Order of steps in repairing photos

From: (JJGrey)

I am always confused as to which order to do the following steps:
1) adjusting contrast and color, such as using levels
2) sharpening
3) resizing and resampling
Can anyone suggest the correct order.

Flip steps 2 and 3 and you’ve got it …

JJGrey wrote:

I am always confused as to which order to do the following steps:
1) adjusting contrast and color, such as using levels
2) sharpening
3) resizing and resampling
Can anyone suggest the correct order.
Thanks,

J. Grey

3
1
2

Uni

There is no correct order…

By "repairing" do you mean touchup and processing for optimum results, or do you mean restoration? Some processes are common to all.

For example, one of the first things to do is establish the largest print size you are going to need and set the print size accordingly. You don HAVE to do this first, but it helps if you only need a smaller file than what you are starting with—eg if you are doing a lot of cropping.

The order of sharpening, color, contrast, touchup, etc. is probably not important.

On 16 Feb 2004 17:35:05 -0800, (JJGrey) wrote:

I am always confused as to which order to do the following steps:
1) adjusting contrast and color, such as using levels
2) sharpening
3) resizing and resampling
Can anyone suggest the correct order.
Thanks,

J. Grey

**************************
Mark Herring, Pasadena, Calif.
Private e-mail: Just say no to "No".

Mark Herring wrote:

There is no correct order…

By "repairing" do you mean touchup and processing for optimum results, or do you mean restoration? Some processes are common to all.
For example, one of the first things to do is establish the largest print size you are going to need and set the print size accordingly. You don HAVE to do this first, but it helps if you only need a smaller file than what you are starting with—eg if you are doing a lot of cropping.

The order of sharpening, color, contrast, touchup, etc. is probably not important.

Probably?

Uni

On 16 Feb 2004 17:35:05 -0800, (JJGrey) wrote:

I am always confused as to which order to do the following steps:
1) adjusting contrast and color, such as using levels
2) sharpening
3) resizing and resampling
Can anyone suggest the correct order.
Thanks,

J. Grey

**************************
Mark Herring, Pasadena, Calif.
Private e-mail: Just say no to "No".

Mark Herring wrote:
The order of sharpening, color, contrast, touchup, etc. is probably not important.

Probably?

Uni

I’m with Uni on this one. It is important and you might want to consider following the suggested order to improve your output. ~Doc

From: Mark Herring

There is no correct order…

The order of sharpening, color, contrast, touchup, etc. is probably not important.

Pretty much everybody in the industry would disagree with this.

The generally accepted flow is to make the global corrections first, then the local corrections, then re-size and finally sharpen.

"Bill Hilton" wrote in message

From: Mark Herring

There is no correct order…

The order of sharpening, color, contrast, touchup, etc. is probably not important.

Pretty much everybody in the industry would disagree with this.
The generally accepted flow is to make the global corrections first, then

the

local corrections, then re-size and finally sharpen.

I concur, however I [sometimes] find that a slight, global contrast tweak is beneficial *after* I see the sharpening results.

—–= Posted via Newsfeeds.Com, Uncensored Usenet News =—– http://www.newsfeeds.com – The #1 Newsgroup Service in the World! —–== Over 100,000 Newsgroups – 19 Different Servers! =—–

On 17 Feb 2004 16:16:00 GMT, (Bill Hilton)
wrote:

From: Mark Herring

There is no correct order…

The order of sharpening, color, contrast, touchup, etc. is probably not important.

Pretty much everybody in the industry would disagree with this.
The generally accepted flow is to make the global corrections first, then the local corrections, then re-size and finally sharpen.

Bill is absolutely right, get in the habit of doing these things in the right order. If you need help with photo repairing, there’s probably no better source than Katrin Eismann’s "Photoshop Restoration & Retouching," now in its 2nd Edn.

The generally accepted flow is to make the global corrections first, then the local corrections, then re-size and finally sharpen.

In a November 2003 Creativepro article

www.creativepro.com/story/feature/20357.html?cprose=4-44

Bruce Fraser argues for a new approach to sharpening that’s based on three imaging phases: capture, creative, and output. In this workflow some sharpening is applied in earlier stages, in addition to the traditional final step of sharpening for output:

1 Capture Sharpening
is applied early in the image-editing process, and just aims to restore any sharpness that was lost in the capture process. 2 Creative Sharpening
is usually applied locally to accentuate specific features in an image-for example, we often give eyes a little extra sharpness in head shots.
3 Output Sharpening
is applied to files that have already had capture and creative sharpening applied, after they’ve been sized to final output resolution, and is tailored to a specific type of output process.

Bob Shomler
www.shomler.com

I am always confused as to which order to do the following steps:
1) adjusting contrast and color, such as using levels
2) sharpening
3) resizing and resampling
Can anyone suggest the correct order.
Thanks,

J. Grey

3
1
2

The problem witht he order in which Uni suggests you do things is that he would have you do color correction AFTER resizing–which means that if you then discover you want the corrected image at a different size, you have to color correct again.

First, do your color correction on your raw image, before you do anything else. Reason? Because then you can create different sizes whose color all will match.

Then, resize and resample. Reason? Sharpening effects will vary depending on image size and intended use 9for example, an image intended for output on an offset printing press may need more sharpening than an image intended for output to the Web). If you sharpen an image, then resample it, the effects of the sharpening may be lost.

Finally, sharpen. This should be the very last step. Why? Because the proper sharpening amount and technique depends on the size and nature of the final output.

—
Rude T-shirts for a rude age: http://www.villaintees.com Art, literature, shareware, polyamory, kink, and more:
http://www.xeromag.com/franklin.html

Bill Hilton wrote:

From: Mark Herring

There is no correct order…

The order of sharpening, color, contrast, touchup, etc. is probably not important.

Pretty much everybody in the industry would disagree with this.
The generally accepted flow is to make the global corrections first, then the local corrections, then re-size and finally sharpen.

I still disagree. Resize mean resampling. You don’t make an image beautiful, then resample it.

Uni

Tacit wrote:

I am always confused as to which order to do the following steps:
1) adjusting contrast and color, such as using levels
2) sharpening
3) resizing and resampling
Can anyone suggest the correct order.
Thanks,

J. Grey

3
1
2

The problem witht he order in which Uni suggests you do things is that he would have you do color correction AFTER resizing–which means that if you then discover you want the corrected image at a different size, you have to color correct again.

First, do your color correction on your raw image, before you do anything else. Reason? Because then you can create different sizes whose color all will match.
Then, resize and resample. Reason? Sharpening effects will vary depending on image size and intended use 9for example, an image intended for output on an offset printing press may need more sharpening than an image intended for output to the Web). If you sharpen an image, then resample it, the effects of the sharpening may be lost.

Finally, sharpen. This should be the very last step. Why? Because the proper sharpening amount and technique depends on the size and nature of the final output.

No, no, no. You do not color correct then resample. Remember, when you resample, you DISCARD information.

Uni

No, no, no. You do not color correct then resample. Remember, when you resample, you DISCARD information.

So? You don’t discard color information the way you discard detail information. A red stop sign, when resampled, will still be red.

If you do as you suggest–resample, then color correct–what happens when you want the large version’s color to match the small version’s color? You have to color correct again.

A large image will, when resampled to a small image, maintain the same relative color balance.

I don’t know what your background is, but I have been doing this professionally since the days of Photoshop 1.0. My advice comes from this experience.

—
Rude T-shirts for a rude age: http://www.villaintees.com Art, literature, shareware, polyamory, kink, and more:
http://www.xeromag.com/franklin.html

Uni wrote:

Pretty much everybody in the industry would disagree with this.
The generally accepted flow is to make the global corrections first, then the local corrections, then re-size and finally sharpen.

I still disagree. Resize mean resampling. You don’t make an image beautiful, then resample it.

Why not? "To make beautiful" means making color corrections. Resampling has nothing to do with that and doesn’t change any colors. There is absolutely no difference between first color correcting, then resampling, or first resampling, then color correcting. It simply doesn’t matter.

—
Johan W. Elzenga johan<<at>>johanfoto.nl Editor / Photographer http://www.johanfoto.nl/

Resampling
has nothing to do with that and doesn’t change any colors. There is absolutely no difference between first color correcting, then resampling, or first resampling, then color correcting. It simply doesn’t matter.

Correct. You’ve made my point exactly. Thank you.

So why do I say color correct first?

Because I’ve been in the industry for a long time. Inevitably, here’s what happens:

You get an image. You resample it. You color correct it. A year later, you discover that you need to use the image again, at the bigger size–and you need the color to match exactly.

Now you’re screwed.

Now you have to go back to the unresampled image and try to remember exactly how you color corrected it a year ago.

If you color correct first, and then SAVE THAT FILE, and then resample, you can go back to the original color corrected version later, and create different sized images whose color matches.

It’s about covering your ass and avoiding repetitive duplication of work.

—
Rude T-shirts for a rude age: http://www.villaintees.com Art, literature, shareware, polyamory, kink, and more:
http://www.xeromag.com/franklin.html

From: (Tacit)

If you color correct first, and then SAVE THAT FILE, and then resample, you can go back to the original color corrected version later, and create

different

sized images whose color matches.

It’s about covering your ass and avoiding repetitive duplication of work.

This is exactly what every other "expert" says (and Tacit is definitely an expert) … do the color and touch-up work on the "master image", resample as required and sharpen the resampled file for the specific output application.

Tacit wrote:

No, no, no. You do not color correct then resample. Remember, when you resample, you DISCARD information.

So? You don’t discard color information the way you discard detail information. A red stop sign, when resampled, will still be red.

Solid colors, yes. However, you DO lose color which causes posterized gradients.

If you do as you suggest–resample, then color correct–what happens when you want the large version’s color to match the small version’s color? You have to color correct again.

So? Short-cutting the process never helps.

A large image will, when resampled to a small image, maintain the same relative color balance.

I don’t know what your background is, but I have been doing this professionally since the days of Photoshop 1.0. My advice comes from this experience.

So is mine.

Regards,
Uni

Johan W. Elzenga wrote:

Uni wrote:

Pretty much everybody in the industry would disagree with this.
The generally accepted flow is to make the global corrections first, then the local corrections, then re-size and finally sharpen.

I still disagree. Resize mean resampling. You don’t make an image beautiful, then resample it.

Why not? "To make beautiful" means making color corrections. Resampling has nothing to do with that and doesn’t change any colors.

Oh, so if I resample an image down to one pixel, it will still contain a million different colors? I think not.

You should work and only work on the final size image. Remember, we cut down the trees from the forest, build houses with them, THEN paint them, so they look beautiful.

Regards,
Uni

There is

absolutely no difference between first color correcting, then resampling, or first resampling, then color correcting. It simply doesn’t matter.

"Uni" wrote in message

Johan W. Elzenga wrote:

Uni wrote:

Pretty much everybody in the industry would disagree with this.
The generally accepted flow is to make the global corrections first, then the local corrections, then re-size and finally sharpen.

I still disagree. Resize mean resampling. You don’t make an image beautiful, then resample it.

Why not? "To make beautiful" means making color corrections. Resampling has nothing to do with that and doesn’t change any colors.

Oh, so if I resample an image down to one pixel, it will still contain a million different colors? I think not.

You should work and only work on the final size image. Remember, we cut down the trees from the forest, build houses with them, THEN paint them, so they look beautiful.

Regards,
Uni

Both examples you just gave are totally irrelevant, and you know it. Nonetheless, you want to make the same corrections multiple times, have at it. You really don’t have to do it this way, but it *is* your prerogative to waste time as you wish.

—–= Posted via Newsfeeds.Com, Uncensored Usenet News =—– http://www.newsfeeds.com – The #1 Newsgroup Service in the World! —–== Over 100,000 Newsgroups – 19 Different Servers! =—–

Flycaster wrote:

"Uni" wrote in message

Johan W. Elzenga wrote:

Uni wrote:

Pretty much everybody in the industry would disagree with this.
The generally accepted flow is to make the global corrections first, then the local corrections, then re-size and finally sharpen.

I still disagree. Resize mean resampling. You don’t make an image beautiful, then resample it.

Why not? "To make beautiful" means making color corrections. Resampling has nothing to do with that and doesn’t change any colors.

Oh, so if I resample an image down to one pixel, it will still contain a million different colors? I think not.

You should work and only work on the final size image. Remember, we cut down the trees from the forest, build houses with them, THEN paint them, so they look beautiful.

Regards,
Uni

Both examples you just gave are totally irrelevant, and you know it. Nonetheless, you want to make the same corrections multiple times, have at it. You really don’t have to do it this way, but it *is* your prerogative to waste time as you wish.

I don’t see YOU or Tacit posting examples of your "excellence".

I’m not afraid to….

http://community.webshots.com/photo/38859810/118879751QdfdDW http://community.webshots.com/photo/42670345/42670526IIeqQz

Regards,
Uni

—–= Posted via Newsfeeds.Com, Uncensored Usenet News =—– http://www.newsfeeds.com – The #1 Newsgroup Service in the World! —–== Over 100,000 Newsgroups – 19 Different Servers! =—–

I don’t see YOU or Tacit posting examples of your "excellence".

http://www.xeromag.com/fvart3.html

—
Rude T-shirts for a rude age: http://www.villaintees.com Art, literature, shareware, polyamory, kink, and more:
http://www.xeromag.com/franklin.html

Tacit wrote:

I don’t see YOU or Tacit posting examples of your "excellence".

http://www.xeromag.com/fvart3.html

Eh, not bad for 2D’er. The last person who challenged me, well……

http://www.angelfire.com/empire/abpsp/uni_stuff/unibasketbal ltoss.gif

🙂

Uni

On 21 Feb 2004 03:07:20 GMT, (Tacit) wrote:

No, no, no. You do not color correct then resample. Remember, when you resample, you DISCARD information.

So? You don’t discard color information the way you discard detail information. A red stop sign, when resampled, will still be red.

If you do as you suggest–resample, then color correct–what happens when you want the large version’s color to match the small version’s color? You have to color correct again.

A large image will, when resampled to a small image, maintain the same relative color balance.

I don’t know what your background is, but I have been doing this professionally since the days of Photoshop 1.0. My advice comes from this experience.

The Uni is the pet troll from the PaintShopPro newsgroup.

It has a copy of PSP7 and Photoshop Elements 1, recommends Photoshop to those who can’t handle PSP and accuses those who have problems with PS to use illegal copies.

Wait for its answer to this post… or better ignore it forever 🙂

Michael

Uni wrote:

Why not? "To make beautiful" means making color corrections. Resampling has nothing to do with that and doesn’t change any colors.

Oh, so if I resample an image down to one pixel, it will still contain a million different colors? I think not.

And if you first resize to one pixel, you CAN retain all the colors? That is a bull shit argument. Of course you loose colors if you downsize to an image that has less pixels the the original has colors. But that will also happen if you work in any other order.

You should work and only work on the final size image.

Mosy experts agree that is the wrong way, because it means you have to do all the corrections all over again if you need another size of the same image. Good chance you won’t be able to get an exact duplicate with only a different size. But if you want to do extra work, be my guest.

Remember, we cut down the trees from the forest, build houses with them, THEN paint them, so they look beautiful.

Again a bull shit argument. You are not putting paint on the outside of an image, you change the colors of the ‘wood’ itself. In many cases, that is indeed what people do with trees as well. We FIRST impregnate the wood, and THEN we build a house out of it.

—
Johan W. Elzenga johan<<at>>johanfoto.nl Editor / Photographer http://www.johanfoto.nl/

"Uni" wrote in message

Flycaster wrote:

[snip]

Oh, so if I resample an image down to one pixel, it will still contain a million different colors? I think not.

You should work and only work on the final size image. Remember, we cut down the trees from the forest, build houses with them, THEN paint them, so they look beautiful.

Regards,
Uni

Both examples you just gave are totally irrelevant, and you know it. Nonetheless, you want to make the same corrections multiple times, have

at

it. You really don’t have to do it this way, but it *is* your

prerogative

to waste time as you wish.

I don’t see YOU or Tacit posting examples of your "excellence".
I’m not afraid to….

http://community.webshots.com/photo/38859810/118879751QdfdDW http://community.webshots.com/photo/42670345/42670526IIeqQz

Serving up "color balanced" point and shoot images as proof of your "excellence" does not help your case whatsoever. Face it, you’re wasting time here. Go on back to the PSP forums from whence you came and leave us in peace. Into the killfile you go.

—–= Posted via Newsfeeds.Com, Uncensored Usenet News =—– http://www.newsfeeds.com – The #1 Newsgroup Service in the World! —–== Over 100,000 Newsgroups – 19 Different Servers! =—–

The generally accepted flow is to make the global corrections first, then the local corrections, then re-size and finally sharpen.

I still disagree. Resize mean resampling. You don’t make an image beautiful, then resample it.

I don’t quite get this argument. If you make two copies of an image and do an extreme ‘colour balance then re-size’ on one and the same extreme ‘re-size then colour balance’ on the other aren’t the two images just slightly different? Look at the Levels dialog box for each image and you’ll see that the one that was ‘re-sized then colour balanced is all broken up and isn’t as continuous as the other image. Or even try pasting one result over the other whilst setting the overlay to ‘Difference’ and you’ll see that it’s not a perfect match.

—

Johan W. Elzenga wrote:

Uni wrote:

Why not? "To make beautiful" means making color corrections. Resampling has nothing to do with that and doesn’t change any colors.

Oh, so if I resample an image down to one pixel, it will still contain a million different colors? I think not.

And if you first resize to one pixel, you CAN retain all the colors? That is a bull shit argument. Of course you loose colors if you downsize to an image that has less pixels the the original has colors. But that will also happen if you work in any other order.

But, you can compensate for it later, IF you do the color correct then.

You should work and only work on the final size image.

Mosy experts agree that is the wrong way, because it means you have to do all the corrections all over again if you need another size of the same image. Good chance you won’t be able to get an exact duplicate with only a different size. But if you want to do extra work, be my guest.

That’s exactly it, these "experts" are just lazy and take short-cuts to save time, not quality.

Remember, we cut down the trees from the forest, build houses with them, THEN paint them, so they look beautiful.

Again a bull shit argument. You are not putting paint on the outside of an image, you change the colors of the ‘wood’ itself. In many cases, that is indeed what people do with trees as well. We FIRST impregnate the wood, and THEN we build a house out of it.

You people need to get a clue about digital image editing. You say to color correct the LARGE image, but DON’T sharpen it. What, does the contrast between pixels vanish when scaled down? Ha!!!

Regards,
Uni

Flycaster wrote:

"Uni" wrote in message

Flycaster wrote:

[snip]

Oh, so if I resample an image down to one pixel, it will still contain a million different colors? I think not.

You should work and only work on the final size image. Remember, we cut down the trees from the forest, build houses with them, THEN paint them, so they look beautiful.

Regards,
Uni

Both examples you just gave are totally irrelevant, and you know it. Nonetheless, you want to make the same corrections multiple times, have

at

it. You really don’t have to do it this way, but it *is* your

prerogative

to waste time as you wish.

I don’t see YOU or Tacit posting examples of your "excellence".
I’m not afraid to….

http://community.webshots.com/photo/38859810/118879751QdfdDW http://community.webshots.com/photo/42670345/42670526IIeqQz

Serving up "color balanced" point and shoot images as proof of your "excellence" does not help your case whatsoever. Face it, you’re wasting time here. Go on back to the PSP forums from whence you came and leave us in peace. Into the killfile you go.

Thank you. You are of little help yourself.

Uni

—–= Posted via Newsfeeds.Com, Uncensored Usenet News =—– http://www.newsfeeds.com – The #1 Newsgroup Service in the World! —–== Over 100,000 Newsgroups – 19 Different Servers! =—–

Xalinai wrote:

On 21 Feb 2004 03:07:20 GMT, (Tacit) wrote:

No, no, no. You do not color correct then resample. Remember, when you resample, you DISCARD information.

So? You don’t discard color information the way you discard detail information. A red stop sign, when resampled, will still be red.

If you do as you suggest–resample, then color correct–what happens when you want the large version’s color to match the small version’s color? You have to color correct again.

A large image will, when resampled to a small image, maintain the same relative color balance.

I don’t know what your background is, but I have been doing this professionally since the days of Photoshop 1.0. My advice comes from this experience.

The Uni is the pet troll from the PaintShopPro newsgroup.

You best get your little behind back there, so you can defend your beloved 19th century, 8-bit image editor.

Ha!

Regards,
Uni

It has a copy of PSP7 and Photoshop Elements 1, recommends Photoshop to those who can’t handle PSP and accuses those who have problems with PS to use illegal copies.

Wait for its answer to this post… or better ignore it forever 🙂
Michael

Uni wrote:

Most experts agree that is the wrong way, because it means you have to do all the corrections all over again if you need another size of the same image. Good chance you won’t be able to get an exact duplicate with only a different size. But if you want to do extra work, be my guest.

That’s exactly it, these "experts" are just lazy and take short-cuts to save time, not quality.

There is no quality difference, because it doesn’t matter if you first color correct and then resize, or vice versa. Only in your mind there is. But quality also means consistency. If a client wants two or more sizes of the same photo, he or she expects identical images apart from the different size. I’m not saying that is impossible if you color correct each size separately, but it’s simply a waste of time and consequently money. Who’s going to pay? Your client, or you?

Again a bull shit argument. You are not putting paint on the outside of an image, you change the colors of the ‘wood’ itself. In many cases, that is indeed what people do with trees as well. We FIRST impregnate the wood, and THEN we build a house out of it.

You people need to get a clue about digital image editing. You say to color correct the LARGE image, but DON’T sharpen it. What, does the contrast between pixels vanish when scaled down? Ha!!!

In this discussion there’s only one who clearly needs to understand digital image editting, and that person calls himself "Uni". The reason you need to sharpen after resizing is because the amount of sharpening depends on the size of the image. Also, resizing usually causes some blurring again. The only way to get the correct amount of sharpening is to do it after resizing. If you do it on the large image, it becomes a ‘hit and miss’ affair, that’s all.

—
Johan W. Elzenga johan<<at>>johanfoto.nl Editor / Photographer http://www.johanfoto.nl/

Johan, save yourself some time. Do a google NG search on this kid and you’ll quickly see what I mean..

—–= Posted via Newsfeeds.Com, Uncensored Usenet News =—– http://www.newsfeeds.com – The #1 Newsgroup Service in the World! —–== Over 100,000 Newsgroups – 19 Different Servers! =—–

Flycaster wrote:

Johan, save yourself some time. Do a google NG search on this kid and you’ll quickly see what I mean..

I don’t see any examples of your work, Flycaster. Or, are you ashamed to post a link, so we can’t see your superior image editing skills?

Uni

—–= Posted via Newsfeeds.Com, Uncensored Usenet News =—– http://www.newsfeeds.com – The #1 Newsgroup Service in the World! —–== Over 100,000 Newsgroups – 19 Different Servers! =—–

Johan W. Elzenga wrote:

Uni wrote:

Most experts agree that is the wrong way, because it means you have to do all the corrections all over again if you need another size of the same image. Good chance you won’t be able to get an exact duplicate with only a different size. But if you want to do extra work, be my guest.

That’s exactly it, these "experts" are just lazy and take short-cuts to save time, not quality.

There is no quality difference, because it doesn’t matter if you first color correct and then resize, or vice versa. Only in your mind there is. But quality also means consistency. If a client wants two or more sizes of the same photo, he or she expects identical images apart from the different size. I’m not saying that is impossible if you color correct each size separately, but it’s simply a waste of time and consequently money. Who’s going to pay? Your client, or you?

Again a bull shit argument. You are not putting paint on the outside of an image, you change the colors of the ‘wood’ itself. In many cases, that is indeed what people do with trees as well. We FIRST impregnate the wood, and THEN we build a house out of it.

You people need to get a clue about digital image editing. You say to color correct the LARGE image, but DON’T sharpen it. What, does the contrast between pixels vanish when scaled down? Ha!!!

In this discussion there’s only one who clearly needs to understand digital image editting, and that person calls himself "Uni". The reason you need to sharpen after resizing is because the amount of sharpening depends on the size of the image.

Oh, I see, now it depend on the "size" of the image. This is a new one. By the way, do you sell your photography scaled down, and then tell your customers they aren’t missing a thing, since you didn’t provide the full size images? Furthermore, sharpening distorts. It also falsely adds colors which were not present in the original image.

Also, resizing usually causes some

blurring again. The only way to get the correct amount of sharpening is to do it after resizing. If you do it on the large image, it becomes a ‘hit and miss’ affair, that’s all.

So, resizing causes blurring of colors.

When you feel up to it, post one of your unedited photos somewhere. I’ll gladly take the challenge from a "chief editor of two Dutch photography magazines and a magazine about digital imaging."

Regards,
Uni

Read this Bruce Fraser article on sharpening.

Powell Hargrave, Nanaimo, B.C. Canada –
Digital Photography Page http://members.shaw.ca/hargravep ****** Remove NOSPAM to Reply ******

Powell Hargrave wrote:

Read this Bruce Fraser article on sharpening.

Would be helpful if you provided a link.

Thanks!

Uni

Powell Hargrave, Nanaimo, B.C. Canada –
Digital Photography Page http://members.shaw.ca/hargravep ****** Remove NOSPAM to Reply ******

On Sun, 22 Feb 2004 16:49:41 -0500, Uni
wrote:

Xalinai wrote:

On 21 Feb 2004 03:07:20 GMT, (Tacit) wrote:

No, no, no. You do not color correct then resample. Remember, when you resample, you DISCARD information.

So? You don’t discard color information the way you discard detail information. A red stop sign, when resampled, will still be red.

If you do as you suggest–resample, then color correct–what happens when you want the large version’s color to match the small version’s color? You have to color correct again.

A large image will, when resampled to a small image, maintain the same relative color balance.

I don’t know what your background is, but I have been doing this professionally since the days of Photoshop 1.0. My advice comes from this experience.

The Uni is the pet troll from the PaintShopPro newsgroup.

You best get your little behind back there, so you can defend your beloved 19th century, 8-bit image editor.

Down! Go to your basket!

BTW: We’re in the 21st century now and PSP and PS have their roots 20est century software. Or do you have a PSP Version running on Babbage’s Analytical Engine?

Michael

Uni wrote:

Oh, I see, now it depend on the "size" of the image. This is a new one.

I’m not surprised that is a new one to you. 😉

Furthermore, sharpening distorts. It also falsely adds
colors which were not present in the original image.

Not if you sharpen luminosity only, but that method is probably new to you as well. Sharpen, go to menu "Fade Unsharp Mask" and set the Mode from Normal to Luminosity.

Also, resizing usually causes some blurring again. The only way to get the correct amount of sharpening is to do it after resizing. If you do it on the large image, it becomes a ‘hit and miss’ affair, that’s all.

So, resizing causes blurring of colors.

Huh? who said that? You can’t blur colors. You can only blur detail. I guess it’s time to stop this discussion. I couldn’t care less what method you use and I don’t want to waste any more time on it.

—
Johan W. Elzenga johan<<at>>johanfoto.nl Editor / Photographer http://www.johanfoto.nl/

Read this Bruce Fraser article on sharpening.

Would be helpful if you provided a link.

This one maybe?
http://www.creativepro.com/story/feature/11242.html

Johan W. Elzenga wrote:

Uni wrote:

Oh, I see, now it depend on the "size" of the image. This is a new one.

I’m not surprised that is a new one to you. 😉

Furthermore, sharpening distorts. It also falsely adds
colors which were not present in the original image.

Not if you sharpen luminosity only, but that method is probably new to you as well. Sharpen, go to menu "Fade Unsharp Mask" and set the Mode from Normal to Luminosity.

If you understood what is happening with sharpening, you’d agree that it does distort. It has to, there is no other way.

Also, resizing usually causes some blurring again. The only way to get the correct amount of sharpening is to do it after resizing. If you do it on the large image, it becomes a ‘hit and miss’ affair, that’s all.

So, resizing causes blurring of colors.

Huh? who said that? You can’t blur colors. You can only blur detail.

Draw two rectangles, touching each other, with different solid colors. We have two unique colors. Now blur the entire image and, count the number of colors. Wow, we managed to blend or blur two colors into many!

I guess it’s time to stop this discussion. I couldn’t care less what method you use and I don’t want to waste any more time on it.

So be it.

Regards,
Uni

"Uni" wrote in message

Flycaster wrote:

Johan, save yourself some time. Do a google NG search on this kid and you’ll quickly see what I mean..

I don’t see any examples of your work, Flycaster. Or, are you ashamed to post a link, so we can’t see your superior image editing skills?

Jesus, not only are you clueless about efficient workflows and how color works, but you now appear to be dumber than mud. He plonked you pal, just like everyone else, so he can’t even read this prissy little reply. Got it?

You managed to piss off a whole crowd of knowledgeable and helpful regulars, all in the same thread. That’s got to be some kind of record, even for Usenet.

Stop posting links to your self worshiping website(s) and just go away.

—–= Posted via Newsfeeds.Com, Uncensored Usenet News =—– http://www.newsfeeds.com – The #1 Newsgroup Service in the World! —–== Over 100,000 Newsgroups – 19 Different Servers! =—–

joe wrote:

"Uni" wrote in message

Flycaster wrote:

Johan, save yourself some time. Do a google NG search on this kid and you’ll quickly see what I mean..

I don’t see any examples of your work, Flycaster. Or, are you ashamed to post a link, so we can’t see your superior image editing skills?

Jesus,

You rang?

http://community.webshots.com/photo/38859810/118879751QdfdDW http://community.webshots.com/photo/42670345/42670526IIeqQz

worshiping website(s)

Good idea.

🙂

Uni

I took a printer test image and opened it in Photoshop.

I added 4 color sampler points. Colors in RGB were:
#1 – 225 0 1
#2 – 0 0 251
#3 – 226 1 253
#4 – 144 144 144

I added a hue/sat layer with Hue set to +25.

Colors changed to:
#1 – 225 92 0
#2 – 104 0 251
#3 – 253 1 174
#4 – 144 144 144

Resized the image to 50%. Colors did not change at all.

Then I took the original image with color samplers, and resized it to 50% without the hue/sat layer. It had the same colors as the original. #1 – 225 0 1
#2 – 0 0 251
#3 – 226 1 253
#4 – 144 144 144

I added a hue/sat layer with the same +25 hue setting. It had the same colors as the other version with the hue/sat layer.
#1 – 225 92 0
#2 – 104 0 251
#3 – 253 1 174
#4 – 144 144 144

So there’s no difference in doing the color change before or after you resize the image. It just depends on if you want to have to do the correction again later on or not.

"Uni" wrote in message

You rang?
http://community.webshots.com/photo/38859810/118879751QdfdDW http://community.webshots.com/photo/42670345/42670526IIeqQz

So you are comparing shots showing auto-colorcast and auto-color balance from Paintshop Pro to your manual adjustments — in this Photoshop Usenet group. Are you suggesting these auto functions in Paintshop Pro are that poor that you have to do manual corrections?

"phucaduck" wrote in message

I took a printer test image and opened it in Photoshop.
I added 4 color sampler points. Colors in RGB were:
#1 – 225 0 1
#2 – 0 0 251
#3 – 226 1 253
#4 – 144 144 144

I added a hue/sat layer with Hue set to +25.

Colors changed to:
#1 – 225 92 0
#2 – 104 0 251
#3 – 253 1 174
#4 – 144 144 144

Resized the image to 50%. Colors did not change at all.

Then I took the original image with color samplers, and resized it to 50% without the hue/sat layer. It had the same colors as the original. #1 – 225 0 1
#2 – 0 0 251
#3 – 226 1 253
#4 – 144 144 144

I added a hue/sat layer with the same +25 hue setting. It had the same colors as the other version with the hue/sat layer.
#1 – 225 92 0
#2 – 104 0 251
#3 – 253 1 174
#4 – 144 144 144

So there’s no difference in doing the color change before or after you resize the image. It just depends on if you want to have to do the correction again later on or not.

Good, I like your approach, let’s figure this out so we know instead of relying on educated guesses. I took a blank RGB image, selected half the image and filled it with red (255/0/0), then inverted the selection and filled it with green (0/255/0). Then I blew up the magnification to 1600 % and took a single pixel point sample on each side. The changeover from red to green occurred in exactly one pixel (in other words no anomalies occurred due to the selection). When I resized it to 88% (this was just an arbitrary choice) the single pixel line between the two changed to 229/26/0. Changing the image size to 88% again created a noticable line at119/136/0, the third time was 150/105/0 and I assume this would continue to degrade the more times I did it. I would also assume this would have a profound effect on gradiated colours. Anyone?

~Doc

"Dr. J. Smith" wrote in message

"phucaduck" wrote in message

SNIP

So there’s no difference in doing the color change before or after you resize the image. It just depends on if you want to have to do the correction again later on or not.

Good, I like your approach, let’s figure this out so we know instead of relying on educated guesses.

I’m not sure what you intend to figure out, as the above seems clear enough.

I took a blank RGB image, selected half the
image and filled it with red (255/0/0), then inverted the selection and filled it with green (0/255/0). Then I blew up the magnification to 1600 % and took a single pixel
point sample on each side. The changeover from red to
green occurred in exactly one pixel (in other words no
anomalies occurred due to the selection). When I
resized it to 88% (this was just an arbitrary choice) the single pixel line between the two changed to 229/26/0.

I’m assuming you used Bi-cubic interpolation (which differs between some PS versions). I also have to assume you used a Color Setting of Blending with Gamma 1 (it’ll make a big difference for some color combinations). The value you get on the transition, depends on the original image size as well. Do the same experiment with an original image, one pixel smaller in size (e.g. crop your original before resizing). All you can conclude is that the transition value is a weighted blend of the two colors being mixed.

Changing the image size to 88% again created a noticable line at119/136/0, the third time was 150/105/0 and I assume this would continue to degrade the more times I did it.

Why do you call it a degradation? You just mix different amounts of colors. The detail is imaged smaller than 1 pixel, so neighboring pixels are weighted in.

I would also assume this would have a profound effect on gradiated colours. Anyone?

Same difference 😉

Try downsizing the following example to e.g. 10% (to make it very clear): http://www.xs4all.nl/~bvdwolf/main/downloads/Rings.gif (1.27MB) That will demonstrate aliasing artifacts, caused by the inability to accurately render detail smaller than one pixel. Simple blending will cause non-existing detail to appear.

Bart

Yeah, it seems that we can investigate this logically and methodically instead of just insulting each other. What a concept for this forum…

Here’s the test image I’m using:
http://www.photo-i.co.uk/Reviews/printers/images/Test.jpg

I’m placing color samplers at the following coordinates:

#1 – x: 830 y: 432 (on the baby’s forehead)
reads – R: 233 G: 198 B: 176

#2 – x: 1750 y: 1250 (on the arm of the soldier)
reads – R: 226 G: 22 B: 31

#3 – x: 780 y: 2080 (on a green leaf in the flower image) reads – R: 83 G: 99 B: 72

I then added the hue-sat adjustment layer with +25 hue and +25 sat.

This gave readings of:
#1 – R: 242 G: 227 B: 166
#2 – R: 248 G: 92 B: 0
#3 – R: 67 G: 103 B: 67

I did two sets of tests. One, I added the hue/sat layer and then did all the resizing. For the second set, I resized and then added the same hue/sat layer each time (deleting it before each subsequent resizing).

Here are the results (for each color sampler, the first set of RGB values are the ‘change color and then resize’ set, and the second set are the ‘resize and then change color’ set)

Sample #1
150% 242 227 166 242 227 166
225% 243 228 167 243 228 167
50% 242 227 166 242 227 166
25% 242 227 166 242 227 166

Sample #2
150% 248 92 0 248 92 0
225% 249 92 0 249 92 0
50% 250 93 0 250 93 0
25% 245 91 0 245 91 0

Sample #3
150% 68 104 68 68 104 68
225% 67 102 68 67 102 68
50% 61 98 62 61 98 62
25% 57 93 57 57 93 57

It looks like it doesn’t matter a bit if you change the color before or after a resize. Because you get the same results either way.

So, it seems that Uni is mistaken.