrectly, but
doesn't fix the reading issue (yet). So, as long as your source CMYK
image
doesn't come from a JPEG file it's ok (with my invert patch).
I'll see if I can spend some time looking at that this weekend.
Kevin.
----- Original Message -----
From: "Cesare Leonardi" <[EMAIL PROTECTED]>
To: <image-sig@python.org>
Sent: Friday, April 14, 2006 9:38 AM
Subject: Re: [Image-SIG] Problem with JPEG and CMYK color space
[EMAIL PROTECTED] ha scritto:
When saving a CMYK file to JPG, PIL seems to invert the colors. I've
submitted a patch to Fredrik already for this, and it should be in the
next major build.
Adding a simple invert to the image before saving will help.
Thanks all for your responses.
Kevin, i've made some tests following what you've suggested and here is
the results.
To obtain the color inversion i've used the function invert() in
ImageChops module. I've added the code and the new sample images in this
page (see "part 2"):
http://www.bernispa.com/pil/index.html
As you can see your suggestion works partially.
Images 02 and 03 (that was particular in "part 1") are the only that
after the inversion looks ok.
All the other images looks less dark, but the color are always different
than the original.
The particular thing to note is that the images produced by the
inversion looks the same as Image.show(). Seems that show() already do
the inversion.
There is another thing that i have noted only now: the images saved by
PIL are often less big than the original. Sometimes much less. For
example:
test01: 1435 KB test01-dst: 198 KB test01-dst-invert: 198 KB
test04: 1237 KB test01-dst: 405 KB test01-dst-invert: 405 KB
test08: 1647 KB test01-dst: 777 KB test01-dst-invert: 776 KB
test09: 1715 KB test01-dst: 665 KB test01-dst-invert: 665 KB
test10: 2032 KB test01-dst: 659 KB test01-dst-invert: 659 KB
I can think that PIL optimize the compression, but that difference are
really big. Isn't it suspicious? I expect that saving an image in a new
file without modifications produces a file very similar to the original,
isn't it?
The original images was not produced by me so the author can have used
low compression. I cannot make tests since i haven't a program that can
save in CMYK (Gimp seems not able).
Regards.
Cesare.
_______________________________________________
Image-SIG maillist - Image-SIG@python.org
http://mail.python.org/mailman/listinfo/image-sig
_______________________________________________
Image-SIG maillist - Image-SIG@python.org
http://mail.python.org/mailman/listinfo/image-sig
#
# The Python Imaging Library.
# $Id: JpegImagePlugin.py 2199 2004-12-18 08:49:05Z fredrik $
#
# JPEG (JFIF) file handling
#
# See "Digital Compression and Coding of Continous-Tone Still Images,
# Part 1, Requirements and Guidelines" (CCITT T.81 / ISO 10918-1)
#
# History:
# 1995-09-09 fl Created
# 1995-09-13 fl Added full parser
# 1996-03-25 fl Added hack to use the IJG command line utilities
# 1996-05-05 fl Workaround Photoshop 2.5 CMYK polarity bug
# 1996-05-28 fl Added draft support, JFIF version (0.1)
# 1996-12-30 fl Added encoder options, added progression property (0.2)
# 1997-08-27 fl Save mode 1 images as BW (0.3)
# 1998-07-12 fl Added YCbCr to draft and save methods (0.4)
# 1998-10-19 fl Don't hang on files using 16-bit DQT's (0.4.1)
# 2001-04-16 fl Extract DPI settings from JFIF files (0.4.2)
# 2002-07-01 fl Skip pad bytes before markers; identify Exif files (0.4.3)
# 2003-04-25 fl Added experimental EXIF decoder (0.5)
# 2003-06-06 fl Added experimental EXIF GPSinfo decoder
# 2003-09-13 fl Extract COM markers
#
# Copyright (c) 1997-2003 by Secret Labs AB.
# Copyright (c) 1995-1996 by Fredrik Lundh.
#
# See the README file for information on usage and redistribution.
#
__version__ = "0.5"
import array, string
import Image, ImageFile, ImageChops
def i16(c,o=0):
return ord(c[o+1]) + (ord(c[o])<<8)
def i32(c,o=0):
return ord(c[o+3]) + (ord(c[o+2])<<8) + (ord(c[o+1])<<16) + (ord(c[o])<<24)
#
# Parser
def Skip(self, marker):
n = i16(self.fp.read(2))-2
ImageFile._safe_read(self.fp, n)
def APP(self, marker):
#
# Application marker. Store these in the APP dictionary.
# Also look for well-known application markers.
n = i16(self.fp.read(2))-2
s = ImageFile._safe_read(self.fp, n)
app = "APP%d" % (marker&15)
self.app[app] = s # compatibility
self.applist.append((app, s))
if marker == 0xFFE0 and s[:4] == "JFIF":
# extract JFIF information
self.info["jfif"] = version = i16(s, 5) # version
self.info["jfif_version"] = divmod(version, 256)
# extract JFIF properties
try:
jfif_unit = ord(s[7])
jfif_density = i16(s, 8), i16(s, 10)
except:
pass
else:
if jfif_unit == 1:
self.info["dpi"] = jfif_density
self.info["jfif_unit"] = jfif_unit
self.info["jfif_density"] = jfif_density
elif marker == 0xFFE1 and s[:5] == "Exif\0":
# extract Exif information (incomplete)
self.info["exif"] = s # FIXME: value will change
elif marker == 0xFFE2 and s[:5] == "FPXR\0":
# extract FlashPix information (incomplete)
self.info["flashpix"] = s # FIXME: value will change
elif marker == 0xFFEE and s[:5] == "Adobe":
self.info["adobe"] = i16(s, 5)
# extract Adobe custom properties
try:
adobe_transform = ord(s[1])
except:
pass
else:
self.info["adobe_transform"] = adobe_transform
def COM(self, marker):
#
# Comment marker. Store these in the APP dictionary.
n = i16(self.fp.read(2))-2
s = ImageFile._safe_read(self.fp, n)
self.app["COM"] = s # compatibility
self.applist.append(("COM", s))
def SOF(self, marker):
#
# Start of frame marker. Defines the size and mode of the
# image. JPEG is colour blind, so we use some simple
# heuristics to map the number of layers to an appropriate
# mode. Note that this could be made a bit brighter, by
# looking for JFIF and Adobe APP markers.
n = i16(self.fp.read(2))-2
s = ImageFile._safe_read(self.fp, n)
self.size = i16(s[3:]), i16(s[1:])
self.bits = ord(s[0])
if self.bits != 8:
raise SyntaxError("cannot handle %d-bit layers" % self.bits)
self.layers = ord(s[5])
if self.layers == 1:
self.mode = "L"
elif self.layers == 3:
self.mode = "RGB"
elif self.layers == 4:
self.mode = "CMYK"
else:
raise SyntaxError("cannot handle %d-layer images" % self.layers)
if marker in [0xFFC2, 0xFFC6, 0xFFCA, 0xFFCE]:
self.info["progression"] = 1
for i in range(6, len(s), 3):
t = s[i:i+3]
# 4-tuples: id, vsamp, hsamp, qtable
self.layer.append((t[0], ord(t[1])/16, ord(t[1])&15, ord(t[2])))
def DQT(self, marker):
#
# Define quantization table. Support baseline 8-bit tables
# only. Note that there might be more than one table in
# each marker.
# FIXME: The quantization tables can be used to estimate the
# compression quality.
n = i16(self.fp.read(2))-2
s = ImageFile._safe_read(self.fp, n)
while len(s):
if len(s) < 65:
raise SyntaxError("bad quantization table marker")
v = ord(s[0])
if v/16 == 0:
self.quantization[v&15] = array.array("b", s[1:65])
s = s[65:]
else:
return # FIXME: add code to read 16-bit tables!
# raise SyntaxError, "bad quantization table element size"
#
# JPEG marker table
MARKER = {
0xFFC0: ("SOF0", "Baseline DCT", SOF),
0xFFC1: ("SOF1", "Extended Sequential DCT", SOF),
0xFFC2: ("SOF2", "Progressive DCT", SOF),
0xFFC3: ("SOF3", "Spatial lossless", SOF),
0xFFC4: ("DHT", "Define Huffman table", Skip),
0xFFC5: ("SOF5", "Differential sequential DCT", SOF),
0xFFC6: ("SOF6", "Differential progressive DCT", SOF),
0xFFC7: ("SOF7", "Differential spatial", SOF),
0xFFC8: ("JPG", "Extension", None),
0xFFC9: ("SOF9", "Extended sequential DCT (AC)", SOF),
0xFFCA: ("SOF10", "Progressive DCT (AC)", SOF),
0xFFCB: ("SOF11", "Spatial lossless DCT (AC)", SOF),
0xFFCC: ("DAC", "Define arithmetic coding conditioning", Skip),
0xFFCD: ("SOF13", "Differential sequential DCT (AC)", SOF),
0xFFCE: ("SOF14", "Differential progressive DCT (AC)", SOF),
0xFFCF: ("SOF15", "Differential spatial (AC)", SOF),
0xFFD0: ("RST0", "Restart 0", None),
0xFFD1: ("RST1", "Restart 1", None),
0xFFD2: ("RST2", "Restart 2", None),
0xFFD3: ("RST3", "Restart 3", None),
0xFFD4: ("RST4", "Restart 4", None),
0xFFD5: ("RST5", "Restart 5", None),
0xFFD6: ("RST6", "Restart 6", None),
0xFFD7: ("RST7", "Restart 7", None),
0xFFD8: ("SOI", "Start of image", None),
0xFFD9: ("EOI", "End of image", None),
0xFFDA: ("SOS", "Start of scan", Skip),
0xFFDB: ("DQT", "Define quantization table", DQT),
0xFFDC: ("DNL", "Define number of lines", Skip),
0xFFDD: ("DRI", "Define restart interval", Skip),
0xFFDE: ("DHP", "Define hierarchical progression", SOF),
0xFFDF: ("EXP", "Expand reference component", Skip),
0xFFE0: ("APP0", "Application segment 0", APP),
0xFFE1: ("APP1", "Application segment 1", APP),
0xFFE2: ("APP2", "Application segment 2", APP),
0xFFE3: ("APP3", "Application segment 3", APP),
0xFFE4: ("APP4", "Application segment 4", APP),
0xFFE5: ("APP5", "Application segment 5", APP),
0xFFE6: ("APP6", "Application segment 6", APP),
0xFFE7: ("APP7", "Application segment 7", APP),
0xFFE8: ("APP8", "Application segment 8", APP),
0xFFE9: ("APP9", "Application segment 9", APP),
0xFFEA: ("APP10", "Application segment 10", APP),
0xFFEB: ("APP11", "Application segment 11", APP),
0xFFEC: ("APP12", "Application segment 12", APP),
0xFFED: ("APP13", "Application segment 13", APP),
0xFFEE: ("APP14", "Application segment 14", APP),
0xFFEF: ("APP15", "Application segment 15", APP),
0xFFF0: ("JPG0", "Extension 0", None),
0xFFF1: ("JPG1", "Extension 1", None),
0xFFF2: ("JPG2", "Extension 2", None),
0xFFF3: ("JPG3", "Extension 3", None),
0xFFF4: ("JPG4", "Extension 4", None),
0xFFF5: ("JPG5", "Extension 5", None),
0xFFF6: ("JPG6", "Extension 6", None),
0xFFF7: ("JPG7", "Extension 7", None),
0xFFF8: ("JPG8", "Extension 8", None),
0xFFF9: ("JPG9", "Extension 9", None),
0xFFFA: ("JPG10", "Extension 10", None),
0xFFFB: ("JPG11", "Extension 11", None),
0xFFFC: ("JPG12", "Extension 12", None),
0xFFFD: ("JPG13", "Extension 13", None),
0xFFFE: ("COM", "Comment", COM)
}
def _accept(prefix):
return prefix[0] == "\377"
##
# Image plugin for JPEG and JFIF images.
class JpegImageFile(ImageFile.ImageFile):
format = "JPEG"
format_description = "JPEG (ISO 10918)"
def _open(self):
s = self.fp.read(1)
if ord(s[0]) != 255:
raise SyntaxError("not a JPEG file")
# Create attributes
self.bits = self.layers = 0
# JPEG specifics (internal)
self.layer = []
self.huffman_dc = {}
self.huffman_ac = {}
self.quantization = {}
self.app = {} # compatibility
self.applist = []
while 1:
s = s + self.fp.read(1)
i = i16(s)
if MARKER.has_key(i):
name, description, handler = MARKER[i]
# print hex(i), name, description
if handler is not None:
handler(self, i)
if i == 0xFFDA: # start of scan
rawmode = self.mode
# patch by Kevin Cazabon to comment this out - nobody should
be using Photoshop 2.5 any more (and it breaks newer versions)
# CMYK images are still inverted, we'll fix that just before
returning.
#if self.mode == "CMYK" and self.info.has_key("adobe"):
# rawmode = "CMYK;I" # Photoshop 2.5 is broken!
self.tile = [("jpeg", (0,0) + self.size, 0, (rawmode, ""))]
# self.__offset = self.fp.tell()
break
s = self.fp.read(1)
elif i == 0 or i == 65535:
# padded marker or junk; move on
s = "\xff"
else:
raise SyntaxError("no marker found")
# patch by Kevin Cazabon to re-invert CMYK JPEG files
if self.mode == "CMYK":
self.im = ImageChops.invert(self).im
def draft(self, mode, size):
if len(self.tile) != 1:
return
d, e, o, a = self.tile[0]
scale = 0
if a[0] == "RGB" and mode in ["L", "YCbCr"]:
self.mode = mode
a = mode, ""
if size:
scale = max(self.size[0] / size[0], self.size[1] / size[1])
for s in [8, 4, 2, 1]:
if scale >= s:
break
e = e[0], e[1], (e[2]-e[0]+s-1)/s+e[0], (e[3]-e[1]+s-1)/s+e[1]
self.size = ((self.size[0]+s-1)/s, (self.size[1]+s-1)/s)
scale = s
self.tile = [(d, e, o, a)]
self.decoderconfig = (scale, 1)
return self
def load_djpeg(self):
# ALTERNATIVE: handle JPEGs via the IJG command line utilities
import tempfile, os
file = tempfile.mktemp()
os.system("djpeg %s >%s" % (self.filename, file))
try:
self.im = Image.core.open_ppm(file)
finally:
try: os.unlink(file)
except: pass
self.mode = self.im.mode
self.size = self.im.size
self.tile = []
def _getexif(self):
# Extract EXIF information. This method is highly experimental,
# and is likely to be replaced with something better in a future
# version.
import TiffImagePlugin, StringIO
def fixup(value):
if len(value) == 1:
return value[0]
return value
# The EXIF record consists of a TIFF file embedded in a JPEG
# application marker (!).
try:
data = self.info["exif"]
except KeyError:
return None
file = StringIO.StringIO(data[6:])
head = file.read(8)
exif = {}
# process dictionary
info = TiffImagePlugin.ImageFileDirectory(head)
info.load(file)
for key, value in info.items():
exif[key] = fixup(value)
# get exif extension
file.seek(exif[0x8769])
info = TiffImagePlugin.ImageFileDirectory(head)
info.load(file)
for key, value in info.items():
exif[key] = fixup(value)
# get gpsinfo extension
try:
file.seek(exif[0x8825])
except KeyError:
pass
else:
info = TiffImagePlugin.ImageFileDirectory(head)
info.load(file)
exif[0x8825] = gps = {}
for key, value in info.items():
gps[key] = fixup(value)
return exif
# --------------------------------------------------------------------
# stuff to save JPEG files
RAWMODE = {
"1": "L",
"L": "L",
"RGB": "RGB",
"RGBA": "RGB",
"RGBX": "RGB",
"CMYK": "CMYK",
"YCbCr": "YCbCr",
}
def _save(im, fp, filename):
try:
rawmode = RAWMODE[im.mode]
except KeyError:
raise IOError("cannot write mode %s as JPEG" % im.mode)
info = im.encoderinfo
dpi = info.get("dpi", (0, 0))
# get keyword arguments
im.encoderconfig = (
info.get("quality", 0),
# "progressive" is the official name, but older documentation
# says "progression"
# FIXME: issue a warning if the wrong form is used (post-1.1.5)
info.has_key("progressive") or info.has_key("progression"),
info.get("smooth", 0),
info.has_key("optimize"),
info.get("streamtype", 0),
dpi[0], dpi[1]
)
if im.mode == "CMYK":
# invert it so it's handled correctly in Photoshop/etc. - Kevin Cazabon.
im = ImageChops.invert(im)
ImageFile._save(im, fp, [("jpeg", (0,0)+im.size, 0, rawmode)])
def _save_cjpeg(im, fp, filename):
# ALTERNATIVE: handle JPEGs via the IJG command line utilities.
import os
file = im._dump()
os.system("cjpeg %s >%s" % (file, filename))
try: os.unlink(file)
except: pass
# -------------------------------------------------------------------q-
# Registry stuff
Image.register_open("JPEG", JpegImageFile, _accept)
Image.register_save("JPEG", _save)
Image.register_extension("JPEG", ".jfif")
Image.register_extension("JPEG", ".jpe")
Image.register_extension("JPEG", ".jpg")
Image.register_extension("JPEG", ".jpeg")
Image.register_mime("JPEG", "image/jpeg")
_______________________________________________
Image-SIG maillist - Image-SIG@python.org
http://mail.python.org/mailman/listinfo/image-sig
