On 3/14/2022 9:40 PM, Dale wrote:

>
> Thank You !!!!
>
> I didn't want to buy Photoshop.
>
> I might install GIMP again ...
>

This is an analysis of the PDF file. Using this package for Windows.

https://www.mupdf.com/releases/index.html

mupdf-1.19.0-windows.zip # The Tesseract version includes OCR.

The first command, removes more compression than necessary, and does not
expose the secrets of Adobe well. But this command still comes in handy.
Even this command does not remove all the compression inside some documents,
but it's a start.

C:\TEMP> mutool convert -F pdf -O decompress,clean -o output.pdf DaleRKelly122717.pdf

This command converts some of the structure of the PDF, into text that Notepad
can read. We can use this one for a look at the structure of the PDF.

C:\TEMP> mutool convert -F pdf -O clean -o output.pdf DaleRKelly122717.pdf

Next, change the file extension on the file, by adding .txt to the end with
the File Explorer rename capability. It helps if the OS and its File Explorer
are set to display the file extension (PDF in this case) so that in the
GUI you can see what is going on.

ren output.pdf output.pdf.txt

notepad output.pdf.txt

You can tack extensions on like that, and it's a way of keeping
track what transformations you have performed.

******

Object number 3 is a raw picture.
Object number 7 is a raw picture.

The pictures need a transform matrix, to position
the picture on the page properly.

When you ask a tool to extract the picture, it calls
the pictures image-0005 and image-0009 because at that
point, they have been instantiated on the paper surface.

And what we came here for, was to verify it uses CCITTFaxDecode as
the compressor. It's their CCITT4 compressor of long ago. All I
wanted to see was "CCITT" to verify a quality compression.

3 0 obj <=== this object stores a picture, byte length 102826
<</DecodeParms<</Columns 5088/Rows 6688/K -1/EndOfBlock true>>/Type/XObject/Subtype/Image/Filter/
CCITTFaxDecode/BitsPerComponent 1/Width 5088/Height 6688/ColorSpace/DeviceGray/Length 102826>>
stream ...

5 0 obj <=== image-0005
<</Length 61>>
stream
1 0 0 -1 0 802.56 cm <=== some sort of transform matrix
-610.56 0 0 802.56 610.56 0 cm
/Img3 Do

endstream
endobj

7 0 obj <=== this object stores a picture, byte length 62500
<</DecodeParms<</Columns 5090/Rows 6672/K -1/EndOfBlock true>>/Type/XObject/Subtype/Image/Filter/
CCITTFaxDecode/BitsPerComponent 1/Width 5090/Height 6672/ColorSpace/DeviceGray/Length 62500>>
stream

9 0 obj <=== image-0009
<</Length 59>>
stream
1 0 0 -1 0 800.64 cm <=== some sort of transform matrix
-610.8 0 0 800.64 610.8 0 cm
/Img7 Do

endstream
endobj

Raw PDF is usually a "binary-looking" format. Using mutool converts
portions of it to readable text, so you can see the commands. PDF files
are hard to edit manually (for a human), because like a banker, the length
of every structure is recorded in a ledger. You can't change the length of
objects, without upsetting some ledger entry.

Paul

Mayayana <mayayana@invalid.nospam> wrote:
> "RJH" <patchmoney@gmx.com> wrote
>
> | > I am working on a second doctorate, Doctor of Divinity in Spiritual
> | > Counseling D.D.
> |
> | Good grief.
> |
> On the other hand, why should any school be able to
> offer a degree in spiritual counseling? What if he got it
> from Harvard? (They offer such degrees. I know people
> personally who have got similar degrees from Harvard
> and Yale.) Would you want a 35-year-old academic
> advising you on life decisions, just because they bought
> the most expensive level of official authorization by reading
> lots of books and theories?

Well, yes. It's called an education. I'd certainly hope and expect that my
GP has read lots of books and be taught theories.

> Then there's the for-profit university trend. You can get
> a degree online from an accredited school if you have a HS
> diploma or GED.

During the pandemic most degrees were run online. Nothing wrong with that
as long as courses are properly run and accredited. Not by some made up
organisations that don't comply with minimum standards.

On 3/15/2022 3:02 AM, Andy Burns wrote:
> Paul wrote:
>
>> CCITT4 is two-dimensional compression.
>
> CCITT T.4 (as used in group 3 fax) is just one dimensional,
> CCITT T.6 (as used in group 4 fax) is two dimensional ...

What I found interesting, is 7ZIP ultra compression could not
match or beat the two dimensional compression, so there is
some benefit from a compression point of view to being 2D.
I thought that maybe the arithmetic encoder could beat
the older compressor, but it didn't quite make it.

My copy of Adobe Distiller 4 for Windows had the CCITT option in it.
And that allowed at least one scanned document (scanned and directed
to PDF) to hit about 50KB per page or so (limited to black and white).

Paul

On 12/03/2022 03:47, Nobody wrote:
> On Fri, 11 Mar 2022 21:21:22 -0500, Dale <dale@dalekelly.org> wrote:
>
>> combine two jpegs?
>>
>> Windows default program? Photos or Paint?
>>
>> Adobe Creative Cloud Express leaves messy text. The jpegs are from a PDF
>> exported from Adobe Acrobat Pro DC
>>
>> PDF is fine, and look fine in Adobe Acrobat Pro DC
>>
>> Photoshop Express or plain Photoshop? Creative Cloud?
>
> Or "stitch" using Irfanview...

There also used to be a program called ICE, MS's own Image Composite
Editor. However I see:

"Please note that the Image Composite Editor download is no longer
available. This project has been retired."

Leave your Hallmark cards - on the passing of ICE - here:

https://www.microsoft.com/en-us/research/product/computational-photography-applications/image-composite-editor/

On 3/13/2022 3:37 PM, Dale wrote:
> On 3/13/2022 1:46 PM, Chris wrote:
>> J. P. Gilliver (John) <G6JPG@255soft.uk> wrote:
>>> On Sun, 13 Mar 2022 at 03:35:37, Dale <dale@dalekelly.org> wrote (my
>>> responses usually FOLLOW):
>>> []
>>>> Used Adobe Creative Cloud Express to turn PDF's into jpegs
>>>> one example ...
>>>>
>>>> https://www.dalekelly.org/images/DaleRKelly122717.pdf
>>>> https://www.dalekelly.org/UMS1.jpg
>>>> https://www.dalekelly.org/UMS2.jpg
>>>>
>>> There's something _very_ odd about those two .jpg files - or the server.
>>> I can _view_ them by entering the URL into Chrome [Version 92.0.4515.159
>>> (Official Build) (32-bit)], but as soon as I, say, try to copy them (to
>>> paste into IrfanView or similar to examine them), the tab crashes.
>>> Repeatably.
>>>>
>>>>
>>>> coded HTML to combine jpegs, one example ...
>>>>
>>>> https://www.dalekelly.org/UMS.html
>>> []
>>> Again, I can _open_ that in Chrome, when I see the menu and top of the
>>> first image, but as soon as I press page down (for some reason I don't
>>> get a scroll bar), the tab crashes.
>>
>> What's most odd is the contents of those images. Any institution that
>> gives
>> out degrees (incl. Masters and PhDs) in only 4 months is bullshit.
>>
>
> the original transcripts were PDF format
>
> https://www.dalekelly.org/images/DaleRKelly122717.pdf
>
> you get about 90 courses
>
> each course has around 25-to-75  8.5x11 pages
>
> many of them have 15-30 minute guided meditations
>
> the designs of the courses are excellent
>
> you can repeat the courses as many times as you need to pass
>
> great support along the way
>
> up to a first doctorate only costs $900 now
>
> https://metaphysicsuniversity.com/
>
> added tremendously to my spiritual progression
>
>
>

here are some other things

you get to keep the *.pdf courses and *.mp3 guided meditations that go
along with the courses

you don't have to go for the degrees online, you can get the courses
printed and the meditations on CDs

I decided to buy the offline courses too

and to give it more credit, the curriculum is tremendous

https://metaphysicsuniversity.com/ums-curriculum/

--
Mystery? -> https://www.dalekelly.org/

On 14 Mar 2022 at 13:07:45 GMT, ""Mayayana"" <mayayana@invalid.nospam> wrote:

> "RJH" <patchmoney@gmx.com> wrote
>
> | > I am working on a second doctorate, Doctor of Divinity in Spiritual
> | > Counseling D.D.
> |
> | Good grief.
> |
> On the other hand, why should any school be able to
> offer a degree in spiritual counseling? What if he got it
> from Harvard? (They offer such degrees. I know people
> personally who have got similar degrees from Harvard
> and Yale.)

My comment was more in the spirit of 'more of the same type of learning',
rather than the subject being covered. So long as it's bound by a level of
rigour and properly overseen, pretty much no topic is out of bounds so far as
I'm concerned.

snip

>
> So... who knows... maybe Dale has a big heart. At least
> he won't be running around telling people they have "ADHD"
> and need to be taking speed. :)
>

Well of course, and least they're not hurting anyone :-)

--
Cheers, Rob

"Paul" wrote:
> On 3/14/2022 9:57 AM, Herbert Kleebauer wrote:
>
>> You should not recompress the pictures as jpeg (a bad format for
>> text) but just store them as gif or png. This retains the original
>> quality of the pdf and the size of both pictures together shouldn't be
>> more than the size of the pdf (165 kB).
>
> We could make a contest out of it :-)

[...]
> Gimp 2.6 Open UMS1.tif
> Save As UMS1a.tif, CCITT4 113,414 bytes
> # Gimp would not offer CCITT4 unless you
> # open a B/W TIF and then just try to save it.

Open the jpegs in Irfanview and save as TIFs, checking "CCITTT Fax 4"
in the save options. That converts to B/W with a bit depth of 1 and
with no loss of info. It's then the same as specified in the PDF:

"/BitsPerComponent 1" and "/K -1"

K less than 0 means "Pure two-dimensional encoding (Group 4)"

UMS1.tif - 102,842 bytes
UMS2.tif - 62,492 bytes

This is the best compression with full size images. I don't think
browsers handle TIFs, so convert to PNG to display in a web page. The
byte count will more than double (more for gif). Better to resize in
Irfanview which will increase the bit depth to preserve appearance.
The byte count will then be about what I said in my earlier post
(resize to 20% gives 112 KB for UMS1).

On 15.03.2022 01:19, Paul wrote:
> On 3/14/2022 9:57 AM, Herbert Kleebauer wrote:
>
>> You should not recompress the pictures as jpeg (a bad format for
>> text) but just store them as gif or png. This retains the original
>> quality of the pdf and the size of both pictures together shouldn't be
>> more than the size of the pdf (165 kB).
>
> We could make a contest out of it :-)

Tried to extract the images without using an external tool
to extract images from pdf files.

1. open the pdf file with Acrobat Reader.

2. Print the 2 pages using the Windows provided printer
"Microsoft XPS Document Writer"
in properties selcet paper size A2 and PNG lossless compression
select custom scale 200%
select "print to file"
print using the file name a.zip

3. Use Windows Explorer to copy the the 32 tif files
a.zip\Documents\1\Resources\Images\ into an empty directory

4. Now you need a program to concatenate the 32 pieces to one
picture. I used convert.exe from ImageMagick:
https://download.imagemagick.org/ImageMagick/download/binaries/ImageMagick-7.1.0-portable-Q16-x64.zip
This is a 150 MB zip file, but all you need is the file convert.exe
stored in the zip file. Copy convert.exe into the directory with
the 32 tif files.

5. Save this batch code as a.bat in the same directory and start it:

@echo off
for /l %%i in (1,1,32) do call set x=%%x%% %%i.tif
..\convert.exe -append -depth 1 -compress group4 %x% all.tif

As a result you will get the file all.tif (162 kB, 4962x13392 pixel)
which contains both pages in a single picture (as the OP wanted) and
it is not larger than the original pdf.

But for sure, a tool to directly extract the images from the
pdf would be preferable. convert.exe should be able to do this
if GhostScript is installed.

On 15.03.2022 13:36, Apd wrote:
> "Paul" wrote:
>> On 3/14/2022 9:57 AM, Herbert Kleebauer wrote:
>>
>>> You should not recompress the pictures as jpeg (a bad format for
>>> text) but just store them as gif or png. This retains the original
>>> quality of the pdf and the size of both pictures together shouldn't be
>>> more than the size of the pdf (165 kB).
>>
>> We could make a contest out of it :-)
>
> [...]
>> Gimp 2.6 Open UMS1.tif
>> Save As UMS1a.tif, CCITT4 113,414 bytes
>> # Gimp would not offer CCITT4 unless you
>> # open a B/W TIF and then just try to save it.
>
> Open the jpegs in Irfanview and save as TIFs, checking "CCITTT Fax 4"
> in the save options. That converts to B/W with a bit depth of 1 and
> with no loss of info. It's then the same as specified in the PDF:

I think it is a real bad idea to start with a jpg (with all the
compression artifacts) and then save it in a lossless format.
When you extract the pictures you have to directly save it a
lossless format (gif, png, tiff).

"Herbert Kleebauer" wrote:
> On 15.03.2022 13:36, Apd wrote:
>> Open the jpegs in Irfanview and save as TIFs, checking "CCITT Fax 4"
>> in the save options. That converts to B/W with a bit depth of 1 and
>> with no loss of info. It's then the same as specified in the PDF:
>
> I think it is a real bad idea to start with a jpg (with all the
> compression artifacts) and then save it in a lossless format.
> When you extract the pictures you have to directly save it a
> lossless format (gif, png, tiff).

I would agree but in this case it makes no difference. The images are
so large (in pixels) that all it does is remove the compression
artifacts. You can also extract direct from the PDF using the free
version of PDF-XChange Viewer by doing "Export to image" and setting
the appropriate options for the format. It has all the necessary
settings for TIFF, PNG and others. Set the page zoom and/or DPI
resolution to get an appropriate size.

On 3/15/2022 11:56 AM, Apd wrote:
> "Herbert Kleebauer" wrote:
>> On 15.03.2022 13:36, Apd wrote:
>>> Open the jpegs in Irfanview and save as TIFs, checking "CCITT Fax 4"
>>> in the save options. That converts to B/W with a bit depth of 1 and
>>> with no loss of info. It's then the same as specified in the PDF:
>>
>> I think it is a real bad idea to start with a jpg (with all the
>> compression artifacts) and then save it in a lossless format.
>> When you extract the pictures you have to directly save it a
>> lossless format (gif, png, tiff).
>
> I would agree but in this case it makes no difference. The images are
> so large (in pixels) that all it does is remove the compression
> artifacts. You can also extract direct from the PDF using the free
> version of PDF-XChange Viewer by doing "Export to image" and setting
> the appropriate options for the format. It has all the necessary
> settings for TIFF, PNG and others. Set the page zoom and/or DPI
> resolution to get an appropriate size.
>
>

I think I can see in the PDF object definition, why an extractor
could behave in a non-intuitive way.

BitsPerComponent 1
ColorSpace/DeviceGray

That's how the JPG ended up with an 8 bit indexed palette, is the
DeviceGray specification. Even though we know from the BPP spec,
that a PBM output file would be more appropriate.

Perhaps if the Adobe tool had used a different "ColorSpace"
that would not have happened.

I didn't even know you could declare an 8-bit palette in a JPG.
I always thought they were 24-bit (or more). Especially given
the propensity for numerical roundoff "spraying" the palette
with fake colours.

*******

One problem with the PDF file, is the image consists of two
objects. The first object is the bitmap/pixmap, stripped of
most of the original filetype info, and in a sense "anonymized".

This makes it pretty hard for an extractor, to do something
sensible, especially if the declarations aren't particularly
hinting strongly in a consistent direction.

A human, upon seeing BitsPerComponent 1, wouldn't be quite as
quick to jump on JPG for that. While PNG is a natural match
(supports 1BPP with your choice of colour), I had a lot of
trouble getting the PNG tools to do that for me. I've done
it before, but I don't remember which PNG tool I used
for the job (things with names like PNGcrush and PNGreduce
and so on, there are a few of them to try).

I think if I was writing the code, I'd do an analysis from
first principles, and derate the header declaration. If the
image came with a palette declared as a discrete table
of colour values, I would use it. That way, I could have a
1BPP PNG where a value of 1 could have a red pixel and a value
of 0 could have a white pixel. But only if a palette was
part of the pixmap info.

Paul

On 12/03/2022 21:28, Bill W wrote:
> On Mar 12, 2022, nospam wrote
> (in article<120320221222134794%nospam@nospam.invalid>):
>
>> In article<t0iko6$8f2$1@dont-email.me>, Dale<dale@dalekelly.org>
>> wrote:
>>
>>>
>>> side by side
>>>
>>> or top and bottom
>>>
>>> top and bottom would be better for printing
>>
>> earlier, you said compositing:
>> In article <t0ij0b$pnr$1@dont-email.me>, Dale<dale@dalekelly.org>
>> wrote:
>>> On 3/12/2022 10:31 AM, Savageduck wrote:
>>>> Do you mean compositing?
>>>
>>> yes
>>
>> now you say tiling.
>>
>> as usual, this is nothing more than a troll.
>
> I’m pretty sure it’s not him, but there is something oh so Arlenesque
> about all of this guy’s posts.
>

No there isn't. 'Arlene' posts faux 'research' links, and his questions
are complicated with pretentions to organise topics. Dale has a simple
question - he wants an image 'stiching' program, but maybe he does not
understand the terminology (tiling, compositing etc) very well. If you
don't know of a program which helps him, just look away rather than
throw accusations of trolling.

"Paul" wrote:
> On 3/15/2022 11:56 AM, Apd wrote:
>> "Herbert Kleebauer" wrote:
>>> I think it is a real bad idea to start with a jpg (with all the
>>> compression artifacts) and then save it in a lossless format.
>>> When you extract the pictures you have to directly save it a
>>> lossless format (gif, png, tiff).
>>
>> I would agree but in this case it makes no difference. The images are
>> so large (in pixels) that all it does is remove the compression
>> artifacts. You can also extract direct from the PDF using the free
>> version of PDF-XChange Viewer by doing "Export to image" and setting
>> the appropriate options for the format. It has all the necessary
>> settings for TIFF, PNG and others. Set the page zoom and/or DPI
>> resolution to get an appropriate size.
>
> I think I can see in the PDF object definition, why an extractor
> could behave in a non-intuitive way.

As I understand it, the jpegs were created from the PDFs so that
format would have been a user choice.

> BitsPerComponent 1
> ColorSpace/DeviceGray
>
> That's how the JPG ended up with an 8 bit indexed palette, is the
> DeviceGray specification. Even though we know from the BPP spec,
> that a PBM output file would be more appropriate.

Except PBM is not compressed. Given the encoding, TIFF would be
better as a default (but not suitable for web).

> Perhaps if the Adobe tool had used a different "ColorSpace"
> that would not have happened.

DeviceGray is stated to be on a scale from black to white but there is
no mention of how many intermediate values there might be, so "none"
must be valid.

> I didn't even know you could declare an 8-bit palette in a JPG.
> I always thought they were 24-bit (or more).

Me neither. It must be the minimum possible for jpeg.

> One problem with the PDF file, is the image consists of two
> objects. The first object is the bitmap/pixmap, stripped of
> most of the original filetype info, and in a sense "anonymized".
>
> This makes it pretty hard for an extractor, to do something
> sensible, especially if the declarations aren't particularly
> hinting strongly in a consistent direction.

The PDF spec should make it clear enough:

| BitsPerComponent
| | If the image stream uses a filter, the value of BitsPerComponent
| shall be consistent with the size of the data samples that the
| filter delivers. In particular, a CCITTFaxDecode or JBIG2Decode
| filter shall always deliver 1-bit samples

So however the colour space is declared, if I'm reading that right,
the bit-depth is fixed for fax decode.

> A human, upon seeing BitsPerComponent 1, wouldn't be quite as
> quick to jump on JPG for that. While PNG is a natural match
> (supports 1BPP with your choice of colour), I had a lot of
> trouble getting the PNG tools to do that for me. I've done
> it before, but I don't remember which PNG tool I used
> for the job (things with names like PNGcrush and PNGreduce
> and so on, there are a few of them to try).

Both Irfanview and PDF-XChange Viewer can save as 1-BPP PNGs.

> I think if I was writing the code, I'd do an analysis from
> first principles, and derate the header declaration. If the
> image came with a palette declared as a discrete table
> of colour values, I would use it. That way, I could have a
> 1BPP PNG where a value of 1 could have a red pixel and a value
> of 0 could have a white pixel. But only if a palette was
> part of the pixmap info.

AS an experiment, I created a PDF with PDF-XChange Viewer using a
CCITT 4 encoded TIFF (which it encoded the same way). This was how it
specified the colour space:

/ColorSpace [/Indexed /DeviceRGB 1 (-values-)]

Where -values- is the two 3-byte RGB colors in binary (0xFFFFFF, 0).

In Irfanview you can edit the palette so I was also able to create
another CCITT 4 TIFF with red text on white background. For some
unknown reason, when I turned that into a PDF, it was encoded the
same but the colours were swapped to white on red.

On 3/11/2022 9:21 PM, Dale wrote:
> combine two jpegs?
>
> Windows default program? Photos or Paint?
>
> Adobe Creative Cloud Express leaves messy text. The jpegs are from a PDF
> exported from Adobe Acrobat Pro DC
>
> PDF is fine, and look fine in Adobe Acrobat Pro DC
>
> Photoshop Express or plain Photoshop? Creative Cloud?
>

okay, so I forgot somethings and wasn't all that familiar with specific
definitions ...

I used to tile images in Photoshop for bitmap output devices like film
recorders. Finding the bytes of the original images, create a big enough
new image for both, then copy/paste the original images into the new
one. All images were open.

I used composite images in Photoshop to form new layers. Like the K
layer to CMY that was converted from RGB. K and a new CMY were derived
by some means like UCR, GCR, etc.

I used to do "page layout" like Quark Express for output to proof, etc.
utilizing a RIP. This was for CMYK rosettes patterns some times.

https://en.wikipedia.org/wiki/Raster_image_processor

I started page layout in HTML about 30 years ago.

The reason I wanted to combine (tile) the two images was to lay them out
in HTML as one image. Decided to lay the two images out in HTML without
tiling them. Don't know the difference on download speed of the two lay
outs. Good enough for now.

I don't do much image manipulation any more. So the free recommendations
might fit.

I was kind of thinking about the manipulation softwares that come with
Windows. "Photos" and "Paint".

I like GIMP on Linux better than on Windows.

--
Mystery? -> https://www.dalekelly.org/