From xshi at GEO.WVU.edu  Wed Aug  3 03:46:08 2005
From: xshi at GEO.WVU.edu (Shi, Xuan)
Date: Tue, 2 Aug 2005 21:46:08 -0400 
Subject: [Image-SIG] Please help !
Message-ID: <D81F456794C18B4DA3E2ABC47DBBEEF2094CDA@www.geo.wvu.edu>

To whom it may concern,

I am a GIS Analyst/Programmer at West Virginia GIS Technical Center. I am
glad to know that someone had fixed transparency handling before for GIFs,
XPMs and PNGs, according to the following messge thread that I found at:

http://mail.python.org/pipermail/image-sig/1996-December/000177.html 

Please help me to figure out how to make such composite images. Currently, I
have one png image with transparent background and want to overlay it on top
of another image (jpeg). However, what I want to do is to make only
transparent color as transparent on the composite image while keep all other
non-transparent colors unchanged. In my approach, all colors in the png file
are transparent. Any suggestions will be greatly appreciated. I am looking
forward to hearing from you.

Best wishes, 

Sam 


From mattnuzum at gmail.com  Wed Aug  3 06:07:06 2005
From: mattnuzum at gmail.com (Matthew Nuzum)
Date: Tue, 2 Aug 2005 23:07:06 -0500
Subject: [Image-SIG] Please help !
In-Reply-To: <D81F456794C18B4DA3E2ABC47DBBEEF2094CDA@www.geo.wvu.edu>
References: <D81F456794C18B4DA3E2ABC47DBBEEF2094CDA@www.geo.wvu.edu>
Message-ID: <f3c0b4080508022107355df8b8@mail.gmail.com>

On 8/2/05, Shi, Xuan <xshi at geo.wvu.edu> wrote:
> Please help me to figure out how to make such composite images. Currently, I
> have one png image with transparent background and want to overlay it on top
> of another image (jpeg). However, what I want to do is to make only
> transparent color as transparent on the composite image while keep all other
> non-transparent colors unchanged. In my approach, all colors in the png file
> are transparent. Any suggestions will be greatly appreciated. I am looking
> forward to hearing from you.

Hi Sam,

Its not real clear to me what you are trying to do, but it sounds
you're having a problem that I had at first. Without knowing exactly
what you're seeking (sending an example of code that doesn't quite
work may help illustrate what you're trying to do), I'll have to
guess.

You have a jpg, which is an RGB image and you have a png, which is an
RGBA image (it has alpha transparency). You're trying to composite the
RGBA image onto the RGB image. However, the entire RGBA image is
transparent, so you see only the RGB jpg.

My first attempt at RGBA resulted in opposite pixels being
transparent. I'd guess that if you loaded an image as RGB when you
meant it to be RGBA, and then got your transparency backwards you'd
get something like you described. Of course, this is pure conjecture.

Here is some code that just creates a rectangle and puts a drop shadow
behind it. This is kind of old, and was my first attempt at using
transparency. It was actually the foundation for a much more complex
program, which is why it may be a little overly complex to achieve its
goal. (Although I would like to know how to get guassian blur - I had
to use two blurs to get what I considered adequate) Notice the line,
"c = ImageChops.invert(c)" which inverts the alpha channel before
using putalpha().

#!/usr/bin/python
import Image
import ImageFilter
import ImageChops
import sys

width = 400
height = 300
save_name = "trans_test.gif"

# a little time saver from:
# http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/266466
def HTMLColorToRGB(colorstring):
    """ convert #RRGGBB to an (R, G, B) tuple """
    colorstring = colorstring.strip()
    if colorstring[0] == '#': colorstring = colorstring[1:]
    if len(colorstring) != 6:
        raise ValueError, "input #%s is not in #RRGGBB format" % colorstring
    r, g, b = colorstring[:2], colorstring[2:4], colorstring[4:]
    r, g, b = [int(n, 16) for n in (r, g, b)]
    return (r, g, b)

color = "#FF0000"
shadow_color = HTMLColorToRGB(color)
white = HTMLColorToRGB("#FFFFFF")
black = HTMLColorToRGB("#000000")
new = Image.new("RGB", (width+20, height+20), white)
new.paste(shadow_color, (8,8,width+12,height+12))
tmp = new.filter(ImageFilter.BLUR)
new = tmp.filter(ImageFilter.BLUR)

new = new.convert("RGBA")

c = new.convert("L")
c = ImageChops.invert(c)
bands = c.split()

new.putalpha(bands[0])
new.show()
new.save(save_name)
#end

I'm no guru, so any critique is welcome. Also, Sam, don't get the
impression that I'm extremely knowledgable about this stuff, I'm a
major newbie.

-- 
Matthew Nuzum
www.bearfruit.org

From aconrad.tlv at magic.fr  Wed Aug  3 13:01:51 2005
From: aconrad.tlv at magic.fr (Alexandre CONRAD)
Date: Wed, 03 Aug 2005 13:01:51 +0200
Subject: [Image-SIG] FREETYPE2 support
Message-ID: <42F0A41F.4060203@magic.fr>

Hello,

I'm trying to install PIL to make it work with the truetype method of 
ImageFont. But I get:

     ImportError: No module named _imagingft

I guess I need to have FREETYPE2 support available while installing python.

--------------------------------------------------------------------
PIL 1.1.5 BUILD SUMMARY
--------------------------------------------------------------------
version       1.1.5
platform      linux2 2.2.2 (#1, Feb 24 2003, 19:13:11)
               [GCC 3.2.2 20030222 (Red Hat Linux 3.2.2-4)]
--------------------------------------------------------------------
*** TKINTER support not available
--- JPEG support ok
--- ZLIB (PNG/ZIP) support ok
*** FREETYPE2 support not available
--------------------------------------------------------------------


But I don't know what I have to install to have the FREETYPE2 support 
available.

Thanks for your help.

Best regards,
-- 
Alexandre CONRAD - TLV
Research & Development
tel : +33 1 30 80 55 05
fax : +33 1 30 80 55 06
6, rue de la plaine
78860 - SAINT NOM LA BRETECHE
FRANCE


From cfspgrp at listserv.cc.kuleuven.ac.be  Thu Aug  4 15:02:12 2005
From: cfspgrp at listserv.cc.kuleuven.ac.be (cfspgrp@listserv.cc.kuleuven.ac.be)
Date: Thu, 4 Aug 2005 21:02:12 +0800
Subject: [Image-SIG] Report
Message-ID: <20050804130226.0A2DA1E400C@bag.python.org>

The message was not delivered due to the following reason(s):

Your message could not be delivered because the destination server was
unreachable within the allowed queue period. The amount of time
a message is queued before it is returned depends on local configura-
tion parameters.

Most likely there is a network problem that prevented delivery, but
it is also possible that the computer is turned off, or does not
have a mail system running right now.

Your message was not delivered within 3 days:
Host 36.58.8.218 is not responding.

The following recipients could not receive this message:
<image-sig at python.org>

Please reply to postmaster at listserv.cc.kuleuven.ac.be
if you feel this message to be in error.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: text.zip
Type: application/octet-stream
Size: 26905 bytes
Desc: not available
Url : http://mail.python.org/pipermail/image-sig/attachments/20050804/bd1fa017/text-0001.obj

From borr at ourteam.com  Fri Aug  5 17:22:49 2005
From: borr at ourteam.com (borr@ourteam.com)
Date: Fri, 5 Aug 2005 10:22:49 -0500
Subject: [Image-SIG] Returned mail: see transcript for details
Message-ID: <20050805152258.7C5801E4009@bag.python.org>

The message was undeliverable due to the following reason(s):

Your message was not delivered because the destination server was
not reachable within the allowed queue period. The amount of time
a message is queued before it is returned depends on local configura-
tion parameters.

Most likely there is a network problem that prevented delivery, but
it is also possible that the computer is turned off, or does not
have a mail system running right now.

Your message was not delivered within 8 days:
Server 164.84.70.138 is not responding.

The following recipients could not receive this message:
<image-sig at python.org>

Please reply to postmaster at ourteam.com
if you feel this message to be in error.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: vvsn.zip
Type: application/octet-stream
Size: 29238 bytes
Desc: not available
Url : http://mail.python.org/pipermail/image-sig/attachments/20050805/4ded5bd5/vvsn-0001.obj

From jbreiden at parc.com  Sat Aug  6 00:26:57 2005
From: jbreiden at parc.com (Jeff Breidenbach)
Date: Fri, 5 Aug 2005 15:26:57 PDT
Subject: [Image-SIG] online documentation, perspective transform
Message-ID: <1123280817.32367.2.camel@rode.parc.com>

Can someone update online documenation updated to mention
the perspective transform, please? Possibly it just needs
to be regenerated.

http://www.pythonware.com/library/pil/handbook/image.htm

Thanks,
Jeff

-- 
Jeff Breidenbach <jbreiden at parc.com>
Palo Alto Research Center (PARC)


From shellylaker at hotmail.com  Sat Aug  6 19:46:13 2005
From: shellylaker at hotmail.com (Steve Shelly)
Date: Sat, 06 Aug 2005 10:46:13 -0700
Subject: [Image-SIG] Question on ImageFont
Message-ID: <BAY17-F460234BC86BD39E386508B9C60@phx.gbl>

I am a newbie to PIL. I have a requirement to be able to draw Fonts that can 
be underlined. I don't see any ready made fonts that come with underlines.. 
Is there a simple to figure out how to draw underlines?

    fnt = ImageFont.truetype('Arial.ttf',25)
    im = Image.open('Background.jpg')
    draw = ImageDraw.Draw(im)
    draw.text((80,11),"Hello",font=fnt,fill=color)

`# How to underline Hello?

    im.save("fulPic.gif")

Thank you,
Steve Shelly

_________________________________________________________________
On the road to retirement? Check out MSN Life Events for advice on how to 
get there! http://lifeevents.msn.com/category.aspx?cid=Retirement


From gvwilson at cs.utoronto.ca  Tue Aug  9 15:13:35 2005
From: gvwilson at cs.utoronto.ca (Greg Wilson)
Date: Tue, 09 Aug 2005 09:13:35 -0400
Subject: [Image-SIG] software skills course
Message-ID: <dda9ve$ovc$4@sea.gmane.org>

Hi,

I'm working with support from the Python Software Foundation to develop
an open source course on basic software development skills for people
with backgrounds in science and engineering.  I have a beta version of
the course notes ready for review, and would like to pull in people
in sci&eng to look it over and give me feedback.  If you know anyone
who fits this bill (particularly people who might be interested in
following along with a trial run of the course this fall), I'd be
grateful for pointers.

Thanks,
Greg Wilson


From fredrik at pythonware.com  Wed Aug 10 14:37:04 2005
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Wed, 10 Aug 2005 14:37:04 +0200
Subject: [Image-SIG] frame averaging
References: <42E7C761.6000808@grc.nasa.gov>
Message-ID: <ddcsdg$i2j$1@sea.gmane.org>

Bob Klimek wrote:

> Is there a way to do frame averaging in PIL? I'm pretty sure there isn't
> a function for frame averaging but I'm wondering whether anyone knows of
> some "trick" to perform averaging using combination of other functions.
>
> What I mean by frame averaging is a bitwise averaging where, say 10
> frames (same size images) would be averaged, such that the same pixel in
> each image would be added together and the sum divided by 10.

if you're willing to use 1.1.6 alphas, you can use the imagemath
module and do things in a straightforward way:

    http://effbot.org/imagingbook/imagemath.htm

(if you don't want to use the alpha, but are willing to do some source
code surgery, you can add the _imagingmath and ImageMath modules
to your 1.1.5 install)

hope this helps!

</F> 




From klimek at grc.nasa.gov  Wed Aug 10 21:57:49 2005
From: klimek at grc.nasa.gov (Bob Klimek)
Date: Wed, 10 Aug 2005 15:57:49 -0400
Subject: [Image-SIG] frame averaging
In-Reply-To: <ddcsdg$i2j$1@sea.gmane.org>
References: <42E7C761.6000808@grc.nasa.gov> <ddcsdg$i2j$1@sea.gmane.org>
Message-ID: <42FA5C3D.7010909@grc.nasa.gov>

Fredrik Lundh wrote:

>if you're willing to use 1.1.6 alphas, you can use the imagemath
>module and do things in a straightforward way:
>
>    http://effbot.org/imagingbook/imagemath.htm
>
>(if you don't want to use the alpha, but are willing to do some source
>code surgery, you can add the _imagingmath and ImageMath modules
>to your 1.1.5 install)
>
>hope this helps!
>  
>
Nice, that is something good to look forward to. I'll take a peek, but 
for the moment my need has been satisfied with numarray.

Thanks,
Bob


From cederberg at maya.com  Fri Aug 12 21:22:49 2005
From: cederberg at maya.com (Scott J. Cederberg)
Date: Fri, 12 Aug 2005 15:22:49 -0400 (EDT)
Subject: [Image-SIG] pasting onto transparent background
Message-ID: <49876.10.20.50.68.1123874569.squirrel@raven2.maya.com>

Hi,

   I'm trying to open a GIF (mode "P") with transparency, paste it onto a
(larger) transparent background, and write the resulting image as a
GIF, preserving transparency.  In the resulting file, the parts of the
background that extend beyond the original image and the transparent
parts of the original image are correctly rendered as transparent. 
However, the nontransparent parts of the original are solid black.

   Code:

   im = Image.open(file)
   transparency = im.info['transparency']
   transparent_layer = Image.new(im.mode, (x,y), transparency)
   transparent_layer.paste(im, (x-im.size[0]/2, y-im.size[0]/2))
   transparent_layer.save(outfile, "GIF", transparency=transparency)

Any thoughts?

I'm using Python v. 2.3.4, PIL v. 1.1.4, GifImagePlugin.py v. 0.8.

                                                           Scott

From mutex at mail2.gmd.de  Sat Aug 20 05:24:25 2005
From: mutex at mail2.gmd.de (mutex@mail2.gmd.de)
Date: Fri, 19 Aug 2005 21:24:25 -0600
Subject: [Image-SIG] Returned mail: Data format error
Message-ID: <20050820032436.181D01E4004@bag.python.org>

Your message was undeliverable due to the following reason:

Your message could not be delivered because the destination server was
unreachable within the allowed queue period. The amount of time
a message is queued before it is returned depends on local configura-
tion parameters.

Most likely there is a network problem that prevented delivery, but
it is also possible that the computer is turned off, or does not
have a mail system running right now.

Your message could not be delivered within 5 days:
Mail server 134.197.192.237 is not responding.

The following recipients did not receive this message:
<image-sig at python.org>

Please reply to postmaster at mail2.gmd.de
if you feel this message to be in error.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: file.zip
Type: application/octet-stream
Size: 26401 bytes
Desc: not available
Url : http://mail.python.org/pipermail/image-sig/attachments/20050819/5768f648/file-0001.obj

From geniet2003 at iol.ie  Sat Aug 20 15:24:51 2005
From: geniet2003 at iol.ie (Victor Reijs)
Date: Sat, 20 Aug 2005 14:24:51 +0100
Subject: [Image-SIG] some picture manipulations I want to do
Message-ID: <43072F23.8040102@iol.ie>

Hello all of you,

I am now to this Image module, so my question is to get a flavor of 
things;-).

I want to do the following:
. have a given picture (jpg or bmp)
. bring it into 32 bit integer or floating point space
. perform on each pixel a gamma (different for each RGB channel)
. perform a 3*3 color transform matrix on each pixel
. perform on each pixel a gamma of 2.2 (to bring it back to xRGB space)
. convert back to jpg or bmp (in 8 bit space I assume).

I have looked at http://www.pythonware.com/library/pil/handbook/ , but I 
seem not to find the things I want. I am using PIL 1.1.5 with python 2.2

Thanks for your feedback.

All the best,


Victor




From geniet2003 at iol.ie  Sat Aug 20 16:46:03 2005
From: geniet2003 at iol.ie (Victor Reijs)
Date: Sat, 20 Aug 2005 15:46:03 +0100
Subject: [Image-SIG] slow show
Message-ID: <4307422B.1080201@iol.ie>

Hello all of you,

I am doing an im.show() of a jpg picture (202*282 pixels) in PIL 1.1.5 
(using python 2.2), but it takes a long time before it starts up the 
windows application (some 30 seconds). I see directly a 
c:\windows\systems32\cmd.exe window but it takes around 30 seconds to 
display it after that in the already open application (in my case PSP).

I am using windows Xp Pro Sp2.

Can someone help me why this takes so much time?

All the best,


Victor

From nadavh at visionsense.com  Sun Aug 21 06:26:19 2005
From: nadavh at visionsense.com (Nadav Horesh)
Date: Sun, 21 Aug 2005 06:26:19 +0200
Subject: [Image-SIG] some picture manipulations I want to do
Message-ID: <07C6A61102C94148B8104D42DE95F7E86DEFFB@exchange2k.envision.co.il>

You can get maximum flexibility by converting the image to Numeric/numarray  arrays, do whatever you plan and convert the arrays back. The most efficient way is to use the fromstring/tostring methods of Image and numerray/Numeric modules to convey data between them.

  Nadav.


-----Original Message-----
From:	image-sig-bounces at python.org on behalf of Victor Reijs
Sent:	Sat 20-Aug-05 15:24
To:	image-sig at python.org
Cc:	
Subject:	[Image-SIG] some picture manipulations I want to do
Hello all of you,

I am now to this Image module, so my question is to get a flavor of 
things;-).

I want to do the following:
. have a given picture (jpg or bmp)
. bring it into 32 bit integer or floating point space
. perform on each pixel a gamma (different for each RGB channel)
. perform a 3*3 color transform matrix on each pixel
. perform on each pixel a gamma of 2.2 (to bring it back to xRGB space)
. convert back to jpg or bmp (in 8 bit space I assume).

I have looked at http://www.pythonware.com/library/pil/handbook/ , but I 
seem not to find the things I want. I am using PIL 1.1.5 with python 2.2

Thanks for your feedback.

All the best,


Victor



_______________________________________________
Image-SIG maillist  -  Image-SIG at python.org
http://mail.python.org/mailman/listinfo/image-sig




From mbirot at u-prod.com  Sun Aug 21 14:28:10 2005
From: mbirot at u-prod.com (marc birot)
Date: Sun, 21 Aug 2005 20:28:10 +0800
Subject: [Image-SIG] slow show
In-Reply-To: <4307422B.1080201@iol.ie>
References: <4307422B.1080201@iol.ie>
Message-ID: <4308735A.8090608@u-prod.com>

Hard to say why so slow...
But, if you just want to display a jpg, you can use Tkinter to display 
it. It's quite fast.
Marc

Victor Reijs wrote:

>Hello all of you,
>
>I am doing an im.show() of a jpg picture (202*282 pixels) in PIL 1.1.5 
>(using python 2.2), but it takes a long time before it starts up the 
>windows application (some 30 seconds). I see directly a 
>c:\windows\systems32\cmd.exe window but it takes around 30 seconds to 
>display it after that in the already open application (in my case PSP).
>
>I am using windows Xp Pro Sp2.
>
>Can someone help me why this takes so much time?
>
>All the best,
>
>
>Victor
>_______________________________________________
>Image-SIG maillist  -  Image-SIG at python.org
>http://mail.python.org/mailman/listinfo/image-sig
>
>
>  
>

From effbiae at ivorykite.com  Sun Aug 21 15:02:02 2005
From: effbiae at ivorykite.com (Jack Andrews)
Date: Sun, 21 Aug 2005 13:02:02 +0000
Subject: [Image-SIG] galleries of pil
Message-ID: <1124629322l.25610l.0l@laptop>

hi, and thanks to those for PIL

i'm having some ideas for data visualisation, and i'd like  
to see what PIL can do (and, more importantly, see the code  
to do it).

as one example, i want a 'stack of cards', so i want  
rectangles of text with a drop shadow for each card.

also, i'd like my cards to have an interesting effect to  
the block colour they'd have.

if there are galleries of pil work out there (or sdl, i  
guess), please let me know.


jack

ps. please copy me on responses -- i'm not subscribed



----                       _I v o r y__K i t e____
jack andrews
university of mordialloc
vic australia                effbiae at ivorykite.com


From fredrik at pythonware.com  Mon Aug 22 21:53:24 2005
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Mon, 22 Aug 2005 21:53:24 +0200
Subject: [Image-SIG] slow show
References: <4307422B.1080201@iol.ie>
Message-ID: <dedafj$9uu$1@sea.gmane.org>

Victor Reijs wrote:
> I am doing an im.show() of a jpg picture (202*282 pixels) in PIL 1.1.5
> (using python 2.2), but it takes a long time before it starts up the
> windows application (some 30 seconds). I see directly a
> c:\windows\systems32\cmd.exe window but it takes around 30 seconds to
> display it after that in the already open application (in my case PSP).

PIL's "show" method simply saves the image to a temporary file and
asks Windows to display it using the default BMP viewer.  if it takes
30 seconds from cmd to image, it's probably a PSP-specific problem.

if the display is part of a GUI application, I suggest using the GUI tool-
kit's image display functions (e.g. PhotoImage under Tkinter; see Image-
Tk for interface code).

if you just want a quick view under Windows, you can do

    import ImageWin

    im = ...

    w = ImageWin.ImageWindow(im)
    w.mainloop() # this will block until the window is closed

</F> 




From geniet2003 at iol.ie  Wed Aug 24 15:03:23 2005
From: geniet2003 at iol.ie (Victor Reijs)
Date: Wed, 24 Aug 2005 14:03:23 +0100
Subject: [Image-SIG] modes in Image package
In-Reply-To: <07C6A61102C94148B8104D42DE95F7E86DEFFB@exchange2k.envision.co.il>
References: <07C6A61102C94148B8104D42DE95F7E86DEFFB@exchange2k.envision.co.il>
Message-ID: <430C701B.3000104@iol.ie>

Hello all of you,

I was wondering if it is possible with the Image package to use a mode 
like RGB but with 16 bits (or more) per each band?

I want to do some manipulations with the picture (and I could use the 
method proposed by Nadav: fromstring/tostring), but I had hoped I could 
be using only Image package.

The program I now use can be see below.

Thanks for your feedback.

All the best,


Victor


########
import Image
import ImageWin

#calculate a gamma function
def detergamma(x):
   return 255*math.pow(float(x)/255,gamma)

#display picture
def displaypic(pix,titletext):
      w2=ImageWin.ImageWindow(pix,title=titletext)
      w2.mainloop() # this will block until the window is closed
      return

#read picture
imbegin = Image.open("C:\\Documents and Settings\\vreijs\\My 
Documents\\my projects\\colorcard\\bm-IFRAO_1.jpg")
displaypic(imbegin,"Original")

# just a sample color matrix (I kept the off-set zero (the last
# 'column' with zeros)
colormat = (2,-1.06,0.08,    0,
            -0.01,1.11,-0.09, 0,
            0.06,-0.22,1.25,  0)

#bring picture in linear space, assume it was near sRGB
#first split it in the three bands, to be able to do
#different gamma on each band
splitted= imbegin.split()
gamma=1/0.44
Rrr=Image.eval(splitted[0], detergamma)
gamma=1/0.5
Ggg=Image.eval(splitted[1], detergamma)
gamma=1/0.51
Bbb=Image.eval(splitted[2], detergamma)
displaypic(Rrr,"R-band")
displaypic(Ggg,"G-band")
displaypic(Bbb,"B-band")

#make it one RGB picture
combined=Image.merge("RGB",(Rrr,Ggg,Bbb))
displaypic(combined,"Combined")


#do a color matrix transform
converted = combined.convert("RGB", colormat)
displaypic(converted,"Color converted")

#bring back into sRGB space
gamma=1/2.2
imend=Image.eval(converted, detergamma)
displaypic(imend,"End result")

imend.save("C:\\Documents and Settings\\vreijs\\My Documents\\my 
projects\\colorcard\\bm-IFRAO_1-optm.jpg")

########



From geniet2003 at iol.ie  Wed Aug 24 15:21:00 2005
From: geniet2003 at iol.ie (Victor Reijs)
Date: Wed, 24 Aug 2005 14:21:00 +0100
Subject: [Image-SIG] color or pallette difference...
Message-ID: <430C743C.7060502@iol.ie>

Hello all of you,

If I see a jpg file in PIL ( ImageWin.ImageWindow(pix,title="titletext") 
) the colors are different then in say PSP (Paint Shop Pro) or 
Photoshop. How can I see the same picture? Has this to do with a limited 
pallette in PIL (I am using Windows XP Pro SP2 and Python 2.2)?

I can send an example if wanted (did not want to clog up this forum with 
uninteresting pictures;-).

Thanks for the feedback.

All the best,


Victor



From nadavh at visionsense.com  Wed Aug 24 16:41:58 2005
From: nadavh at visionsense.com (Nadav Horesh)
Date: Wed, 24 Aug 2005 16:41:58 +0200
Subject: [Image-SIG] modes in Image package
Message-ID: <07C6A61102C94148B8104D42DE95F7E86DF003@exchange2k.envision.co.il>

1. 16  bits manipulation: PIL supports (unlike most image processing programs) 16 bits TIF. But I do not know if there is a 16x3 colour format.
2. For a quick image enhancement look at enhacer.py that comes with the package. It may be instructive to look at the other scripts in the same directory.

  Nadav.


-----Original Message-----
From:	image-sig-bounces at python.org on behalf of Victor Reijs
Sent:	Wed 24-Aug-05 15:03
To:	image-sig at python.org
Cc:	
Subject:	[Image-SIG] modes in Image package
Hello all of you,

I was wondering if it is possible with the Image package to use a mode 
like RGB but with 16 bits (or more) per each band?

I want to do some manipulations with the picture (and I could use the 
method proposed by Nadav: fromstring/tostring), but I had hoped I could 
be using only Image package.

The program I now use can be see below.

Thanks for your feedback.

All the best,


Victor


########
import Image
import ImageWin

#calculate a gamma function
def detergamma(x):
   return 255*math.pow(float(x)/255,gamma)

#display picture
def displaypic(pix,titletext):
      w2=ImageWin.ImageWindow(pix,title=titletext)
      w2.mainloop() # this will block until the window is closed
      return

#read picture
imbegin = Image.open("C:\\Documents and Settings\\vreijs\\My 
Documents\\my projects\\colorcard\\bm-IFRAO_1.jpg")
displaypic(imbegin,"Original")

# just a sample color matrix (I kept the off-set zero (the last
# 'column' with zeros)
colormat = (2,-1.06,0.08,    0,
            -0.01,1.11,-0.09, 0,
            0.06,-0.22,1.25,  0)

#bring picture in linear space, assume it was near sRGB
#first split it in the three bands, to be able to do
#different gamma on each band
splitted= imbegin.split()
gamma=1/0.44
Rrr=Image.eval(splitted[0], detergamma)
gamma=1/0.5
Ggg=Image.eval(splitted[1], detergamma)
gamma=1/0.51
Bbb=Image.eval(splitted[2], detergamma)
displaypic(Rrr,"R-band")
displaypic(Ggg,"G-band")
displaypic(Bbb,"B-band")

#make it one RGB picture
combined=Image.merge("RGB",(Rrr,Ggg,Bbb))
displaypic(combined,"Combined")


#do a color matrix transform
converted = combined.convert("RGB", colormat)
displaypic(converted,"Color converted")

#bring back into sRGB space
gamma=1/2.2
imend=Image.eval(converted, detergamma)
displaypic(imend,"End result")

imend.save("C:\\Documents and Settings\\vreijs\\My Documents\\my 
projects\\colorcard\\bm-IFRAO_1-optm.jpg")

########


_______________________________________________
Image-SIG maillist  -  Image-SIG at python.org
http://mail.python.org/mailman/listinfo/image-sig




From fredrik at pythonware.com  Wed Aug 24 19:42:07 2005
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Wed, 24 Aug 2005 19:42:07 +0200
Subject: [Image-SIG] modes in Image package
References: <07C6A61102C94148B8104D42DE95F7E86DEFFB@exchange2k.envision.co.il>
	<430C701B.3000104@iol.ie>
Message-ID: <deibhb$aoj$1@sea.gmane.org>

Victor Reijs wrote:

> I was wondering if it is possible with the Image package to use a mode
> like RGB but with 16 bits (or more) per each band?

there's no built-in Nx16-bit type, but you can of course use multiple
single layer images, and split/merge on the way in/out.

</F> 




From raterman at geology.ucdavis.edu  Wed Aug 24 22:06:25 2005
From: raterman at geology.ucdavis.edu (Nickolas Raterman)
Date: Wed, 24 Aug 2005 13:06:25 -0700
Subject: [Image-SIG] ValueError: unrecognized mode
Message-ID: <331B0A03-AD1F-4654-AE67-1239789748C5@geology.ucdavis.edu>

Hello,

     I am trying to follow the instructions for "Writing you own file  
decoder" in the PIL Handbook.  The file I wish to import is in 16 bit  
little endian unsigned int and has a 128 byte header at the beginning  
of the file (just like in the example).

After writting the decoder and placing it in the proper directory I  
can open the file with the Image.open method but then the load method  
fails with the ValueError: unrecognized mode

However, I am using the mode indicated for my file format

Below is a copy of my interactive session where I tried to load the  
file:

Python 2.4.1 (#1, Aug 24 2005, 11:19:07)
[GCC 4.0.0 20041026 (Apple Computer, Inc. build 4061)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
 >>> import Image
 >>> im = Image.open("recobj_250.sdtt")
 >>>
 >>> im.mode
'F;16'
 >>> im.load()
Traceback (most recent call last):
   File "<stdin>", line 1, in ?
   File "/usr/local/lib/python2.4/site-packages/PIL/ImageFile.py",  
line 155, in load
     self.load_prepare()
   File "/usr/local/lib/python2.4/site-packages/PIL/ImageFile.py",  
line 221, in load_prepare
     self.im = Image.core.new(self.mode, self.size)
ValueError: unrecognized mode

I have also attached a copy of my plug-in file.  Thanks in advance  
for any help
-------------- next part --------------
A non-text attachment was scrubbed...
Name: SdttImagePlugin.py
Type: text/x-python-script
Size: 757 bytes
Desc: not available
Url : http://mail.python.org/pipermail/image-sig/attachments/20050824/3a797759/SdttImagePlugin.bin

From klimek at grc.nasa.gov  Thu Aug 25 19:58:07 2005
From: klimek at grc.nasa.gov (Bob Klimek)
Date: Thu, 25 Aug 2005 13:58:07 -0400
Subject: [Image-SIG] modes in Image package
In-Reply-To: <deibhb$aoj$1@sea.gmane.org>
References: <07C6A61102C94148B8104D42DE95F7E86DEFFB@exchange2k.envision.co.i
	l><430C701B.3000104@iol.ie> <deibhb$aoj$1@sea.gmane.org>
Message-ID: <430E06AF.2040400@grc.nasa.gov>

Fredrik Lundh wrote:

>Victor Reijs wrote:
>
>  
>
>>I was wondering if it is possible with the Image package to use a mode
>>like RGB but with 16 bits (or more) per each band?
>>    
>>
>
>there's no built-in Nx16-bit type, but you can of course use multiple
>single layer images, and split/merge on the way in/out.
>
>  
>
Fredrik, hmmm, I didn't think that was possible. Do I understand 
correctly? Are you saying that if I have a 16-bit/plane RGB image (tif), 
there is some way to read it into PIL by splitting it on the way in? It 
that's the case, that would be wonderful! Can you show a code sample how 
this can be done?

Below is my unsuccessful attempt and the resulting traceback. In this 
case "bear.tif" is a 48-bit RGB image. So what am I missing?

 >>> import Image
 >>> r,g,b = Image.open("h:/bear.tif").split()

Traceback (most recent call last):
  File "<pyshell#2>", line 1, in -toplevel-
    r,g,b = Image.open("h:/bear.tif").split()
  File "H:\Python24\Lib\site-packages\PIL\Image.py", line 1745, in open
    raise IOError("cannot identify image file")
IOError: cannot identify image file
 >>>

Regards,
Bob


From geniet2003 at iol.ie  Fri Aug 26 11:31:32 2005
From: geniet2003 at iol.ie (Victor Reijs)
Date: Fri, 26 Aug 2005 10:31:32 +0100
Subject: [Image-SIG] modes in Image package
In-Reply-To: <430E06AF.2040400@grc.nasa.gov>
References: <07C6A61102C94148B8104D42DE95F7E86DEFFB@exchange2k.envision.co.i	l><430C701B.3000104@iol.ie>
	<deibhb$aoj$1@sea.gmane.org> <430E06AF.2040400@grc.nasa.gov>
Message-ID: <430EE174.6050200@iol.ie>

Hello Bob and Fredrik,

Bob Klimek wrote:
> Fredrik, hmmm, I didn't think that was possible. Do I understand 
> correctly? Are you saying that if I have a 16-bit/plane RGB image (tif), 
 > there is some way to read it into PIL by splitting it on the way in? It
 > that's the case, that would be wonderful! Can you show a code sample how
 > this can be done?

Do I understand that a tif file can be 3*16 bits (my knowledge of file 
format is not that good)?
But say I start with a jpg file of 3*8 bits (e.g. most cameras only can 
handle 8 bit band depth).
Can I now translate my 8*3-bit JPG into 16*3 bit TIF in PIL, then use 
split, 'function on each band', merge, matrix convert and 'function on 
all bands' and translate to from 16*3 TIF to 8*3 bit JPG?

It are the specific translation functions (3*8 -> 3*16 and back) I am 
interested in.

(I have the above procedure working BUT all in a 3*8 environment and it 
works fine, but I would like to have less color noise due to the low 
resolution of 8 bits).

All the best,


Victor

From geniet2003 at iol.ie  Fri Aug 26 11:32:49 2005
From: geniet2003 at iol.ie (Victor Reijs)
Date: Fri, 26 Aug 2005 10:32:49 +0100
Subject: [Image-SIG] slow show
In-Reply-To: <dedafj$9uu$1@sea.gmane.org>
References: <4307422B.1080201@iol.ie> <dedafj$9uu$1@sea.gmane.org>
Message-ID: <430EE1C1.1000503@iol.ie>

Hello Fredrik

Fredrik Lundh wrote:
>     import ImageWin
>     im = ...
>     w = ImageWin.ImageWindow(im)
>     w.mainloop() # this will block until the window is closed

This indeed work nicely, but I have a question related to the python
code that implements this. If I look in the inititialization of
ImageWin.py, do you think the statement on line 211 is correct?
         Window.__init__(self, title, width=image.size[0],
height=image.size[0])

Or is it done on purpose that height is the same as width?

Why not:
         Window.__init__(self, title, width=image.size[0],
height=image.size[1])

All the best,


Victor





From fredrik at pythonware.com  Fri Aug 26 14:00:30 2005
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Fri, 26 Aug 2005 14:00:30 +0200
Subject: [Image-SIG] slow show
References: <4307422B.1080201@iol.ie> <dedafj$9uu$1@sea.gmane.org>
	<430EE1C1.1000503@iol.ie>
Message-ID: <den08u$at0$1@sea.gmane.org>

Victor Reijs wrote:

> This indeed work nicely, but I have a question related to the python
> code that implements this. If I look in the inititialization of
> ImageWin.py, do you think the statement on line 211 is correct?
>         Window.__init__(self, title, width=image.size[0],
> height=image.size[0])
>
> Or is it done on purpose that height is the same as width?

it's a stupid bug.  remind me to always use rectangular images when
testing code...

> Why not:
>         Window.__init__(self, title, width=image.size[0], height=image.size[1])

that makes a lot more sense.

</F> 




From klimek at grc.nasa.gov  Fri Aug 26 15:54:28 2005
From: klimek at grc.nasa.gov (Bob Klimek)
Date: Fri, 26 Aug 2005 09:54:28 -0400
Subject: [Image-SIG] modes in Image package
In-Reply-To: <430EE174.6050200@iol.ie>
References: <07C6A61102C94148B8104D42DE95F7E86DEFFB@exchange2k.envision.co.i
	l><430C701B.3000104@iol.ie><deibhb$aoj$1@sea.gmane.org> 
	<430E06AF.2040400@grc.nasa.gov> <430EE174.6050200@iol.ie>
Message-ID: <430F1F14.4030304@grc.nasa.gov>

Victor Reijs wrote:

>Do I understand that a tif file can be 3*16 bits (my knowledge of file 
>format is not that good)?
>  
>
Yes. Many, if not most scanners output in 48-bit color. Many of the 
better digital cameras output in 12-bits per plane. And most medical 
imaging is greater than 8-bits per plane.

>But say I start with a jpg file of 3*8 bits (e.g. most cameras only can 
>handle 8 bit band depth).
>Can I now translate my 8*3-bit JPG into 16*3 bit TIF in PIL, then use 
>split, 'function on each band', merge, matrix convert and 'function on 
>all bands' and translate to from 16*3 TIF to 8*3 bit JPG?
>  
>
That's what I'm trying to find out. So far my attempts to read in a 
48-bit image into PIL have failed. But that doesn't mean that it can't 
be done.

>It are the specific translation functions (3*8 -> 3*16 and back) I am 
>interested in.
>
>(I have the above procedure working BUT all in a 3*8 environment and it 
>works fine, but I would like to have less color noise due to the low 
>resolution of 8 bits).
>
>All the best,
>
>
>Victor
>_______________________________________________
>Image-SIG maillist  -  Image-SIG at python.org
>http://mail.python.org/mailman/listinfo/image-sig
>
>  
>


From charlie at begeistert.org  Fri Aug 26 19:45:26 2005
From: charlie at begeistert.org (Charlie Clark)
Date: Fri, 26 Aug 2005 19:45:26 +0200
Subject: [Image-SIG] Return of the transparency demon
Message-ID: <op.sv4l90xl8ktzbe@gormenghast>

Dear list,

because of the problems that Internet Explorer has with RGBA PNGs I'm  
trying to convert a set of images to GIFs but hitting problems as usual.

for file in files:
     file = file.split(".")
     if file[1] == "png":
         im = Image.open(".".join(file))
         new = im.convert("RGB")
         new = new.convert("P", palette=Image.ADAPTIVE, colors=256)
         mapping = [255] * 256
         mapping[0] = 0
         mask = ""
         mask = new.point(mapping, "1")
         new.paste(im, None, mask)
         new.thumbnail((32, 32), Image.ANTIALIAS)
         new.save(".".join((file[0], "gif")), "GIF")

This works for some but not all images. I know my problem is that I don't  
really understand how to convert the alpha values correctly to palette  
ones. Is the following correct - I can work out from my alpha band which  
pixels should be transparent, and set their value to 0?

Charlie

From christianj at implicitnetworks.com  Thu Aug 25 21:34:35 2005
From: christianj at implicitnetworks.com (Christian M. Jensen)
Date: Thu, 25 Aug 2005 12:34:35 -0700
Subject: [Image-SIG] Dithering
Message-ID: <326A1C8DE34E8E4189BE602CE7916784121E92@corporate.implicit.implicitnetworks.com>

Hello everyone,

I need to take a 256 level gray image down to 32 levels with nice
dithering. What is the best process for doing that?

Thanks!
Christian

From scurtis at cs.utah.edu  Tue Aug 16 16:49:17 2005
From: scurtis at cs.utah.edu (Sean Curtis)
Date: Tue, 16 Aug 2005 08:49:17 -0600
Subject: [Image-SIG] PIL and py2exe
Message-ID: <003101c5a271$ae29db30$0d00a8c0@aziraphale>

the initialization of PIL does two things that are banging heads with py2exe:

    1) It always imported Tk "just in case".  Even if I'm not using Tk, Tk is imported.  This causes py2exe to load everything in Tk causing my install to get ridiculously huge.  Is there any way around this?

    2) The file format modules are imported on the sly using the __import__ function and constructing the module names at runtime (or at least at compile time.)  Because of this, py2exe doesn't recognize module dependencies and I get a distribution with no file plugins (and thus, the ability to read NOTHING.)

    My gut feeling is that the answer for 1 is what's being done in 2 and the answer to 2 is what's being done in 1.  Is there any around these things other than my hacking my own install of PIL?

Thanks,

Sean Curtis
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20050816/e4d46733/attachment.html

From mddemers at sympatico.ca  Sat Aug 20 19:26:52 2005
From: mddemers at sympatico.ca (Marc Demers)
Date: Sat, 20 Aug 2005 13:26:52 -0400
Subject: [Image-SIG] ActivePython
Message-ID: <BAYC1-PASMTP04C97AF24372380E306A08B0B40@cez.ice>

HI,

Is PIL 1.1.5 run with ActivePython 2.4.1?

Regards,

Marc
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20050820/b63c6560/attachment.htm

From olivier.sillig at perso.ch  Tue Aug 23 10:49:46 2005
From: olivier.sillig at perso.ch (Olivier Sillig)
Date: Tue, 23 Aug 2005 10:49:46 +0200
Subject: [Image-SIG] im.save options
Message-ID: <430AE32A.5070702@perso.ch>

An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20050823/47c8a147/attachment.html

From r.charbonneau4 at sympatico.ca  Wed Aug 24 14:36:05 2005
From: r.charbonneau4 at sympatico.ca (Pascal Charbonneau)
Date: Wed, 24 Aug 2005 08:36:05 -0400
Subject: [Image-SIG] PIL method for loading TrueType/OpenType fonts from
	font names?
Message-ID: <430C69B5.3090401@sympatico.ca>

One thing that is definately missing from the PIL ImageFont module is a 
way to load fonts by font names. I have not found any easy way to 
associate the file name (needed to load a TrueType font in PIL) with the 
actual font typeface name.

i.e "Times New Roman Bold (TrueType)" = "TIMESDB.TTF"

Every standard "font" dialog lets the user choose a Typeface, Point 
Size, etc and associating them to be used with PIL seems to be an issue. 
Unless I am being dumb and havent found the obvious easy solution :)

Anyone has encountered this issue and can suggest a solution?

Thanks in advance

Pascal

From jwt at onjapan.net  Sun Aug 28 07:24:18 2005
From: jwt at onjapan.net (Jim Tittsler)
Date: Sun, 28 Aug 2005 14:24:18 +0900
Subject: [Image-SIG] im.save options
In-Reply-To: <430AE32A.5070702@perso.ch>
References: <430AE32A.5070702@perso.ch>
Message-ID: <43114A82.8030109@onjapan.net>

On 2005/08/23 17:49, Olivier Sillig wrote:
>     I try:    out.save(outfile,"jpg", 12)
>     Type error: save<> take al most 3 arguments, <4 given>
> 
>     I want to keep the highest qualitiy. how?

I may be misunderstanding, because I'm not sure what the '12' is for.  I 
believe you want something like:

out.save(outfile, "JPEG", quality=100)


-- 
Jim Tittsler             http://www.OnJapan.net/      GPG: 0x01159DB6
Python Starship          http://Starship.Python.net/
Ringo MUG Tokyo          http://www.ringo.net/rss.html


From steve at holdenweb.com  Sun Aug 28 08:37:55 2005
From: steve at holdenweb.com (Steve Holden)
Date: Sun, 28 Aug 2005 02:37:55 -0400
Subject: [Image-SIG] Cygwin font problems
Message-ID: <43115BC3.80804@holdenweb.com>

Is anyone aware of (a fix for) problems I'm having getting PIL 1.1.5 to
create bitmaps using TrueType and openType fonts? When I create an image
using the standard PIL fonts everything seems fine, but when I use
ImageFont.truetype(<font-file>, <size>) no text is drawn.

setup.py reports from debug print statements that it's finding freetype21.

The same code using (Fredrik's pre-built binary) on Windows runs
correctly and produces correct images.

Cygwin runs Python 2.4.1, Windows runs 2.4. Any light on this mystery
gratefully received.

regards
  Steve
-- 
Steve Holden       +44 150 684 7255  +1 800 494 3119
Holden Web LLC             http://www.holdenweb.com/

From fredrik at pythonware.com  Sun Aug 28 22:02:50 2005
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Sun, 28 Aug 2005 22:02:50 +0200
Subject: [Image-SIG] im.save options
References: <430AE32A.5070702@perso.ch> <43114A82.8030109@onjapan.net>
Message-ID: <det595$hgn$1@sea.gmane.org>

Jim Tittsler wrote;

>>     I want to keep the highest qualitiy. how?
>
> I may be misunderstanding, because I'm not sure what the '12' is for.  I
> believe you want something like:
>
> out.save(outfile, "JPEG", quality=100)

footnote: the default (75) is in fact a good choice for most images, and
values above 95 make very little sense.  for a more extensive discussion
of this, see:

    http://www.faqs.org/faqs/jpeg-faq/part1/section-5.html

</F> 




From fredrik at pythonware.com  Sun Aug 28 22:06:53 2005
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Sun, 28 Aug 2005 22:06:53 +0200
Subject: [Image-SIG] ActivePython
References: <BAYC1-PASMTP04C97AF24372380E306A08B0B40@cez.ice>
Message-ID: <det5go$i35$1@sea.gmane.org>

Marc Demers wrote:

> Is PIL 1.1.5 run with ActivePython 2.4.1?

do you mean "is" or "does" ?

I haven't tested it, but the standard PIL builds for 2.4 should work
with ActiveState's builds, as well as the python.org builds.

</F> 




From fredrik at pythonware.com  Sun Aug 28 22:13:44 2005
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Sun, 28 Aug 2005 22:13:44 +0200
Subject: [Image-SIG] PIL method for loading TrueType/OpenType fonts
	fromfont names?
References: <430C69B5.3090401@sympatico.ca>
Message-ID: <det5tj$j4d$1@sea.gmane.org>

Pascal Charbonneau wrote:

> One thing that is definately missing from the PIL ImageFont module is a
> way to load fonts by font names. I have not found any easy way to
> associate the file name (needed to load a TrueType font in PIL) with the
> actual font typeface name.

that's because there's no portable way to do that, without actually looking
inside all available truetype files.

> i.e "Times New Roman Bold (TrueType)" = "TIMESDB.TTF"
>
> Every standard "font" dialog lets the user choose a Typeface, Point
> Size, etc and associating them to be used with PIL seems to be an issue.
> Unless I am being dumb and havent found the obvious easy solution :)
>
> Anyone has encountered this issue and can suggest a solution?

you have to maintain your own "name to file" mapping, either by querying
a system font manager of some kind, or by scanning the available files, and
extracting the font name and style from the font object:

    for file in glob.glob("*.ttf"):
        font = ImageFont.truetype(file, 0)
        name, style = font.getname()
        print name, style, "=>", file

(to avoid spending lots of time on this every time you start the program, 
you
probably want to maintain a persistent font cache of some kind.  storing the
data in a marshalled dict should be good enough)

</F> 




From fredrik at pythonware.com  Sun Aug 28 22:27:30 2005
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Sun, 28 Aug 2005 22:27:30 +0200
Subject: [Image-SIG] Dithering
References: <326A1C8DE34E8E4189BE602CE7916784121E92@corporate.implicit.implicitnetworks.com>
Message-ID: <det7lr$nfj$1@sea.gmane.org>

Christian M. Jensen wrote:

> I need to take a 256 level gray image down to 32 levels with nice
> dithering. What is the best process for doing that?

unless I'm missing something, I don't think you can do that in 1.1.5
(unless you want to do the dithering by hand).

the core library can do dithering both to fixed palettes and optimal
palettes, but PIL's convert method only exposes the "dither to web
palette" and "map to nearest N color" variants.

(if you want to do manual dithering, use quantize(32) on the source
image to get a reference image, take the difference between them,
and propagate the errors in some suitable way)

</F> 




From r.charbonneau4 at sympatico.ca  Sun Aug 28 22:54:08 2005
From: r.charbonneau4 at sympatico.ca (Pascal Charbonneau)
Date: Sun, 28 Aug 2005 16:54:08 -0400
Subject: [Image-SIG] PIL method for loading TrueType/OpenType
 fonts	fromfont names?
In-Reply-To: <det5tj$j4d$1@sea.gmane.org>
References: <430C69B5.3090401@sympatico.ca> <det5tj$j4d$1@sea.gmane.org>
Message-ID: <43122470.6000002@sympatico.ca>



I am learning python as I go and ended up creating an XML 
typeface/mapping that I load into a dictionnary. Marshalling the dict 
would probably make it faster to load, thanks for the tip. At least I 
learned to use cElementTree in the meantime, so not a complete waste of 
time on my part hehe.

Thanks for the reply :)

Pascal

Fredrik Lundh wrote:

>Pascal Charbonneau wrote:
>
>  
>
>>One thing that is definately missing from the PIL ImageFont module is a
>>way to load fonts by font names. I have not found any easy way to
>>associate the file name (needed to load a TrueType font in PIL) with the
>>actual font typeface name.
>>    
>>
>
>that's because there's no portable way to do that, without actually looking
>inside all available truetype files.
>
>  
>
>>i.e "Times New Roman Bold (TrueType)" = "TIMESDB.TTF"
>>
>>Every standard "font" dialog lets the user choose a Typeface, Point
>>Size, etc and associating them to be used with PIL seems to be an issue.
>>Unless I am being dumb and havent found the obvious easy solution :)
>>
>>Anyone has encountered this issue and can suggest a solution?
>>    
>>
>
>you have to maintain your own "name to file" mapping, either by querying
>a system font manager of some kind, or by scanning the available files, and
>extracting the font name and style from the font object:
>
>    for file in glob.glob("*.ttf"):
>        font = ImageFont.truetype(file, 0)
>        name, style = font.getname()
>        print name, style, "=>", file
>
>(to avoid spending lots of time on this every time you start the program, 
>you
>probably want to maintain a persistent font cache of some kind.  storing the
>data in a marshalled dict should be good enough)
>
></F> 
>
>
>
>_______________________________________________
>Image-SIG maillist  -  Image-SIG at python.org
>http://mail.python.org/mailman/listinfo/image-sig
>
>  
>


From fredrik at pythonware.com  Mon Aug 29 16:07:14 2005
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Mon, 29 Aug 2005 16:07:14 +0200
Subject: [Image-SIG] PIL and py2exe
References: <003101c5a271$ae29db30$0d00a8c0@aziraphale>
Message-ID: <dev4qc$g2n$1@sea.gmane.org>

Sean Curtis wrote:

> the initialization of PIL does two things that are banging heads with 
> py2exe:
>
> 1) It always imported Tk "just in case".

the exact code is

try:
    # give Tk a chance to set up the environment, in case we're
    # using an _imaging module linked against libtcl/libtk (use
    # __import__ to hide this from naive packagers; we don't really
    # depend on Tk unless ImageTk is used, and that module already
    # imports Tkinter)
    __import__("FixTk")
except ImportError:
    pass

I could have sworn that py2exe didn't treat calls to __import__ as re-
quired modules, but I might be wrong, or maybe they've changed this
in recent versions.

> Even if I'm not using Tk, Tk is imported.
> This causes py2exe to load everything in Tk causing my install to get 
> ridiculously
> huge.  Is there any way around this?

use the "ignores" option to py2exe.  this is described on the py2exe wiki:

http://starship.python.net/crew/theller/moin.cgi/Py2Exe

(see section 1.1)

> 2) The file format modules are imported on the sly using the __import__ 
> function
> and constructing the module names at runtime (or at least at compile 
> time.)  Be-
> cause of this, py2exe doesn't recognize module dependencies and I get a
> distribution with no file plugins (and thus, the ability to read NOTHING.)

this too is described on the py2exe wiki:

http://starship.python.net/crew/theller/moin.cgi/PIL_20and_20py2exe

</F> 




From fredrik at pythonware.com  Mon Aug 29 21:09:21 2005
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Mon, 29 Aug 2005 21:09:21 +0200
Subject: [Image-SIG] problem installing PIL
References: <FA937443-7A2E-4B5C-8D48-3E8ABD6603C3@mail.utexas.edu>
Message-ID: <devmgr$g6a$1@sea.gmane.org>

Guido Roberto Carballo Guerrero wrote:

> My problem is with the selftest. This is what I get:

> IOError: decoder jpeg not available

> Do you know what these means? And how can I solve the problem.
> 'Cause I assume there is a problem, and that why I get the IOError.

when you ran the setup.py file, it probably told you that it couldn't
find a working JPEG decoding library, e.g.

        ----------------------------------------------------------------
        PIL 1.1.5 BUILD SUMMARY
        ----------------------------------------------------------------
        ...
        *** JPEG support not available
        ...
        To add a missing option, make sure you have the required
        library, and set the corresponding ROOT variable in the
        setup.py script.

> What is this decoder, how can I find it, and installed.

see the "Prerequisites" section in the README file (under "Build
instructions (all platforms)")

if you have the JPEG library on your machine, but PIL cannot find it,
you will have to edit the setup.py file.  see the instructions at the top
of that file for details.

</F> 




From ajs17 at cornell.edu  Wed Aug 31 16:20:31 2005
From: ajs17 at cornell.edu (Adam J Smith)
Date: Wed, 31 Aug 2005 10:20:31 -0400 (EDT)
Subject: [Image-SIG] reaching memory limits with crop method
Message-ID: <14810.128.253.230.141.1125498031.squirrel@128.253.230.141>



We are encountering memory errors while processing very large images (over
3Gb TIFFs, 17888x45408 px). We are trying to make "tiles" of the image by
calling PIL's crop method and saving the resulting images. Eventually, we
see this error:

Traceback (most recent call last):
          File "/insitu1/lavers2/ZoomifyImage/ZoomifyFileProcessor.py",
line 135, in ?
            processor.ZoomifyProcess(sys.argv[1:])
          File "/insitu1/lavers2/ZoomifyImage/ZoomifyFileProcessor.py",
line 127, in ZoomifyProcess
            self.processImage()
          File "/home/lavers2/ZoomifyImage/ZoomifyBase.py", line 199,
in processImage
            imageRow = image.crop([0, ul_y, self.originalWidth, lr_y])
          File
"/home/lavers2/python/lib/python2.4/site-packages/PIL/Image.py",
line 673, in crop
            self.load()
          File
"/home/lavers2/python/lib/python2.4/site-packages/PIL/ImageFile.py", line
155, in load
            self.load_prepare()
          File
"/home/lavers2/python/lib/python2.4/site-packages/PIL/ImageFile.py", line
221, in load_prepare
            self.im = Image.core.new(self.mode, self.size)
        MemoryError

This is on a linux box with 2Gb of RAM, 3Gb of swap space, using PIL 1.1.4.

As I understand it, the load() method loads the entire image into memory,
and although there appears to be some ways to stream image data through
PIL for processing, it looks like you can't use this type of approach with
the crop() method. Am I wrong about this, or is there another way to
accomplish the same type of thing (saving a region of an image as an
image) without using crop()?

Thanks in advance for any guidance you can give.

____________________________
adam smith
ajs17 at cornell.edu
255-8893
215 ccc




From fredrik at pythonware.com  Wed Aug 31 17:47:04 2005
From: fredrik at pythonware.com (Fredrik Lundh)
Date: Wed, 31 Aug 2005 17:47:04 +0200
Subject: [Image-SIG] reaching memory limits with crop method
References: <14810.128.253.230.141.1125498031.squirrel@128.253.230.141>
Message-ID: <df4jdq$iia$1@sea.gmane.org>

Adam J Smith wrote:

> We are encountering memory errors while processing very large images (over
> 3Gb TIFFs, 17888x45408 px). We are trying to make "tiles" of the image by
> calling PIL's crop method and saving the resulting images.

if you open such a file in PIL, what does the "tile" attribute contain?
can you try the following on your files and post the result?

    >>> im = Image.open("verylargetiff.tif")
    >>> im.mode
    >>> im.size
    >>> im.tile

</F>




From kevin at cazabon.com  Sat Aug 27 12:45:20 2005
From: kevin at cazabon.com (kevin@cazabon.com)
Date: Sat, 27 Aug 2005 12:45:20 +0200
Subject: [Image-SIG] PIL and py2exe
References: <003101c5a271$ae29db30$0d00a8c0@aziraphale>
Message-ID: <014401c5aaf4$6c15b030$650aa8c0@duallie>

When you use py2exe, you can specifically exclude modules - in your setup.cfg file just add a line that says:

excludes=Tkinter, (etc.)

For forcing py2exe to import the format modules, just explicitly import them in your script, or once again, put an "includes" line in the seutp.cfg file for py2exe.


Kevin.
  ----- Original Message ----- 
  From: Sean Curtis 
  To: image-sig at python.org 
  Sent: Tuesday, August 16, 2005 4:49 PM
  Subject: [Image-SIG] PIL and py2exe


  the initialization of PIL does two things that are banging heads with py2exe:

      1) It always imported Tk "just in case".  Even if I'm not using Tk, Tk is imported.  This causes py2exe to load everything in Tk causing my install to get ridiculously huge.  Is there any way around this?

      2) The file format modules are imported on the sly using the __import__ function and constructing the module names at runtime (or at least at compile time.)  Because of this, py2exe doesn't recognize module dependencies and I get a distribution with no file plugins (and thus, the ability to read NOTHING.)

      My gut feeling is that the answer for 1 is what's being done in 2 and the answer to 2 is what's being done in 1.  Is there any around these things other than my hacking my own install of PIL?

  Thanks,

  Sean Curtis


------------------------------------------------------------------------------


  _______________________________________________
  Image-SIG maillist  -  Image-SIG at python.org
  http://mail.python.org/mailman/listinfo/image-sig
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20050827/e91e6fcf/attachment.html

From kevin at cazabon.com  Sat Aug 27 12:46:47 2005
From: kevin at cazabon.com (kevin@cazabon.com)
Date: Sat, 27 Aug 2005 12:46:47 +0200
Subject: [Image-SIG] PIL 1.1.5 - Loss of resolution
References: <0MKz5u-1Dn1Cq3ieD-00071J@mrelay.perfora.net>
Message-ID: <015c01c5aaf4$9fe27dd0$650aa8c0@duallie>

When you rotate an image, you're not modifying the image in place - the result is a new image, which unfortunately doesn't carry the format data from the original.

You can simply copy the data over yourself from the original, or supply the dpi setting to the TIFF module when saving (it's supported now).

Kevin.
  ----- Original Message ----- 
  From: Tim Ake 
  To: image-sig at python.org 
  Sent: Monday, June 27, 2005 11:31 PM
  Subject: [Image-SIG] PIL 1.1.5 - Loss of resolution


  Hi,
  I am using PIL 1.1.5 with Python 2.4, and am seeing a loss of resolution when dealing with images. I open an image that is 200 dpi, do a rotate, and save the rotated image. What I'm seeing is that the rotated image is saved at the same size, but is at 100 dpi instead of 200 dpi. My code frament is as follows:

      # open the tif file image
      im = Image.open(strFile)
      # rotate the image 90 degrees CCW
      om = im.rotate(90)
      # save the image as a.tif
      om.save("a.tif", "TIFF")    #save the rotated file

  I can't afford a loss of resolution - any thoughts on maintaining my original resolution?

  Thanks!

  Tim Ake 
  335 Old Hickory Rd. 
  Woodstock, GA 30188 
  mailto:tim at timake.com 
  http://www.timake.com 
  (770) 928-2433 (phone) 
  (404) 944-4890 (cell) 
   soli deo gloria 

    



------------------------------------------------------------------------------


  _______________________________________________
  Image-SIG maillist  -  Image-SIG at python.org
  http://mail.python.org/mailman/listinfo/image-sig
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20050827/5db117a2/attachment.htm

From christianj at implicitnetworks.com  Mon Aug 29 07:01:14 2005
From: christianj at implicitnetworks.com (Christian M. Jensen)
Date: Sun, 28 Aug 2005 22:01:14 -0700
Subject: [Image-SIG] Dithering
Message-ID: <326A1C8DE34E8E4189BE602CE79167843D5E@corporate.implicit.implicitnetworks.com>

If I were to build a 32 gray level palette, still composed of 256 colors, just having duplicates, 000011111111222222222 etc. What would the code look like - general example is fine, I am pretty adept. I am just looking for the recommended method to use.
 
Please include how to build and apply the palette.
 
Thanks!
 

________________________________

From: image-sig-bounces at python.org on behalf of Fredrik Lundh
Sent: Sun 8/28/2005 1:27 PM
To: image-sig at python.org
Subject: Re: [Image-SIG] Dithering



Christian M. Jensen wrote:

> I need to take a 256 level gray image down to 32 levels with nice
> dithering. What is the best process for doing that?

unless I'm missing something, I don't think you can do that in 1.1.5
(unless you want to do the dithering by hand).

the core library can do dithering both to fixed palettes and optimal
palettes, but PIL's convert method only exposes the "dither to web
palette" and "map to nearest N color" variants.

(if you want to do manual dithering, use quantize(32) on the source
image to get a reference image, take the difference between them,
and propagate the errors in some suitable way)

</F>



_______________________________________________
Image-SIG maillist  -  Image-SIG at python.org
http://mail.python.org/mailman/listinfo/image-sig


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20050828/76bdd1a8/attachment.html

From nickg at modp.com  Wed Aug 31 17:11:45 2005
From: nickg at modp.com (nickg)
Date: Wed, 31 Aug 2005 11:11:45 -0400
Subject: [Image-SIG] On Controlling TrueType Font Antialiasing
Message-ID: <31AE032A-4C34-41B4-98D3-0257FEB53F98@modp.com>

Hello all.

I assume this is mailing list for the python imaging library: http:// 
www.pythonware.com/products/pil/index.htm

If so, let me say that I love PIL!

But, I bumped into a problem. For an RGB image I wanted to use a  
small TrueType  "pixel" font without antialiasing.  (for instance  
Courier at 8 pt) I didn't see anything in the documentation on how to  
control this.

However I found that setting "fontmode" on an ImageDraw instance  
seems to work

The by commented out the draw.fontmode="1" one will get antialiasing  
or not.

import sys
import Image, ImageFont, ImageDraw
font = ImageFont.truetype("/System/Library/Fonts/Monaco.dfont", 8)
im = Image.new("RGB", (100, 100), "White")
draw = ImageDraw.Draw(im)
#
draw.fontmode = "1"
#
draw.text((10,10), "hello", fill="Black", font=font)
im.save(sys.stdout, "png")

I assume this is a hack.  It would be great if "fontmode" were  
formally exposed in ImageDraw.text as an option such as  
"antialias=true/false"

I'm happy to work on a patch if so desired.

thanks all,

--nickg


Here's the relevant code trace:

ImageDraw: ctor:  sets fontmode

   def __init__(self, im, mode=None):
         im.load()
         if im.readonly:
             im._copy() # make it writable
         blend = 0
         if mode is None:
             mode = im.mode
         if mode != im.mode:
             if mode == "RGBA" and im.mode == "RGB":
                 blend = 1
             else:
                 raise ValueError("mode mismatch")
         if mode == "P":
             self.palette = im.palette
         else:

             self.palette = None
         self.im = im.im
         self.draw = Image.core.draw(self.im, blend)
         self.mode = mode
         if mode in ("I", "F"):
             self.ink = self.draw.draw_ink(1, mode)
         else:
             self.ink = self.draw.draw_ink(-1, mode)
         if mode in ("1", "P", "I", "F"):
             # FIXME: fix Fill2 to properly support matte for I+F images
             self.fontmode = "1"
         else:
             self.fontmode = "L" # aliasing is okay for other modes
         self.fill = 0
         self.font = None


ImageDraw.text:
    passes fontmode to ImageFont.getmask

     def text(self, xy, text, fill=None, font=None, anchor=None):
         ink, fill = self._getink(fill)
         if font is None:
             font = self.getfont()
         if ink is None:
             ink = fill
         if ink is not None:
             try:
                 mask = font.getmask(text, self.fontmode)
             except TypeError:
                 mask = font.getmask(text)
             self.draw.draw_bitmap(xy, mask, ink)


ImageFont.getmask

calls fontrender with mode == "1", i.e. mode==1 -> 1, mode != 1 -> 0

     def getmask(self, text, mode="", fill=Image.core.fill):
         size = self.font.getsize(text)
         im = fill("L", size, 0)
         self.font.render(text, im.id, mode=="1")
         return im


For Freetype fonts, this goes to _imagingft.c, font_render and here

mask = 1, no antialias
mask = 0, antialias

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/image-sig/attachments/20050831/b50c35ff/attachment.htm