JPEG Questions: Loss In Quality When "Saving"

I’m confused about quality deterioration when repeatedly opening and "saving" JPEG images. I’ve read that every time a JPEG image is "saved," some quality is lost due to the lossy compression format.
My questions are….

1) What is meant by "saving" a JPEG image?

2) If I simply "open" a JPEG image in Microsoft Photo Editor, Adobe Photoshop CS or any other image editor, View the JPEG image and then simply exit out of viewing the JPEG, am I actually "saving" it again and therefore loosing quality?

3) If in Windows I right click on a JPEG and make a duplicate copy of the JPEG, is the copy (with a different name) "saved" and therefore the quality of the copy diminished.

4) If I open a JPEG in a photo editor and "save as" (instead of simply exiting) and use the same name as the JPEG I just opened, is the quality diminished?

When I do all of the above things, I don’t notice any decrease in quality and the JPEG seems to be exactly the same size so I’m guessing I haven’t degraded the original. If I’m wrong, I’m guessing I should make each JPEG file "read only" or burn them to a CD or DVD so they can’t be changed when I view them.

If I open up a JPEG in a photo editing program and make any changes to the JPEG at all, I can understand that the quality degrades. Any advice on my questions are appreciated.

1) What is meant by "saving" a JPEG image?

2) If I simply "open" a JPEG image in Microsoft Photo Editor, Adobe

CS or any other image editor, View the JPEG image and then simply exit

viewing the JPEG, am I actually "saving" it again and therefore

quality?

3) If in Windows I right click on a JPEG and make a duplicate copy of

JPEG, is the copy (with a different name) "saved" and therefore the

the copy diminished.

4) If I open a JPEG in a photo editor and "save as" (instead of

exiting) and use the same name as the JPEG I just opened, is the

diminished?

When I do all of the above things, I don’t notice any decrease

quality and the JPEG seems to be exactly the same size so I’m guessing

haven’t degraded the original.

If I’m wrong, I’m guessing I should make each
JPEG file "read only" or burn them to a CD or DVD so they can’t be

I view them.

If I open up a JPEG in a photo editing program and make any

the JPEG at all, I can understand that the quality degrades.

It doesn’t matter if you change it or not. If you open it up and then re-save it, you are causing a decompress and recompress of the file, which will result in quality degradation. Obviously, the level of compression will determine just how much. If you save at Maximum quality, you will not lose nearly as much quality as you would in lower settings. Any compression setting will still lose SOME quality, however, especially if you re-save multiple times, so this is why you want to leave images uncompressed until they’re completely final.

"Darrel Hoffman" wrote in message

It doesn’t matter if you change it or not. If you open it up and then re-save it, you are causing a decompress and recompress of the file, which will result in quality degradation. Obviously, the level of compression will determine just how much. If you save at Maximum quality, you will not lose nearly as much quality as you would in lower settings. Any compression setting will still lose SOME quality, however, especially if you re-save multiple times, so this is why you want to leave images uncompressed until they’re completely final.

Sorry but this is wrong.
Opening and restoring a JPEG will NOT cause any degradation at all. You lose data if you change something to your file and recompress it or, of course, if you don’t change anything to it but recompress it at a lower setting.
Try it and see by yourself: open and close a JPEG as many times as you want and try to find any sign of data loss.

Stephan

Sorry, you did not understand what Darrel wrote, he wrote: "if you

it" whch is right, whether you change it or not

Sorry, you did not understand what Darrel wrote, he wrote: "if you

re-save

it" whch is right, whether you change it or not

Yeah, what he said. If you open and SAVE a jpg, whether or not you make any changes, you will losw quality. Saving a file is different from just closing a file.

Darrel Hoffman wrote:

Sorry, you did not understand what Darrel wrote, he wrote: "if you

re-save

it" whch is right, whether you change it or not

Yeah, what he said. If you open and SAVE a jpg, whether or not you make any changes, you will losw quality. Saving a file is different from just closing a file.

if the same settings are used to recompress a decompressed Jpeg – it doesn’t loose any data. try it – open a pix and then resave it, the file sizes match. at least the code in the jpeg consortiums comp/decomp routines do ( as far as i’v gotten into them ). it’s because the decomp just reverses the comp alogrithim and since all the data fits the huff table, it doesn’t throw any out. this wouldn’t hold true if the compressor wasn’t the same as the decomp.

it’s kinda like a 3D jigsaw puzzle that fits in a box all the pieces get trimed to fit the box and will fit back together in the box as long as no piece gets changed and the same person repacks the box! a diierent pakager will not put it back together the same way and will ‘trim’ a piece or two to fit.

Sorry, you did not understand what Darrel wrote, he wrote: "if you

re-save

it" whch is right, whether you change it or not

Yeah, what he said. If you open and SAVE a jpg, whether or not you make any changes, you will losw quality. Saving a file is different from just closing a file.

Stephan wrote:

Sorry, you did not understand what Darrel wrote, he wrote: "if you re-save it" whch is right, whether you change it or not

No difference.
Saving without changes is the same as just closing.
Try it!

Stephan

I did I created a 100 by 100 image of a gradient fill. Then I saved named original.jpg. at that point it was 9,673 bytes.

Then I saved it as original not.jpg and proceeded to open it and save it

repeat 6 times. With out making any changes. At that point the new file, Original not’s size was 9,688 bytes.

I used the same compression ratio every time, 4. While the difference was small the file is small also.

Was the visual difference noticeable? Barely but yes.

I did I created a 100 by 100 image of a gradient fill. Then I saved

original.jpg. at that point it was 9,673 bytes.

Then I saved it as original not.jpg and proceeded to open it and

and

repeat 6 times. With out making any changes. At that point the new

Original not’s size was 9,688 bytes.

I used the same compression ratio every time, 4. While the

small the file is small also.

Was the visual difference noticeable? Barely but yes.

Interesting, Can you show it?

Interesting, Can you show it?

Interesting, Can you show it?

I’ve posted an example in alt.binaries.comp-graphics. Look for thread titled "Results of JPG compression".

Interesting, Can you show it?

I’ve posted an example in alt.binaries.comp-graphics. Look for thread titled "Results of JPG compression".

"Darrel Hoffman" wrote in message

Interesting, Can you show it?

I’ve posted an example in alt.binaries.comp-graphics. Look for thread titled "Results of JPG compression".

I tried on Google groups:
Your search – Results of JPG compression group:alt.binaries.comp-graphics – did not match any messages posted between Mar 12, 2004 and today. Where is it?

On Mon, 12 Apr 2004 23:29:43 GMT, "Stephan"
scribbled:

"Darrel Hoffman" wrote in message

Interesting, Can you show it?

I’ve posted an example in alt.binaries.comp-graphics. Look for thread titled "Results of JPG compression".

I tried on Google groups:
Your search – Results of JPG compression

did not match any messages posted between Mar 12, 2004 and today. Where is it?

Google doesn’t carry the binary groups.

OK, makes sense: I can’t see it then because my ISP can’t find the

either

I’ve posted an example in *alt.binaries.comp-graphics*. Look for

titled "Results of JPG compression".

Just as an FYI:
news:alt.binaries.comp-graphics

Will take most complaint news readers right there.

I’ve posted an example in *alt.binaries.comp-graphics*. Look for

thread

titled "Results of JPG compression".

Just as an FYI:
news:alt.binaries.comp-graphics

Will take most complaint news readers right there.

Is there a reason you just told me what I obviously just figured out on my own?

And while I do have many complaints about Outlook, I assume you meant "compliant"?

I’ve posted an example in *alt.binaries.comp-graphics*. Look for

thread

titled "Results of JPG compression".

Just as an FYI:
news:alt.binaries.comp-graphics

Will take most complaint news readers right there.

Is there a reason you just told me what I obviously just figured out

my own?

How is it obvious you figured it out on your own? If you figured it out on your own why didn’t you include the news: link in the first place?

I’ve posted an example in *alt.binaries.comp-graphics*. Look for

thread

titled "Results of JPG compression".

Just as an FYI:
news:alt.binaries.comp-graphics

Will take most complaint news readers right there.

Is there a reason you just told me what I obviously just figured out

on

my own?

How is it obvious you figured it out on your own? If you figured it out on your own why didn’t you include the news: link in the first place?

Note asterisks in previous quote… You told me to post it to a group after I just said I’d posted to that exact group.

OK, makes sense: I can’t see it then because my ISP can’t find the

group

either

Well, anyways, it’s not hard to reproduce the results yourself. Take any image, (I used a black-on-white line drawing), and save as JPG, quality 4*. Now open it, and save it again under a new name, same settings. Open the new file, and re-save it. (Use "Save as…" and save on the same file name. Just doing a Ctrl-S won’t do anything, because Photoshop knows you haven’t changed anything, so it won’t actually save.) Repeat this step 4 or 5 times. Now open your original saved JPG and compare it to the one you’ve saved repeatedly. I guarantee you will notice a difference.

*(You’ll lose quality over time even with maximum quality, but it’s much more obvious with a lower setting. If you use the highest quality it may take 15-20 iterations for the compression artifacts to be noticeable.)

I don’t know why you have to waste your time to do that kind of test,

why you need a test to know it will show some difference.
Since Photoshop has 12 as MAX quality or 100% so "4" is 1/3 or about 33.33% of the original. So you only need to save ONCE at 33% to see

difference, and you don’t even need to save once to know 33.33% won’t

good as 100%

Or you save the 2nd time as 100% of 33.33% you will only get 100% of 33.33% *not* 100% of the original, so your test won’t make any sence.

I don’t know why you have to waste your time to do that kind of test,

and

why you need a test to know it will show some difference.
Since Photoshop has 12 as MAX quality or 100% so "4" is 1/3 or about 33.33% of the original. So you only need to save ONCE at 33% to see

the

difference, and you don’t even need to save once to know 33.33% won’t

be as

good as 100%

Or you save the 2nd time as 100% of 33.33% you will only get 100% of 33.33% *not* 100% of the original, so your test won’t make any sence.

Test repeated at Maximum quality of 12 every save. Yes, it takes many more repetitions for the quality degradation to be obvious, but it’s still there. And no, a quality level of 4 is not a strict 33% of original quality. In this case, the uncompressed bmp file is 900kb, and the jpg after only 1 compression is 63kb. After 40 repetitions it is 58kb. Don’t tell me there’s no data loss. The file went from 63kb to 58kb without any changes. Where did those 5kb go? Not to mention how it was reduced to 1/15th the uncompressed size. JPG causes quality loss. Plain and simple. You can minimize the degree of loss by cranking the quality up, but you’ll still lose quality every time you de-compress and re-compress it.

If 4 out of 12 isn’t 1/3 then beat me, cuz I don’t know what else to say.

On Wed, 14 Apr 2004 17:04:38 -0500, (Joe) wrote:

If 4 out of 12 isn’t 1/3 then beat me, cuz I don’t know what else to say.

It assumes that the scale is arithmetical. IIRC, it isn’t.