Better JPEG program - minimized JPEG degredation

I was following the discussion and started doing some searching.

I found this app called Better JPEG that says it does some key
operations in a method that absolutley minimizes recompression.

http://www.betterjpeg.com/

Has anyone heard about this? It seems a smart way to do things,
but I'm not sure if it is BS or if it is real.

If you have any experience on this app, please share some
feedback if you would.

Thanks

Paul D. Sullivan wrote:
I was following the discussion and started doing some searching.

I found this app called Better JPEG that says it does some key
operations in a method that absolutley minimizes recompression.

http://www.betterjpeg.com/

Has anyone heard about this? It seems a smart way to do things,
but I'm not sure if it is BS or if it is real.

If you have any experience on this app, please share some
feedback if you would.

Thanks

With the right software, you can perform cropping and 90-degree step
rotations on JPEG images /without/ incurring any extra loss due to
re-compression. E.g. Jpegcrop:

http://sylvana.net/jpegcrop/

David

"David J Taylor"
wrote in
message ...
Paul D. Sullivan wrote:
I was following the discussion and started doing some searching.

I found this app called Better JPEG that says it does some key
operations in a method that absolutley minimizes recompression.

http://www.betterjpeg.com/

Has anyone heard about this? It seems a smart way to do things,
but I'm not sure if it is BS or if it is real.

If you have any experience on this app, please share some
feedback if you would.

Thanks

With the right software, you can perform cropping and 90-degree step
rotations on JPEG images /without/ incurring any extra loss due to
re-compression. E.g. Jpegcrop:

http://sylvana.net/jpegcrop/

David

Irfanview (FREE) also supports "lossless" jpg rotations.

mikey

Today, David J Taylor made these interesting comments ...

Paul D. Sullivan wrote:
I was following the discussion and started doing some
searching.

I found this app called Better JPEG that says it does some
key operations in a method that absolutley minimizes
recompression.

http://www.betterjpeg.com/

Has anyone heard about this? It seems a smart way to do
things, but I'm not sure if it is BS or if it is real.

If you have any experience on this app, please share some
feedback if you would.

Thanks

With the right software, you can perform cropping and
90-degree step rotations on JPEG images /without/ incurring
any extra loss due to re-compression. E.g. Jpegcrop:

http://sylvana.net/jpegcrop/

David
adept cropping and 90 deg rotates can indeed result with no image
degradation but how does one then save the image without incurring
at least some loss, unless you've figured out an algorithm for
altering the compression factor, chroma subsampling, etc. that
minimizes damage? I've always been taught that as soon as you re-
save at least some damage is done, but NOT necessarily enough to be
concerned about or to even see

--
HP, aka Jerry

Paul D. Sullivan wrote:
I was following the discussion and started doing some
searching.

I found this app called Better JPEG that says it does some
key operations in a method that absolutley minimizes
recompression.

http://www.betterjpeg.com/

Has anyone heard about this? It seems a smart way to do
things, but I'm not sure if it is BS or if it is real.

If you have any experience on this app, please share some
feedback if you would.

Thanks

With the right software, you can perform cropping and
90-degree step rotations on JPEG images /without/ incurring
any extra loss due to re-compression. E.g. Jpegcrop:

http://sylvana.net/jpegcrop/

David
adept cropping and 90 deg rotates can indeed result with no
image degradation but how does one then save the image without
incurring at least some loss, unless you've figured out an
algorithm for altering the compression factor, chroma
subsampling, etc. that minimizes damage? I've always been
taught that as soon as you re- save at least some damage is
done, but NOT necessarily enough to be concerned about or to
even see

I was impressed by their language on their site in way of
explanation. They say they optimize it so only the actual pixels
that have changed are re-saved. It does not process any other
blocks of pixels. Sounds quite smart.

I think I'm going to try that program out more in-depth. Red Eye
Reduction with almost no hit in quality sounds quite good indeed.

Today, Paul D. Sullivan made these interesting comments ...

Paul D. Sullivan wrote:
I was following the discussion and started doing some
searching.

I found this app called Better JPEG that says it does some
key operations in a method that absolutley minimizes
recompression.

http://www.betterjpeg.com/

Has anyone heard about this? It seems a smart way to do
things, but I'm not sure if it is BS or if it is real.

If you have any experience on this app, please share some
feedback if you would.

Thanks

With the right software, you can perform cropping and
90-degree step rotations on JPEG images /without/ incurring
any extra loss due to re-compression. E.g. Jpegcrop:

http://sylvana.net/jpegcrop/

David
adept cropping and 90 deg rotates can indeed result with no
image degradation but how does one then save the image
without incurring at least some loss, unless you've figured
out an algorithm for altering the compression factor, chroma
subsampling, etc. that minimizes damage? I've always been
taught that as soon as you re- save at least some damage is
done, but NOT necessarily enough to be concerned about or to
even see

I was impressed by their language on their site in way of
explanation. They say they optimize it so only the actual
pixels that have changed are re-saved. It does not process
any other blocks of pixels. Sounds quite smart.

I think I'm going to try that program out more in-depth. Red
Eye Reduction with almost no hit in quality sounds quite good
indeed.

Can't comment on a specific program which may have been written
to minimize image damage under certain well-defined situations as
I haven't tried it/them. I was talking in general terms using my
knowledge - or lack thereof - of who the JPEG spec was designed
and how it is implemented in software.

For me, I try never to re-edit the same image, if I can, I will
go back to the original unedited camera image which I always
save. However, both the 80/20 Rule and the Law of Diminishing
Returns get in the way of that rigid a rule, so I do re-edit my
own or others images. However, I carefully examine the image to
see what damage is already there - it may be slight or
considerable - and usually go to a lower compression rate and/or
alter chroma sub-sampling to achieve best possible results.

Depending on the types of damage I may see, such as jaggies,
posterization, artefacts, noise, etc. etc., I may apply mild-to-
agresssive corrective action before re-saving. But, one thing I
ALWAYS do, no matter if it is a first-time save or a multiple
edit/save/edit/save cycle, is immedately open the just saved (or
re-saved) image an relook for damage. Sometime I see considerable
damage even on a 1st time save, for which I alter my technique as
described above.

Since I am not a pro nor do I print to large sizes, I can afford
compromizes that others may find to be unacceptable. Thus, unless
I am specifically saving proprietary items such as layers vector
data, I usually don't save to a non-compressed format.

Just one man's opinion, YMMV ...

--
HP, aka Jerry

Good points you make. What with data storage being so cheap
compared to the old days, I can get 7 meg JPG into 14 meg PSP and
store 'em in that native Paint Shop Pro format without much
trouble.

Can't comment on a specific program which may have been written
to minimize image damage under certain well-defined situations
as I haven't tried it/them. I was talking in general terms
using my knowledge - or lack thereof - of who the JPEG spec
was designed and how it is implemented in software.

For me, I try never to re-edit the same image, if I can, I will
go back to the original unedited camera image which I always
save. However, both the 80/20 Rule and the Law of Diminishing
Returns get in the way of that rigid a rule, so I do re-edit my
own or others images. However, I carefully examine the image to
see what damage is already there - it may be slight or
considerable - and usually go to a lower compression rate
and/or alter chroma sub-sampling to achieve best possible
results.

Depending on the types of damage I may see, such as jaggies,
posterization, artefacts, noise, etc. etc., I may apply
mild-to- agresssive corrective action before re-saving. But,
one thing I ALWAYS do, no matter if it is a first-time save or
a multiple edit/save/edit/save cycle, is immedately open the
just saved (or re-saved) image an relook for damage. Sometime
I see considerable damage even on a 1st time save, for which I
alter my technique as described above.

Since I am not a pro nor do I print to large sizes, I can
afford compromizes that others may find to be unacceptable.
Thus, unless I am specifically saving proprietary items such
as layers vector data, I usually don't save to a
non-compressed format.

Just one man's opinion, YMMV ...

Today, Paul D. Sullivan made these interesting comments ...

Good points you make. What with data storage being so cheap
compared to the old days, I can get 7 meg JPG into 14 meg PSP
and store 'em in that native Paint Shop Pro format without
much trouble.

Since I seldome use vector data or extensive layering, PSPimage
isn't that important to me, nor have the many advantages of RAW
been enough to overcome the steep learning curve.

For your example of a 7 MB JPEG into an 8MB PSP, that seems
extreme. What pixel size was used in this comparo? When I do
that, the ratio is more like 10:1 in favor JPEG.

Can't comment on a specific program which may have been
written to minimize image damage under certain well-defined
situations as I haven't tried it/them. I was talking in
general terms using my knowledge - or lack thereof - of who
the JPEG spec was designed and how it is implemented in
software.

For me, I try never to re-edit the same image, if I can, I
will go back to the original unedited camera image which I
always save. However, both the 80/20 Rule and the Law of
Diminishing Returns get in the way of that rigid a rule, so I
do re-edit my own or others images. However, I carefully
examine the image to see what damage is already there - it
may be slight or considerable - and usually go to a lower
compression rate and/or alter chroma sub-sampling to achieve
best possible results.

Depending on the types of damage I may see, such as jaggies,
posterization, artefacts, noise, etc. etc., I may apply
mild-to- agresssive corrective action before re-saving. But,
one thing I ALWAYS do, no matter if it is a first-time save
or a multiple edit/save/edit/save cycle, is immedately open
the just saved (or re-saved) image an relook for damage.
Sometime I see considerable damage even on a 1st time save,
for which I alter my technique as described above.

Since I am not a pro nor do I print to large sizes, I can
afford compromizes that others may find to be unacceptable.
Thus, unless I am specifically saving proprietary items such
as layers vector data, I usually don't save to a
non-compressed format.

Just one man's opinion, YMMV ...







--
HP, aka Jerry

HEMI-Powered wrote:
[]
adept cropping and 90 deg rotates can indeed result with no image
degradation but how does one then save the image without incurring
at least some loss, unless you've figured out an algorithm for
altering the compression factor, chroma subsampling, etc. that
minimizes damage? I've always been taught that as soon as you re-
save at least some damage is done, but NOT necessarily enough to be
concerned about or to even see

In programs such as JPEGcrop, although the image is presented to the
viewer in an uncompressed form for examination and selection of the
cropping region (i.e. as 24-bit RGB), internally the 8 x 8, or 16 x 16
blocks comprising the original JPEG are retained, and the rotation or crop
are made on these blocks, and /not/ on the RGB data.

So, for example, crop simply consists of writing out only the blocks you
need, and altering the file headers to reflect the new number of pixels.
The content of the blocks - the compressed JPEG data - is not altered, so
no new compression loss is incurred because there is no recompression.

Lossless rotation (only at 90 degree angles) is achieved by similar
mathematical operations on the data in the blocks, but again without any
decompression and re-compression.

It's a neat idea which works very well.

David

Since I seldome use vector data or extensive layering, PSPimage
isn't that important to me, nor have the many advantages of RAW
been enough to overcome the steep learning curve.

I use tons of layers, so it's a very convenient format for me.