JPEG to PDF... lost of qualty
T. Parker
I'm trying to put jpeg files into PDF. There is a lost
of quality. Why and how to retain the original
jpeg quality when seen inside PDF?
Thanks.
Parker
Bigguy
Set for high quality images (Press quality). This makes the .pdf files
much larger of course.
Guy
T. Parker
I'm using many other software to convert jpeg to pdf. When
zoomed, the converted jpeg always has poorer quality in
the texts for example. Could it be that whenever jpeg is
converted to pdf. There is automatically a slight loss of
quality because of the conversion algorithm? When
you mentioned Adobe Acrobat? You mean the Adobe
Acrobat Professional.. not just Reader? It can convert
a batch of jpegs into one pdf??
Parker
T. Parker
I ran Adobe Acrobat Professional 7.0. I went to the Create PDF
from File and chose the jpeg I got from google books. The
PDF file converted still has loss of quality compared to the
original jpeg. There seems to be no settings to adjust it.
I tried PDF Optimizer but it doesn't improve the quality. Is
it really a fact that any conversion from jpeg to pdf would
automatically entail loss of quality without anything that
can be done?
Parker
T. Parker
Oh I figured it out. The Adobe reads the jpeg as original
quality. I was trying this software IrfanView in which when
you try to zoom the image, there is no pixelization,
the resolution keeps getting bigger. Do you know what
the process is called? Know how I can integrate it in
Adobe so that when I tried to zoom in, instead of
pixalization of the original image, the software would
fill it producing solid text files (I got from goggle books)??
Thanks.
Parker
bugbear
> I was trying this software IrfanView in which when
> you try to zoom the image, there is no pixelization,
> the resolution keeps getting bigger. Do you know what
> the process is called?
If it's as you report
it's called "a miracle"
BugBear
T. Parker
Here's a screen capture of the difference.
The following is from original jpeg zoomed to 1600% in
Adobe Reader or Photoshop.
http://www.pbase.com/image/110861624/original
The following is from the program IrFanView.
http://www.pbase.com/image/110861650/original
The image is much smoother here. What feature is that
called. Aliasing? Anyone?
Also in movies, there is always shown a photo not clear
and they did something that can bring out the image of
the person. Maybe it has the same feature as the above
Irfanview stunt?
Parker
Mark Thomas
Interpolation.
The ragged sample you posted has obviously used a very low quality
algorithm, like linear or nearest neighbour - how did you enlarge it
exactly?.. Whereas Irfanview used something like Lanczos (I'm
guessing?) which is a very intelligent interpolation algorithm.
There is no free lunch with enlarging, all you can really do is smooth
the image and guess the intervening pixels. There are some systems that
guess quite intelligently.. You can also correct an unfocussed image
*slightly* (see 'deconvolution'), but there is no magic of the kind you
see in CSI where they magnify a ratty video-grab twenty times and can
suddenly read that tiny number plate....
If the detail wasn't there in the original (or at least a strong hint of
it), it won't be in the enlargement either.
There are some interesting comparisons of enlarging algorithms here:
http://www.general-cathexis.com/interpolation/index.html
(takes a while to fully load)
T. Parker
The ragged sample was enlarged using Adobe Reader,
even Adobe Acrobat doesn't show any improvement,
nor Photoshop, Windows Photo Gallery, etc. Only
IrFanView shows it much better. What other viewers
use Lanczos? How come Adobe Reader doesn't use
it when it deails with texts and reading material. The
original jpeg in the pictures came off a page in google
books. Even at slight zooming, the text using IrFanview
is much clearer. Do you know how I can turn off the
feature in IrFanview, the Lanczos thing?
Parker
> There is no free lunch with enlarging, all you can really do is smooth
> the image and guess the intervening pixels. There are some systems that
> guess quite intelligently.. You can also correct an unfocussed image
> *slightly* (see 'deconvolution'), but there is no magic of the kind you
> see in CSI where they magnify a ratty video-grab twenty times and can
> suddenly read that tiny number plate....
>
> If the detail wasn't there in the original (or at least a strong hint of
> it), it won't be in the enlargement either.
>
> There are some interesting comparisons of enlarging algorithms here:
>
> http://www.general-cathexis.com/interpolation/index.html
>
> (takes a while to fully load)- Hide quoted text -
>
> - Show quoted text -
T. Parker
>
> - Show quoted text -
Btw... the following is the original jpeg taken from google
free book. See how clearly you can zoom in the word "simple"
in the middle at 1600% zoom level:
http://www.pbase.com/image/110863838/original
parker
bugbear
> On Apr 2, 7:59 pm, bugbear <bugbear@trim_papermule.co.uk_trim> wrote:
>> T. Parker wrote:
>>> I was trying this software IrfanView in which when
>>> you try to zoom the image, there is no pixelization,
>>> the resolution keeps getting bigger. Do you know what
>>> the process is called?
>> If it's as you report
>> it's called "a miracle"
>>
>> BugBear
>
> Here's a screen capture of the difference.
>
> The following is from original jpeg zoomed to 1600% in
> Adobe Reader or Photoshop.
>
> http://www.pbase.com/image/110861624/original
>
> The following is from the program IrFanView.
>
> http://www.pbase.com/image/110861650/original
>
> The image is much smoother here.
I blurred the Adobe version with a 25 pixel Gaussian
(in Gimp) and it then looked a lot like the irfanview version.
Not identical, but similar.
BugBear
Matt Ion
> Also in movies, there is always shown a photo not clear
> and they did something that can bring out the image of
> the person. Maybe it has the same feature as the above
> Irfanview stunt?
Don't believe for a second what you see in the movies - it's all
trickery, the proverbial smoke-and-mirrors.
I run into this problem regularly in the CCTV/surveillance camera
business - people see all this amazing stuff in CSI and whatnot and want
systems for their homes and small stores that will do the same thing...
and then want to know why the cheap $500 system they bought at Walmart
doesn't measure up.
John McWilliams
>
> Btw... the following is the original jpeg taken from google
> free book. See how clearly you can zoom in the word "simple"
> in the middle at 1600% zoom level:
>
> http://www.pbase.com/image/110863838/original
Both are horrid. You should not be making JPEGs out of text.
--
john mcwilliams
Keith Nuttle
image viewed in an image viewer and an Adobe image to the fact that to
convert an image to Adobe it is "printed" to the Adobe file. So the
Adobe image is dependent on the "print" driver used to make the
conversion.
ray
Please excuse my ignorance, but why would one want to make a PDF from an
image file? Why not just use the image?
nospam
wrote:
> Please excuse my ignorance, but why would one want to make a PDF from an
> image file? Why not just use the image?
one reason would be if they're writing a paper and want to include a
few images.
ray
Then why convert to PDF first? I would think they would simply include
the images in the document.
Ofnuts
The first question is the size in pixels of the pictures, and the actual
print size you are tying to achieve with them...
--
Bertrand
Martin Brown
The text itself is meaningless new age gibberish, but the JPEG encoding
of it isn't all that bad and could be better with optimised quantisation
parameters. JPEG can encode line art surprisingly well if you choose the
encoding quantisation matrix optimally. It is a bit weak with the
default photographic image encoding.
The reconstructed image would be even better if Xerox didn't have an
obstructive patent on the mathematical identity
X + (-X) = 0
The OPs original question is answered by interpolation or anti aliassing
depending on which literature you want to search. It effectively rounds
the corners of smooth curves and keeps text more legible.
But the IrfanView solution isn't optimal if you actually know "a priori"
that your target image is black print on mostly white paper.
Various non-linear methods are capable of obtaining up to 3x super
resolution depending on the signal to noise ratio although most of them
have been optimised for astronomy where the sky is black with a few
bright spots and nebulae.
Regards,
Martin Brown
Matt Clara
news:eb768a51-0724-4946...@b7g2000pre.googlegroups.com...
I'm going to go out on a limb and guess you didn't check the help files even
once.
tomcas
What you are describing sounds like a vector format such as windows
metafile. Being that it's vector based and not raster based there is no
pixelization at high zoom. It's a great format for inserting drawings
into office documents and has a very small file size.
T. Parker
wrote:
> John McWilliams wrote:
> > T. Parker wrote:
>
> >> Btw... the following is the original jpeg taken from google
> >> free book. See how clearly you can zoom in the word "simple"
> >> in the middle at 1600% zoom level:
>
> >>http://www.pbase.com/image/110863838/original
>
> > Both are horrid. You should not be making JPEGs out of text.
>
> The text itself is meaningless new age gibberish, but the JPEG encoding
> of it isn't all that bad and could be better with optimised quantisation
> parameters. JPEG can encode line art surprisingly well if you choose the
> encoding quantisation matrix optimally. It is a bit weak with the
> default photographic image encoding.
The text came from
It's a free book that is written by Nobel Committee member Bjorn
Nordenstrom. All 368 pages are browseable. Check it out
and see if you would call it new age gibberish again. It's
cutting edge research.
>
> The reconstructed image would be even better if Xerox didn't have an
> obstructive patent on the mathematical identity
> X + (-X) = 0
>
> The OPs original question is answered by interpolation or anti aliassing
> depending on which literature you want to search. It effectively rounds
> the corners of smooth curves and keeps text more legible.
>
> But the IrfanView solution isn't optimal if you actually know "a priori"
> that your target image is black print on mostly white paper.
But how come CS Photoshop that costs thousands more can't
duplicate the smoothness produced by shareware IrFanview? All other
software produced the ragged edge text. But then in zooming,
isn't it that you simply magnify the image meaning whatever
is there in the original pixel is simply made bigger? IrFanview
could be doing something else to it extra like smoothing
the ragged edged that others don't. Maybe there is a function
inside Photoshop that can do the trick too? Also the IrFanview
output I shared is saved in jpeg from original screen capture
in bmp so it looks blur but the original capture is clearer.
To other posters questions why I'd love to put them in
jpeg. Well. In the world of e-book. PDF is the gem, so
naturally you want to put hundreds of pages of jpegs in
a neat single PDF file.
Parker
Mark Thomas
Why would you want to?
Adobe reader is a *reader* - it is not really designed for enlarging
stuff on screen, so it uses a crude algorithm. IrfanView, just like
Photoshop or any other decent image *editor* will use more sophisticated
methods to actually enlarge images (although, they may still use crude
methods if you are simply 'zooming', ie not changing the file).
T. Parker
>
> - Show quoted text -
Know how to use Photoshop to do the same thing IrFanview
does? I notice that in zooming using IrFanview, it is slower,
so many hard disc activity like maybe it's changing the file.
I wonder how to initiate it using photoshop to remove
the ragged edged. Anyone?
Parker
Matt Clara
news:5N9Bl.131826$RJ7....@newsfe18.iad...
I open the image and select Print and then choose to print to PDF. Go into
the printer setup, and then Properties, and select Edit to the right of the
Default Settings row, when that dialogue opens, choose Image, and turn off
all compression. It will likely INCREASE the size of your image file.
Chris Malcolm
> But how come CS Photoshop that costs thousands more can't
> duplicate the smoothness produced by shareware IrFanview?
Because in that respect it's inferior to Irfanview. The idea that the
most expensive software is the best is a very profitable idea, which
is why so many salesmen work hard to convince their victims of it. But
in some cases even completely free software can give superior results
to anything you can buy at any price. But for obvious reasons no
salesman will ever admit that :-)
--
Chris Malcolm
T. Parker
> "Matt Clara" <n...@myexpense.com> wrote in message
>
> news:5N9Bl.131826$RJ7....@newsfe18.iad...
>
>
>
>
>
> > "T. Parker" <tomparke...@gmail.com> wrote in message
> >news:eb768a51-0724-4946...@b7g2000pre.googlegroups.com...
> >> Hi,
>
> >> I'm trying to put jpeg files into PDF. There is a lost
> >> of quality. Why and how to retain the original
> >> jpeg quality when seen inside PDF?
>
> >> Thanks.
>
> >> Parker
>
> > I'm going to go out on a limb and guess you didn't check the help files
> > even once.
>
> I open the image and select Print and then choose to print to PDF. Go into
> the printer setup, and then Properties, and select Edit to the right of the
> Default Settings row, when that dialogue opens, choose Image, and turn off
>
> - Show quoted text -
I tried that. But no, it's not that the jpeg is degraded when
converted to
PDF but I'm using Irfanview to view it against adobe reader so I
thought it degraded the file when Irfanview simply enhanced it.
It is noticeable even at small zoom like 125% so I wonder why
adobe reader doesn't integrate the superior interpolation used
by Irfanview when its supposed to be the e-book standard reader.
Parker
John McWilliams
> John McWilliams wrote:
>> T. Parker wrote:
>>>
>>> Btw... the following is the original jpeg taken from google
>>> free book. See how clearly you can zoom in the word "simple"
>>> in the middle at 1600% zoom level:
>>>
>>> http://www.pbase.com/image/110863838/original
>>
>> Both are horrid. You should not be making JPEGs out of text.
>
> The text itself is meaningless new age gibberish, but the JPEG encoding
> of it isn't all that bad and could be better with optimised quantisation
> parameters. JPEG can encode line art surprisingly well if you choose the
> encoding quantisation matrix optimally. It is a bit weak with the
> default photographic image encoding.
No argument about the content of the text, but JPEG izing it in a PDF is
a rotten choice. PDF's can scale (most) text well.
--
john mcwilliams
T. Parker
>
> - Show quoted text -
I just found out that in books.google.com their online text reader
also uses excellent interpolation (zoom it to maximum for better
eye relief). However, when the ebook is read as PDF. It losses
quality and the text becomes more blurred. Since Adobe Reader
is the leader in e-book technology. I wonder why they can't use
the excellent interpolation of IrFanView and Google books,
maybe because the method is patented??
Parker
Ray Fischer
More likely because they don't need to. They don't use compressed
image files for text. They use the actual text. Perfect sharpness at
any size. Embedded images can be added for illustrations and it's
left to the author to decide what size image to is needed.
--
Ray Fischer
rfis...@sonic.net
T. Parker
>
> - Show quoted text -
Hey, that's a brilliant idea. Adobe Reader needing the
original file and not jpegs. This would indeed create
good quality zooms. But if you can only extract jpegs
from google books. You only have the jpegs to start
with. Do you know what other software uses Lanczos
interpolation besides Qimage Pro which doesn't even
display them on screen but just printer wise. I'd like
to try Lanczos on other image reader besides Irfanview
to see if Lanczos is really the factor. How come Photoshop
doesn't have Lanczos but even lowly interpolation algorithm.
Parker
T. Parker
>
> - Show quoted text -
That may be true for Adobe Reader where people import
texts directly from Word or other software that is not jpegs.
What about Photoshop CS4? I got to download the 30
days Trial Version and the interpolation methods are
still Bilinear and Bicubic. How come they can't give
other options like Lanczos when people got the software
to edit pictures??
Parker
Ray Fischer
>original file and not jpegs. This would indeed create
>good quality zooms. But if you can only extract jpegs
>from google books.
Tha's because Google is solving a different problem.
> You only have the jpegs to start
>with. Do you know what other software uses Lanczos
>interpolation besides Qimage Pro which doesn't even
Why are there always some obsessive nutters who are so fanatical about
non problems?
--
Ray Fischer
rfis...@sonic.net
Chris Malcolm
> Hey, that's a brilliant idea. Adobe Reader needing the
> original file and not jpegs. This would indeed create
> good quality zooms. But if you can only extract jpegs
> from google books. You only have the jpegs to start
> with. Do you know what other software uses Lanczos
> interpolation besides Qimage Pro which doesn't even
> display them on screen but just printer wise. I'd like
> to try Lanczos on other image reader besides Irfanview
> to see if Lanczos is really the factor. How come Photoshop
> doesn't have Lanczos but even lowly interpolation algorithm.
Because it doesn't care enough that particular feature, or because
marketing think it's not important enough to pay good enough
programmers to do it well, or maybe just simple ignorance. Like a
Swiss Army knife, Photoshop is very good at being one program that can
do nearly everything you might want to do, but like a Swiss Army
knife, for any one single job there's usually a choice of better
tools.
If you like Lanczos, note that it's not a single algorithm, but a
family, and the more complex ones usually do better but take
longer. Editors often only implement one or two members of the family,
not necessarily the same ones.
--
Chris Malcolm
T. Parker
But Lanczos is clearly more superior than Photoshop Bilinear
or Bicubic Interpolation. Or maybe in some pictures, Binliear,
Bicubic is better than Lanczos? Also maybe Photoshop
doesn't include Lanczos maybe because it belongs to a
rival company or association of rivals of some kind?
Parker
Ray Fischer
Why would Adobe spend hundreds of thousands of dollars for a feature
that nobody cares about?
--
Ray Fischer
rfis...@sonic.net
Chris Malcolm
What has that got to with Lanczos resampling, which is a freely
available published method the virtues and defects of which have been
the topic of much published comparative research by both academic
interpolation experts and practical working photographers?
It's obviously true that most photographers don't care, but it's
equally obvious that some do. That's why there's already so many
different methods avaiable, and why researching better methods
continues.
--
Chris Malcolm
Ray Fischer
>Ray Fischer <rfis...@sonic.net> wrote:
>> T. Parker <tompa...@gmail.com> wrote:
>>>But Lanczos is clearly more superior than Photoshop Bilinear
>>>or Bicubic Interpolation. Or maybe in some pictures, Binliear,
>>>Bicubic is better than Lanczos?
>
>> Why would Adobe spend hundreds of thousands of dollars for a feature
>> that nobody cares about?
>
>What has that got to with Lanczos resampling,
It's a feature that nobody cares about.
> which is a freely
>available published method
Irrelevant. Turning a method into actual software costs money.
Counting design, implementation, and testing (don't forget UI), that's
roughly 1-2 man years. There's $300,000 of expense. Given that
Photoshop costs about $400 (ballpark average of 1st-time sales and
upgrades) that means that the feature has to sell an additional 800
copies of Photoshop.
>the virtues and defects of which have been
Nobody cares.
>It's obviously true that most photographers don't care, but it's
>equally obvious that some do.
No, it isn't obvious. DO you think that there are 800 people out
there will to spend an average of $400 to get that feature?
--
Ray Fischer
rfis...@sonic.net