Author | Thread |
|
03/16/2009 10:27:27 PM · #1 |
I've been trying to use it on this photo but not sure I'm getting the effect right.
I resized to 1500px, then applied sharpening 2x, and resized to 720px. Am I missing something? |
|
|
03/16/2009 10:42:58 PM · #2 |
It looks like the shot has a lot of detail to start with, try sharpening just the once then resizing. |
|
|
03/16/2009 10:58:23 PM · #3 |
1. Initial resize should be exactly twice the size the web image will be, and you should use plain "Bicubic", not "Bicubic Sharper", as your algorithm.
2. In this case, that would be 1440 pixels. After resizing, flatten image, then dupe BG layer and sharpen. Call it "sharp1"
3. Now dupe "sharp1" and call it "sharp2" and sharpen THAT.
4. Now dupe "sharp2", call it "sharp3", and sharpen THAT.
5. Now resize to 720 pixels, again using plain "bicubic" algorithm. Set view to "actual pixels".
6. At this point the top layer is sharpened 3x; if you turn it off, you're looking at the 2x layer - turn THAT one off and you're looking at the 1x layer. Choose the best one, or possibly one more than the best one and fade that top layer opacity until it looks right.
I suspect with this image 1x or maybe 1.5 x will be sufficient. Remember, if processing from RAW, do NOT sharpen image in RAW processing. If working in JPG, set camera to zero sharpening in-camera. This parts important, this should be the only sharpening you do for web viewing.
R.
Message edited by author 2009-03-16 23:06:55.
|
|
|
03/17/2009 06:53:36 AM · #4 |
Bear - where do you get/did you get that it should be exactly twice the size you want for the web on the first resize?
"Sharpening for the Web—Marc Adamus
- Take full res image and reduce size to between 1000 and 1300 pixels. As much as 1300 for detailed forest scenes and such, 1000 for rock and more simple images. Apply "Sharpen" filter (easy enough right?) twice or even three times at 1300-1600 pixels.
(stuff in between not copied)
- Play around with this technique and make it work for you. Every image is different. Sometimes I'll apply a sharpen filter 3 times at 1500 pixels and then reduce for an even finer sharpness, but other times it has too much of an adverse effect on color and resaturation is necessary. Sometimes I'll just lasso an area and apply the filter more heavily in one section before reducing size. Just experiment!"
Above taken from here.
I have been using the method as it is written above with experimentation and found the results to be quite good a lot of times. I will try on my shot for the open challenges at the twice the web size tonight.
Message edited by author 2009-03-17 07:07:27. |
|
|
03/17/2009 07:22:07 AM · #5 |
Rob and Bear- thanks!
This is what I got by following Bear's method with 1.5x sharpening
Not sure if it's any better than before.
Maybe too much detail here and the sun flare doesn't help? |
|
|
03/17/2009 07:27:51 AM · #6 |
Just remember that if you're doing this for an open challenge under basic editing, you cannot duplicate the background layers and sharpen them as Bear suggests for your challenge entry. You CAN do it to see how much sharpening is needed, but when it comes to the challenge entry, just resize and sharpen your base image. (Bear will correct me if I'm wrong!) |
|
|
03/17/2009 07:38:21 AM · #7 |
IMO...the second attempt is better - more contrast, less of an over-sharpened look in general. |
|
|
03/17/2009 08:08:18 AM · #8 |
Originally posted by CEJ: IMO...the second attempt is better - more contrast, less of an over-sharpened look in general. |
Agreed. Though I think the degree of saturation is more problematic. |
|
|
03/17/2009 08:11:28 AM · #9 |
Originally posted by Melethia: Just remember that if you're doing this for an open challenge under basic editing, you cannot duplicate the background layers and sharpen them as Bear suggests for your challenge entry. You CAN do it to see how much sharpening is needed, but when it comes to the challenge entry, just resize and sharpen your base image. (Bear will correct me if Im wrong!) |
This is good to know!!! |
|
|
03/17/2009 08:39:09 AM · #10 |
Make sure I'm right - but I think duplicating layers introduces *new pixel data* which working on a base layer or with adjustment layers does not. It's the new pixel data that needs to be avoided for basic, I think. |
|
|
03/17/2009 08:45:12 AM · #11 |
Originally posted by Melethia: Make sure I'm right - but I think duplicating layers introduces *new pixel data* which working on a base layer or with adjustment layers does not. It's the new pixel data that needs to be avoided for basic, I think. |
I think you're right
-"use only Adjustment Layers (or their equivalent). An Adjustment Layer is a special type of layer containing no image data that lets you experiment with color and tonal adjustments without permanently modifying the pixels. Adjustment Layers must be applied in Normal mode".
Sharpening by duplicating layers would involve layers containing image data. Right?
Message edited by author 2009-03-17 08:45:53. |
|
|
03/17/2009 08:57:41 AM · #12 |
Originally posted by smichener: Originally posted by Melethia: Make sure I'm right - but I think duplicating layers introduces *new pixel data* which working on a base layer or with adjustment layers does not. It's the new pixel data that needs to be avoided for basic, I think. |
I think you're right
-"use only Adjustment Layers (or their equivalent). An Adjustment Layer is a special type of layer containing no image data that lets you experiment with color and tonal adjustments without permanently modifying the pixels. Adjustment Layers must be applied in Normal mode".
Sharpening by duplicating layers would involve layers containing image data. Right? |
Yep, it would. What you can do to incorporate Bear's trial and error method is work on your file, make a duplicate FILE at the time you're ready to sharpen, do the duplicate LAYER bit with successive sharpening on that one and see what you like. Then delete the duplicate FILE (so you don't get confused) and go back and sharpen your basic layer to the level you determined. |
|
|
03/17/2009 09:00:07 AM · #13 |
With regard to Basic and layer duplication, some time ago this was discussed (both publicly and by SC) and the ruling was that it is permissible to duplicate a layer to save an edited state prior to further editing. The key is whether more than one layer participates in the final image.
In Robert's example, only one of the final layers will be used for the final image *unless* the opacity or fill sliders are used. If only one layer is visible during the final save, it should be legal, since you are just choosing one result our of several "trials."
Disclaimer:
This is based on my interpretation of past rulings, and should be confirmed by the SC. |
|
|
03/17/2009 09:20:19 AM · #14 |
Originally posted by kirbic: With regard to Basic and layer duplication, some time ago this was discussed (both publicly and by SC) and the ruling was that it is permissible to duplicate a layer to save an edited state prior to further editing. The key is whether more than one layer participates in the final image.
In Robert's example, only one of the final layers will be used for the final image *unless* the opacity or fill sliders are used. If only one layer is visible during the final save, it should be legal, since you are just choosing one result our of several "trials."
Disclaimer:
This is based on my interpretation of past rulings, and should be confirmed by the SC. |
if this is legal or not, i don't think the SC would ever be able to know if there was more than 1 layer with pixels present but turned off when you flatten.
to be legal, the sharpening should be done on the background layer only.
in bears example, only 1 layer would be used, but i even that is against the basic editing rules, but i think this is un-enforcable (is that a word??) as the final result would be the same!
|
|
|
03/17/2009 09:24:13 AM · #15 |
Originally posted by kirbic: With regard to Basic and layer duplication, some time ago this was discussed (both publicly and by SC) and the ruling was that it is permissible to duplicate a layer to save an edited state prior to further editing. The key is whether more than one layer participates in the final image.
In Robert's example, only one of the final layers will be used for the final image *unless* the opacity or fill sliders are used. If only one layer is visible during the final save, it should be legal, since you are just choosing one result our of several "trials."
Disclaimer:
This is based on my interpretation of past rulings, and should be confirmed by the SC. |
In basic editing, I'll run through the workflow as described, with layers, note the result desired, then back off to the non-sharpened state and apply the requisite sharpening in multiple passes on the same layer. If I find that my desired final result involves the ultimate layer being faded, then I'll do an "edit>fade sharpening" on the final pass, which is the equivalent of doing it in a separate layer and fading the layer. In this way I'll be adhering to the letter of the rules, and the result will be the same.
The multiple layers sharpening approach just allows one to critically examine the results more efficiently. One could, for example, do it without layers by duplicating the image twice and saving as two new file names, sharpening the respective images once, twice, and three times in the respective images, and toggling between them to find the best one for further work, but that's a bunch of extra steps.
R.
|
|
|
03/17/2009 09:28:34 AM · #16 |
Originally posted by CEJ: Bear - where do you get/did you get that it should be exactly twice the size you want for the web on the first resize? |
It's just from personal experience; I have had a couple images where it seems to make a difference, that the final resize of the sharpened master worked noticeably better when it was exactly a 50% resize, as opposed to, say, 61%. Might just be an illusion, but it seems to work for me, so as a matter of course I do the first resize to 1280 or 1440 pixels, in basic and advanced respectively.
As an aside, I can't wait 'til they finally wise up and make basic 720 pixels to match advanced, so I can stop having to use two separate sets of actions in my processing, LOL.
R.
|
|
|
03/17/2009 09:50:52 AM · #17 |
Originally posted by Bear_Music: Originally posted by CEJ: Bear - where do you get/did you get that it should be exactly twice the size you want for the web on the first resize? |
It's just from personal experience; I have had a couple images where it seems to make a difference, that the final resize of the sharpened master worked noticeably better when it was exactly a 50% resize...
|
FWIW, the reason it works better with exactly 50% as the final resize is because it means that precisely four pixels (a two-by-two pixel block) are used for each single pixel in the final image. |
|
|
03/17/2009 09:58:33 AM · #18 |
Originally posted by kirbic: FWIW, the reason it works better with exactly 50% as the final resize is because it means that precisely four pixels (a two-by-two pixel block) are used for each single pixel in the final image. |
I was gonna SAY that was my theory, but I didn't want to sound stupid if it was wrong :-)
R.
|
|
|
03/17/2009 10:48:12 AM · #19 |
Ahh...makes sense and now feel stupid that I did not make the connection. I do like the method for a lot of shots and will definitely use the size ratio tweak this evening. |
|
|
03/17/2009 11:09:17 AM · #20 |
Pretty interesting. And its the first time I realise that Adapus Sharpenning is a procedure. |
|
|
03/17/2009 11:56:47 AM · #21 |
yeah, that Adapus technique is ok, but most use the Adamus version. it seems to work a little better. :)
Message edited by author 2009-03-17 14:55:12. |
|
|
03/17/2009 12:26:18 PM · #22 |
I think you can only resize once for challenges anyways. At least of I remember correctly I saw one get DQ'd for resizing multiple times about 6 months ago. |
|
|
03/17/2009 12:31:18 PM · #23 |
Originally posted by Patrick_R: I think you can only resize once for challenges anyways. At least of I remember correctly I saw one get DQ'd for resizing multiple times about 6 months ago. |
i believe that is only the minimal editing rule set!
|
|
|
03/17/2009 12:42:01 PM · #24 |
Originally posted by dainmcgowan: Originally posted by Patrick_R: I think you can only resize once for challenges anyways. At least of I remember correctly I saw one get DQ'd for resizing multiple times about 6 months ago. |
i believe that is only the minimal editing rule set! |
Correct.
R.
|
|
|
03/17/2009 12:52:35 PM · #25 |
i'm glad that particular is finally recognized....
i always have, and always will edit on a duplicate of the original layer.
way - back the rules seemed to point at an exact replicate of the original layer - as a wroking layer - was illegal. i've always found that nit-pick silly.
BTW - i notice you're not an SC anymore... :(
Originally posted by kirbic: With regard to Basic and layer duplication, some time ago this was discussed (both publicly and by SC) and the ruling was that it is permissible to duplicate a layer to save an edited state prior to further editing. The key is whether more than one layer participates in the final image.
In Robert's example, only one of the final layers will be used for the final image *unless* the opacity or fill sliders are used. If only one layer is visible during the final save, it should be legal, since you are just choosing one result our of several "trials."
Disclaimer:
This is based on my interpretation of past rulings, and should be confirmed by the SC. |
|
|
Home -
Challenges -
Community -
League -
Photos -
Cameras -
Lenses -
Learn -
Help -
Terms of Use -
Privacy -
Top ^
DPChallenge, and website content and design, Copyright © 2001-2025 Challenging Technologies, LLC.
All digital photo copyrights belong to the photographers and may not be used without permission.
Current Server Time: 08/01/2025 06:40:03 AM EDT.