Best Image -- Image Size vs Compression

I use a slide show package that will display images at full screen,
regardless of the size/resolution that the monitor is set at. Since
the slide shows, which are run from a CD, contain many images, there
is a need to keep the image file sizes below 600KB or so. The images
are edited in PS, and stored using PS's jpg save. PS can save jpgs at
MAX (levels 10, 11, and 12) or HIGH (7, 8, or 9).

What I am seeking is to get the most sharpness in the images while
keeping the file sizes below 600KB. The trade-off I am looking at is
a larger image at higher compression, or slightly smaller image at
lower compression. On my test image I came up with the following
combinations that fall within my desired file size range:

1200 at 10 437KB
1000 at 11 476KB
800 at 12 494KB
1100 at 11 572KB

where, for example, 1200 is size of longest side, and 10 is PS
compression setting

In closely examining the images on the screen, it appeared to me that
the 1100 @ 11 was the best, while the 800 @ 12 was the least sharp.
Is there some known truth about size vs compression in terms of jpg
quality, or does one simply arrive at some general truth by this kind
of testing.

Does anyone have any alternatives to recommend? Thanks in advance.

john chapman wrote:

I use a slide show package that will display images at full screen,
regardless of the size/resolution that the monitor is set at. Since
the slide shows, which are run from a CD, contain many images, there
is a need to keep the image file sizes below 600KB or so. The images
are edited in PS, and stored using PS's jpg save. PS can save jpgs at
MAX (levels 10, 11, and 12) or HIGH (7, 8, or 9).

What I am seeking is to get the most sharpness in the images while
keeping the file sizes below 600KB. The trade-off I am looking at is
a larger image at higher compression, or slightly smaller image at
lower compression. On my test image I came up with the following
combinations that fall within my desired file size range:

1200 at 10 437KB
1000 at 11 476KB
800 at 12 494KB
1100 at 11 572KB

where, for example, 1200 is size of longest side, and 10 is PS
compression setting

In closely examining the images on the screen, it appeared to me that
the 1100 @ 11 was the best, while the 800 @ 12 was the least sharp.
Is there some known truth about size vs compression in terms of jpg
quality, or does one simply arrive at some general truth by this kind
of testing.

Does anyone have any alternatives to recommend? Thanks in advance.

You will find that compression ratios are strongly dependent on the
actual image data. Try compressing a picture with a lot of grass and
trees and you will see the effect. Unfortunately, there is no answer
that will work for all pictures. You will just about have to do each
one on its own merits to get the best results. Frankly, I would match
the compression to the original image size, and compress less if the
picture is filled with complex shapes, such as grass and trees, and more
if it is large areas of color. You may see more artifacts, but they
will be less distracting.

john chapman wrote:

I use a slide show package that will display images at full screen,
regardless of the size/resolution that the monitor is set at. Since
the slide shows, which are run from a CD, contain many images, there
is a need to keep the image file sizes below 600KB or so. The images
are edited in PS, and stored using PS's jpg save. PS can save jpgs at
MAX (levels 10, 11, and 12) or HIGH (7, 8, or 9).

What I am seeking is to get the most sharpness in the images while
keeping the file sizes below 600KB. The trade-off I am looking at is
a larger image at higher compression, or slightly smaller image at
lower compression. On my test image I came up with the following
combinations that fall within my desired file size range:

1200 at 10 437KB
1000 at 11 476KB
800 at 12 494KB
1100 at 11 572KB

where, for example, 1200 is size of longest side, and 10 is PS
compression setting

In closely examining the images on the screen, it appeared to me that
the 1100 @ 11 was the best, while the 800 @ 12 was the least sharp.
Is there some known truth about size vs compression in terms of jpg
quality, or does one simply arrive at some general truth by this kind
of testing.

Does anyone have any alternatives to recommend? Thanks in advance.

You will find that compression ratios are strongly dependent on the
actual image data. Try compressing a picture with a lot of grass and
trees and you will see the effect. Unfortunately, there is no answer
that will work for all pictures. You will just about have to do each
one on its own merits to get the best results. Frankly, I would match
the compression to the original image size, and compress less if the
picture is filled with complex shapes, such as grass and trees, and more
if it is large areas of color. You may see more artifacts, but they
will be less distracting.

In message , john
chapman writes
I use a slide show package that will display images at full screen,
regardless of the size/resolution that the monitor is set at. Since
the slide shows, which are run from a CD, contain many images, there
is a need to keep the image file sizes below 600KB or so. The images
are edited in PS, and stored using PS's jpg save. PS can save jpgs at
MAX (levels 10, 11, and 12) or HIGH (7, 8, or 9).

What I am seeking is to get the most sharpness in the images while
keeping the file sizes below 600KB. The trade-off I am looking at is
a larger image at higher compression, or slightly smaller image at
lower compression. On my test image I came up with the following
combinations that fall within my desired file size range:

1200 at 10 437KB
1000 at 11 476KB
800 at 12 494KB
1100 at 11 572KB

where, for example, 1200 is size of longest side, and 10 is PS
compression setting

A single test image will lead you seriously astray unless by chance you
happened to have picked an unusually representative one.

Do the test on a directory of about 100 images and you will get a better
idea of how the size quality trade off works with a range of images.

The ones to inspect really carefully are a couple of typical ones and
the two with the most extreme smallest and largest file sizes after
compression.

In closely examining the images on the screen, it appeared to me that
the 1100 @ 11 was the best, while the 800 @ 12 was the least sharp.
Is there some known truth about size vs compression in terms of jpg
quality, or does one simply arrive at some general truth by this kind
of testing.

You would be well advised to chose a size that is nicely commensurate
with typical screen sizes of 800x600, 1024x768, 1280x960, 1440x1080,
1600x1200. Some display drivers can be relied upon to make a hash of
displaying an image full screen with unusual sizes like e.g. 1100x825.

Does anyone have any alternatives to recommend? Thanks in advance.

You would have a lot more control of the compression settings with
something based on the JIG codec. IrfanView springs to mind as
reasonable quality free software with excellent batch rescaling and JPEG
encoding functions.

In addition it will not store a small essay on PhotoShop and various
assorted colour management dross which will probably save you around
50kb per file.

Regards,
--
Martin Brown

In message , john
chapman writes
I use a slide show package that will display images at full screen,
regardless of the size/resolution that the monitor is set at. Since
the slide shows, which are run from a CD, contain many images, there
is a need to keep the image file sizes below 600KB or so. The images
are edited in PS, and stored using PS's jpg save. PS can save jpgs at
MAX (levels 10, 11, and 12) or HIGH (7, 8, or 9).

What I am seeking is to get the most sharpness in the images while
keeping the file sizes below 600KB. The trade-off I am looking at is
a larger image at higher compression, or slightly smaller image at
lower compression. On my test image I came up with the following
combinations that fall within my desired file size range:

1200 at 10 437KB
1000 at 11 476KB
800 at 12 494KB
1100 at 11 572KB

where, for example, 1200 is size of longest side, and 10 is PS
compression setting

A single test image will lead you seriously astray unless by chance you
happened to have picked an unusually representative one.

Do the test on a directory of about 100 images and you will get a better
idea of how the size quality trade off works with a range of images.

The ones to inspect really carefully are a couple of typical ones and
the two with the most extreme smallest and largest file sizes after
compression.

In closely examining the images on the screen, it appeared to me that
the 1100 @ 11 was the best, while the 800 @ 12 was the least sharp.
Is there some known truth about size vs compression in terms of jpg
quality, or does one simply arrive at some general truth by this kind
of testing.

You would be well advised to chose a size that is nicely commensurate
with typical screen sizes of 800x600, 1024x768, 1280x960, 1440x1080,
1600x1200. Some display drivers can be relied upon to make a hash of
displaying an image full screen with unusual sizes like e.g. 1100x825.

Does anyone have any alternatives to recommend? Thanks in advance.

You would have a lot more control of the compression settings with
something based on the JIG codec. IrfanView springs to mind as
reasonable quality free software with excellent batch rescaling and JPEG
encoding functions.

In addition it will not store a small essay on PhotoShop and various
assorted colour management dross which will probably save you around
50kb per file.

Regards,
--
Martin Brown

(john chapman) wrote in message . com...
I use a slide show package that will display images at full screen,
regardless of the size/resolution that the monitor is set at. Since
the slide shows, which are run from a CD, contain many images, there
is a need to keep the image file sizes below 600KB or so. The images
are edited in PS, and stored using PS's jpg save. PS can save jpgs at
MAX (levels 10, 11, and 12) or HIGH (7, 8, or 9).

What I am seeking is to get the most sharpness in the images while
keeping the file sizes below 600KB. The trade-off I am looking at is
a larger image at higher compression, or slightly smaller image at
lower compression. On my test image I came up with the following
combinations that fall within my desired file size range:

1200 at 10 437KB
1000 at 11 476KB
800 at 12 494KB
1100 at 11 572KB

where, for example, 1200 is size of longest side, and 10 is PS
compression setting

In closely examining the images on the screen, it appeared to me that
the 1100 @ 11 was the best, while the 800 @ 12 was the least sharp.
Is there some known truth about size vs compression in terms of jpg
quality, or does one simply arrive at some general truth by this kind
of testing.

Does anyone have any alternatives to recommend? Thanks in advance.

Why not downsize to the resolution of the media first? In general,
you can downsize a Bayer image (most cameras) to 25% of the original
area with no optical loss whatsoever, since they are all digitally
upscaled 400% as output (they have an interpolated value inserted
between all their measured values, by a computer, after the shutter
closes).

(john chapman) wrote in message . com...
I use a slide show package that will display images at full screen,
regardless of the size/resolution that the monitor is set at. Since
the slide shows, which are run from a CD, contain many images, there
is a need to keep the image file sizes below 600KB or so. The images
are edited in PS, and stored using PS's jpg save. PS can save jpgs at
MAX (levels 10, 11, and 12) or HIGH (7, 8, or 9).

What I am seeking is to get the most sharpness in the images while
keeping the file sizes below 600KB. The trade-off I am looking at is
a larger image at higher compression, or slightly smaller image at
lower compression. On my test image I came up with the following
combinations that fall within my desired file size range:

1200 at 10 437KB
1000 at 11 476KB
800 at 12 494KB
1100 at 11 572KB

where, for example, 1200 is size of longest side, and 10 is PS
compression setting

In closely examining the images on the screen, it appeared to me that
the 1100 @ 11 was the best, while the 800 @ 12 was the least sharp.
Is there some known truth about size vs compression in terms of jpg
quality, or does one simply arrive at some general truth by this kind
of testing.

Does anyone have any alternatives to recommend? Thanks in advance.

Why not downsize to the resolution of the media first? In general,
you can downsize a Bayer image (most cameras) to 25% of the original
area with no optical loss whatsoever, since they are all digitally
upscaled 400% as output (they have an interpolated value inserted
between all their measured values, by a computer, after the shutter
closes).

(john chapman) wrote in message . com...
I use a slide show package that will display images at full screen,
regardless of the size/resolution that the monitor is set at. Since
the slide shows, which are run from a CD, contain many images, there
is a need to keep the image file sizes below 600KB or so. The images
are edited in PS, and stored using PS's jpg save. PS can save jpgs at
MAX (levels 10, 11, and 12) or HIGH (7, 8, or 9).

What I am seeking is to get the most sharpness in the images while
keeping the file sizes below 600KB. The trade-off I am looking at is
a larger image at higher compression, or slightly smaller image at
lower compression. On my test image I came up with the following
combinations that fall within my desired file size range:

1200 at 10 437KB
1000 at 11 476KB
800 at 12 494KB
1100 at 11 572KB

where, for example, 1200 is size of longest side, and 10 is PS
compression setting

In closely examining the images on the screen, it appeared to me that
the 1100 @ 11 was the best, while the 800 @ 12 was the least sharp.
Is there some known truth about size vs compression in terms of jpg
quality, or does one simply arrive at some general truth by this kind
of testing.

Does anyone have any alternatives to recommend? Thanks in advance.

Why not downsize to the resolution of the media first? In general,
you can downsize a Bayer image (most cameras) to 25% of the original
area with no optical loss whatsoever, since they are all digitally
upscaled 400% as output (they have an interpolated value inserted
between all their measured values, by a computer, after the shutter
closes).

On 7 Aug 2004 07:18:11 -0700, (George
Preddy) wrote:

Why not downsize to the resolution of the media first? In general,
you can downsize a Bayer image (most cameras) to 25% of the original
area with no optical loss whatsoever, since they are all digitally
upscaled 400% as output (they have an interpolated value inserted
between all their measured values, by a computer, after the shutter
closes).

What nonsense is this?

Hans-Georg

--
No mail, please.

I do not understand the terminology. Regardless, my images are all
scanned from slides and edited in Photoshop.

(George Preddy) wrote in message . com...
(john chapman) wrote in message . com...
1100 at 11 572KB

Why not downsize to the resolution of the media first? In general,
you can downsize a Bayer image (most cameras) to 25% of the original
area with no optical loss whatsoever, since they are all digitally
upscaled 400% as output (they have an interpolated value inserted
between all their measured values, by a computer, after the shutter
closes).