carl_miller23@hotmail.com (Carl Miller) wrote in message news:<00040823211103.OUI90.carl_mi
ller23@hotmail.com>...
> A straight-forward answer to this might be out there, but I haven't
> found it yet. (All file sizes are approximations, but representative of
> actual experience.)
>
> When saving JPEG's, the Quality drop-down menu gives you the choice of
> Low, Medium, High, Maximum. Choosing Maximum gives a setting of 10.
> However, you can increase the setting to 12. I've noticed that if I open
> a JPEG that is, say, about 1.3mb and save it at the "default" Maximum
> setting of 10, it compresses it to, say, about 900k. If I save it at the
> "maximum" Maximum setting of 12, I actually end up with a larger file
> than I started out with of about 3mb. Apparently it is uncompressing the
> original compressed jpeg?
>
> What's up with this, and practically speaking, should I be saving at 10
> or 12?
>
> (Take it as read that I know about tif being non-lossy, and jpeg being
> lossy, etc. I'm just wondering about this Photoshop jpeg Quality thing.)
>
> Thanks!!

I've had the exact questions. I was hoping for a straight forward
answer, but I must agree with Carl, no one directly answered the
questions; e.g. what about Max = 10 but 12 is available. Why is 12
larger that the original (details)? If someone who really understand
P/Shop JPEG save routing could/would answer it would be appreciated.

FWIW I don't know exactly either

Urban myths are:
resaving JPEG to JPEG at low levels cummulatively reduces (?) image size and
hence quality
JPEG compression algorithms are independent steps
for example
(a)image save at JPEG level 5 then image reopened and saved at JPEG level 12
=> this will not return the image to its original appearance because JPEG is
lossy BUT saving and resaving at JPEG level 12 will not cummulatively reduce
(?) image quality

(b) image saved a JPEG level 5 then image reopned and saved at JPEG level 5
=> a second application of the JPEG compression algorithms => a further loss
in image quality

I suppose (a) and (b) may be tested empirically (I can be bothered at the
mo')

My own preference on working images is to take JPEG to PSD and keep
everytthing in PSD form flattening out into JPEG at level 12 setting until I
am pleased with result.  Mike will probably confirm that I am easily pleased
with results - possibly too easily pleased LOL

Saving to JPEG at anyother setting - I try to do this from the PSD file
rather than JPEG to JPEG (I can't remember why tho :-)

Artie

"Martin" <metpx3c@hotmail.com> wrote in message
news:3a16084b.0409260857.646fd4de@posting.google.com...
> carl_miller23@hotmail.com (Carl Miller) wrote in message
> news:<00040823211103.OUI90.carl_miller23@hotmail.com>... 
>
> I've had the exact questions. I was hoping for a straight forward
> answer, but I must agree with Carl, no one directly answered the
> questions; e.g. what about Max = 10 but 12 is available. Why is 12
> larger that the original (details)? If someone who really understand
> P/Shop JPEG save routing could/would answer it would be appreciated.

"Martin" <metpx3c@hotmail.com> wrote in message news:3a16084b.0409260857.646fd4de@posting.g
oogle.com...
> carl_miller23@hotmail.com (Carl Miller) wrote in message news:<00040823211
103.OUI90.carl_miller23@hotmail.com>... 
>
> I've had the exact questions. I was hoping for a straight forward
> answer, but I must agree with Carl, no one directly answered the
> questions; e.g. what about Max = 10 but 12 is available. Why is 12
> larger that the original (details)? If someone who really understand
> P/Shop JPEG save routing could/would answer it would be appreciated.

It's been explained to you in plain English. What part of this
do you not understand:

The JPEG compression routine doesn't know or care whether
an image has been previously compressed.  In other words,
when you JPEG compress an image, the routine doesn't know
or care what the original file size was, or what quality settings
were used for previous compressions.  At quality 12 you're
getting a less compressed image, therefore larger in size,
although it will still be compressed, and will still be lossy
relative to the original.

Rick

> Why is 12
>larger that the original (details)?

Very, very simple.

You have an image. Let us say for the sake of example that the image, when
UNCOMPRESSED, is one megabyte.

You save it as a JPEG. You use quality, say, 4. It saves to about 200
kilobytes.

You open the JPEG, It uncompresses. The image is 1 megabyte uncompressed.

You save it as JPEG with quality 12. It compresses again. You saved it with 
a
higher quality so this time it compresses to 500 kilobytes.

Make sense?

--
Art, literature, shareware, polyamory, kink, and more:
http://www.xeromag.com/franklin.html

>Why is 12
>larger that the original (details)? If someone who really understand
>P/Shop JPEG save routing could/would answer it would be appreciated.

It's very, very simple, and has already been explained.

Let us say you have an image that is 1 megabyte in size uncompresssed. Let's
say you save it as a JPEG with quality 4. So it saves to a file that's, for
instance, about 150 K in size.

Now you open that 150 K JPEG. It uncompresses. Now it is 1 megabyte again.

Now you save it again. You save as JPEG with quality 12. So it compresses
again, this time to 500 kilobytes.

Make sense?

--
Art, literature, shareware, polyamory, kink, and more:
http://www.xeromag.com/franklin.html

metpx3c@hotmail.com (Martin) writes:

> I've had the exact questions. I was hoping for a straight forward
> answer, but I must agree with Carl, no one directly answered the
> questions; e.g. what about Max = 10 but 12 is available. Why is 12
> larger that the original (details)? If someone who really understand
> P/Shop JPEG save routing could/would answer it would be appreciated.

That second question *was* specifically answered, I thought, but I'll
go through it again.

When you open an image in Photoshop, it reads whatever format the
image file is originally in and decodes it into computer memory where
it's stored as an uncompressed bitmap.  (Yes, if any of the Photoshop
programmers hang out here I know there are additional complexities,
not nearly well enough to explain them, so I'm not trying.)  This is
the *full-sized* bitmap -- if it's a 1600x1200 digital original in
24-bit color, that's a 5,760,000 byte image, even though the jpeg
might have been 512KB.

So, there's this bitmap image in memory.  Big.  You edit it some,
change various bits, whatever.  (Or just leave it alone, not change
any bits).

Now you tell Photoshop to save the image.  It pops up the dialog for
you to select the quality level, and then starts coding the image into
a jpeg using the setting you gave it and writing the result to disk.

If the new quality level you give is the same as was used in the
original jpeg, the resulting file will be about the same size (if you
haven't changed the bitmap extensively).  If you pick a higher quality
level, the resulting file will be larger -- it won't compress it as
much.  If you pick a lower quality level, the resulting file will be
smaller -- it will compress more.

(The photoshop quality levels are a layer above the actual parameters
to the jpeg compression algorithm, by the way; that's why every
program has a different way of expressing degree of jpeg compression,
and they're not easy to compare to each other.)

Have I managed to make it make sense?  Sorry if I still haven't.

As to your *first* question, about the "level 12" quality, I think I
know a tiny bit about why that's so.  I believe that the range was
1-10 in earlier versions.  They decided to add some options at the
top, and rather than changing the meaning of the previous options,
they just added more numbers (11 and 12) to describe the new options.
I *don't* really know exactly how those differ from 10, though.
--
David Dyer-Bennet, <mailto:dd-b@dd-b.net>, <http://www.dd-b.net/dd-b/>
RKBA: <http://noguns-nomoney.com/> <http://www.dd-b.net/carry/>
Pics: <http://dd-b.lighthunters.net/> <http://www.dd-b.net/dd-b/SnapshotAl.../dragaera.info/>

On September 24 2004, Eric Gill <ericvgill@yahoo.com> wrote:
> The answer is simply that you shouldn't be re-saving in JPEG format.

Not ever?

So how about explaining this to most of the printing services that
PREFER you submit your digital files in JPEG. (MPIX for instance, a
division of Miller, a HIGHLY RESPECTED company in the professional
photography world.)

See, I LOVE the internet because people who DON'T KNOW the answer to the
question always seem to feel the need to phrase something that they DO
KNOW so that it SOMEHOW REMOTELY RELATES to the original question,
thereby SEEMING TO PROVE to themselves and everyone listening that they
DO INDEED KNOW SOMETHING!

--
Carl Miller
carl_miller23@hotmail.com
http://www.ezinfocenter.com/8557444

Tacit wrote: 
>
>
> Very, very simple.
>
> You have an image. Let us say for the sake of example that the image, when
> UNCOMPRESSED, is one megabyte.
>
> You save it as a JPEG. You use quality, say, 4. It saves to about 200
> kilobytes.
>
> You open the JPEG, It uncompresses. The image is 1 megabyte uncompressed.
>
> You save it as JPEG with quality 12. It compresses again. You saved it wit
h a
> higher quality so this time it compresses to 500 kilobytes.
>
> Make sense?
>

No problem, as long as you understand that Photoshop isn't restoring any
of that lossy information that JPEG got rid of when it compressed. PS is
only opening what is there.

However, it opens everything the same. So, this shade of red takes up
just as much memory as that shade of red. That is true even if the
shades of red are the same. That's why the memory size is the big size,
not the little file size. The file size makes certain shades of red all
the same and then treats them as one small piece of data rather than all
of them separately. [Highly simplified]

So, if PS were just cutting down the number of reds when it resaves in
JPEG, the theory it wouldn't change a thing. Of course, YOU may have
changed something in the picture. Then it isn't changing the same thing.
So, it compresses on top of the previous.

Also, you don't know what settings were used for the first JPEG
compression. The odds are that yours won't be the same. So, PS will be
compressing it differently. That means that it's likely to add to lossiness.

Now, if those losses mean anything or are actually visible are another
question completely. You will have to test that to your own satisfaction.

Clyde

Carl Miller wrote:

> On September 24 2004, Eric Gill <ericvgill@yahoo.com> wrote: 
>
> Not ever?
>
> So how about explaining this to most of the printing services that
> PREFER you submit your digital files in JPEG. (MPIX for instance, a
> division of Miller, a HIGHLY RESPECTED company in the professional
> photography world.)
>
> See, I LOVE the internet because people who DON'T KNOW the answer to
> the question always seem to feel the need to phrase something that
> they DO KNOW so that it SOMEHOW REMOTELY RELATES to the original
> question, thereby SEEMING TO PROVE to themselves and everyone
> listening that they DO INDEED KNOW SOMETHING!

Whatever you do, you should handle the decision whether it is time to
save as JPG with the same care as you would the decision to save a text
document in a printer's native format.

You surely would keep a text version - so keep an uncompressed version
of your images.

When you save an image that is in your computer's memory as a JPG image
the compression algorithm does not know anything about the image expect
that it is a rectangular array of pixels.

IF your image was overcompressed earlier, saving at a lower compression
setting will take care to represent all detail fount in the currently
loaded image - including the artifacts created by the earlier save
process. So your image file may become bigger - it won't become better.

If you use very low compression you can be lucky that untrained eyes do
not see the degradation even avter several saves - and if you send
images to printing services that are sufficiently large for a high
number of pixels per inch then artifacts will be small enough to become
invisible. As JPG artifacts occur in blocks of 8x8 pixels, with 300dpi
your artifact size will be 37.5 blocks per inch - not everybody will
notice them.

Printing services love JPG as it reduces their transfer volume.

Michael

>So how about explaining this to most of the printing services that
>PREFER you submit your digital files in JPEG. (MPIX for instance, a
>division of Miller, a HIGHLY RESPECTED company in the professional
>photography world.)

MPIX is a consumer-level printing service that takes images online. They pre
fer
JPEG because it's easiest to transfer JPEG images over the Web, and as a
consumer-level service, they know that the JPEG artifacts will probably not 
be
noticeable or objectionable to their clients.


--
Art, literature, shareware, polyamory, kink, and more:
http://www.xeromag.com/franklin.html