What program is best at JPEG compression?

On Thu, 19 Jul 2007 22:38:17 -0400, Jonathan wrote:

What is xv? I think I must have missed something somewhere. Like another
poster related. I only use Jpeg for uploading to usenet and I am not that
concerned much beyond that point.

X-Newsreader: Microsoft Outlook Express 6.00.2900.3028


Jon,

Don't worry your pretty little head about it. :-)

In article ,
Scott W wrote:
wrote:
Are some graphics programs better than others for compressing JPEG?
A friend swears that Image Ready is the best for getting the
smallest file with same quality.
But I see a few companies have made utilities purely for
compressing pictures, and claim they obtain better
compression than the top tier graphics programs.
I presume there are different numerical routines used by
different companies in their software, but I would guess
maybe 5% variation in file size at best.
I know there was a comparison done several years ago,
but the website with the pictures seems to be kaput.
(www.imagecompress.com ?)
Is there a recent comparison of different programs to see
who offers the best compression?
That is, best quality for same size, or smallest size
for same quality.

I have a program called Advanced JPEG Compressor, it gives you far more
control in how the compression is does and allow you to use different
amounts of compression in different areas of the image. It can do a
better job, but takes a fair bit of use to use and the gains are fairly
small. I only use it when posting to the web and where I am really
trying to limit the file size of the photo.

The xat.com Image Optimizer, and JPEG Wizard (Pegasus Systems) are two
similar products that work somewhat differently, and have differing
additional feature sets - but also allow direct visual manipulation of the
trade-off between file size and image quality on a per-area basis within
the image. As you say, mainly for web posting of where size is of absolute
importance (like when we email photos to relatives using dial-up
connections).

Gordon Freeman wrote:

I like Irfanview too -- it is a great image and slideshow viewer --
but it does not control chroma subsampling.

You can turn subsampling off in Irfanview by ticking an option during
saving.

Do you mean "Disable color subsampling" option in the JPEG/GIF dialog?
I had not noticed it before, but upon testing, it only switches between
2x2 (default) and 1x1 chroma subsampling. It can't set 4:2:2 like GIMP.

I would agree with the person who said that Photoshop is amongst the worst
- I find I can always shrink a Photoshop-saved JPEG significantly by
running it through a lossless optimiser (such as Irfanview's) afterwards.

And Photoshop's Save For Web produces remarkably lousy quality, allowing
(if I'm not mistaken) quality 99-100 with 2x2 chroma subsampling, which is
just stupid. Anything above about Q 85 should use 1x1 or 4:2:2.

Jonathan wrote:

Xnview is available for Linux, apparently. Do you prefer it to xv?

What is xv? I think I must have missed something somewhere.

XV is an old Unix/X11 program by Jon Bradley that's much like Irfanview,
though obviously not as modern. For example, xv 3.10 lacks EXIF support.
Irfanview is better in some ways, worse in others. I use xv on Linux
because Irfanview runs only on Windows, and GIMP (like Photoshop) is more
an editor than a viewer.

Like another poster related. I only use Jpeg for uploading to Usenet
and I am not that concerned much beyond that point.

In theory that is the best plan, but my friends are often sending me
JPEG files from digital cameras incapable of producing RAW. So I have
learned how to edit JPEG with minimal damage.

On 20 Jul 2007 15:44:07 -0700, Bill Tuthill wrote:
I use xv on Linux
because Irfanview runs only on Windows, .........

Irfanview runs just fine under wine.

On Thu, 19 Jul 2007 09:57:39 -0500, "CSM1"
wrote:

wrote in message
roups.com...
Are some graphics programs better than others for compressing JPEG?
A friend swears that Image Ready is the best for getting the
smallest file with same quality.
But I see a few companies have made utilities purely for
compressing pictures, and claim they obtain better
compression than the top tier graphics programs.
I presume there are different numerical routines used by
different companies in their software, but I would guess
maybe 5% variation in file size at best.
I know there was a comparison done several years ago,
but the website with the pictures seems to be kaput.
(www.imagecompress.com ?)
Is there a recent comparison of different programs to see
who offers the best compression?
That is, best quality for same size, or smallest size
for same quality.


Jpeg compression is a big subject, since JPEG is a lossy compression scheme,
there are always trade offs. File size vs. amount of compression and
artifacts.

The better Graphic programs give you a choice of the amount of compression
used.
The de facto best Photo editor it Adobe Photoshop. The full version is about
$650 or $150 for the upgrade.

Photoshop Elements is a consumer grade Photo editor that costs much less
money ($100) and does most of what Photoshop CS does.
http://www.adobe.com/products/photoshopelwin/


I've always thought jpeg and its compression algorithm were subject to
a general standard and not to any program makers' fantasy ..????

imbsysop wrote:
[] I've always thought jpeg and its compression algorithm were subject to
a general standard and not to any program makers' fantasy ..????

There is a standard, but with a large number of different choices. For
example, you can change the colour resolution relative to the luminance
resolution. How different programmers interpret "95% quality" is up to
them, so it's entirely possible that different programs will better suit
different images.

Cheers,
David

"imbsysop" wrote in message
...
I've always thought jpeg and its compression algorithm
were subject to
a general standard and not to any program makers' fantasy
..????

As with MPEG, JPEG decoders are deterministic (given the
same input, they all produce the same output), but the
encoders are not. Choices can be made during encoding even
for algorithms that produce the same output file size. MPEG
encoding has even more variance and there can be enormous
quality differences in the decoded output of a given average
bit rate. Coders for DVDs do a good job because the encoding
does not have to be done in real time and multipass encoding
and larger GOP sizes are used.

David

On Jul 23, 1:47 pm, "David" wrote:
"imbsysop" wrote in message

... I've always thought jpeg and its compression algorithm
were subject to
a general standard and not to any program makers' fantasy
..????

The standard has a large number of free parameters in it. And although
many applications use a scaled version of the original canonical
quantisation table given in the standard other encoders have made
different choices. And in some cases an encoder with all the time in
the world can find an optimised quantisation table for a specific
target image. This helps encode physically small dimensioned images in
a more compact JPEG file.

As with MPEG, JPEG decoders are deterministic (given the
same input, they all produce the same output), but the
encoders are not. Choices can be made during encoding even
for algorithms that produce the same output file size.

Sadly even though the specification of JPEG is fairly precise there is
still enough latitude for the decoder that there is plenty of scope
for different decoders producing different images from the same JPEG
encoded file. The most commonly noticeable implementation differences
are in the chroma subsampling treatment (seriously broken in PSPro 9's
encoder). And where for example the IJG codec uses interpolation
whereas Adobe uses pixel replication - both of these methods have
their merits and demerits.

There are other options dating from the bad old days when computing a
DCT was still hard work that permit various speed for accuracy trade-
offs (much less of an issue these days with fast hardware floating
point support).

Regards,
Martin Brown

"imbsysop" wrote in message
...
On Thu, 19 Jul 2007 09:57:39 -0500, "CSM1"
wrote:

wrote in message
groups.com...
Are some graphics programs better than others for compressing JPEG?
A friend swears that Image Ready is the best for getting the
smallest file with same quality.
But I see a few companies have made utilities purely for
compressing pictures, and claim they obtain better
compression than the top tier graphics programs.
I presume there are different numerical routines used by
different companies in their software, but I would guess
maybe 5% variation in file size at best.
I know there was a comparison done several years ago,
but the website with the pictures seems to be kaput.
(www.imagecompress.com ?)
Is there a recent comparison of different programs to see
who offers the best compression?
That is, best quality for same size, or smallest size
for same quality.


Jpeg compression is a big subject, since JPEG is a lossy compression
scheme,
there are always trade offs. File size vs. amount of compression and
artifacts.

The better Graphic programs give you a choice of the amount of compression
used.
The de facto best Photo editor it Adobe Photoshop. The full version is
about
$650 or $150 for the upgrade.

Photoshop Elements is a consumer grade Photo editor that costs much less
money ($100) and does most of what Photoshop CS does.
http://www.adobe.com/products/photoshopelwin/


I've always thought jpeg and its compression algorithm were subject to
a general standard and not to any program makers' fantasy ..????


Have you read the JPEG specification?
http://www.w3.org/Graphics/JPEG/jfif3.pdf

If you are talking about Digital cameras, you use the EXIF 2.2 spec.
http://www.exif.org/Exif2-2.PDF

There is a lot of room for interpolation of the application of JPEG
compression.
Many factors that have to be balanced for the best result.

Since Jpeg is lossy, it throws away a lot of important information that is
gone forever, if you choose wrong.


--
CSM1
http://www.carlmcmillan.com
--