Posted by Scott Buckel on March 17, 2006 at 3:48 PM
I have been wondering about this. I have not been able to make good jpgs from some of my images. Recently, I posted an image that really needs to have lots of detail, and the posted jpg does not have it. SO I completely agree with the "critiquers".
The part that does bother me is that the image has a huge amount of detail. The image has been printed to 12x18 and is in my office here at work and people are amazed that every single feather shows up in detail. The usual comment about the print is wow look at the detail in the feathers.
I usually just size the image down from my master file using using the image size dialogue and then go to save for web. Am I doing something wrong?
Did you sharpen after resizing? If not, you need to. Sizing an image, especially down, will blur detail. You have to sharpen it to get the detail back.
There are several articles in the NPN Digital Archives that may help you with workflow.
Posted by Scott Buckel on March 17, 2006 at 6:24 PM
Bob - I did sharpen. The detail was gone and sharpening did not seem to do very much. Thanks for hte reference to the archives, I'll have to take a look.
If you sharpen via USM, it can sometimes be tricky to set the radius right. For those cases, the High Pass method gives better visual control, IMHO. (Google "high pass sharpen" and you should get some good info) You can easily see the amount of detail being enchanced (by adding local contrast) in the grey image that you get after applying the High Pass filter. If you have found the radius that keeps the most detail, you can either proceed with the High Pass method or memorize the radius value and use it in the USM dialogue.
good luck!
Posted by Roger Rouch on March 18, 2006 at 8:56 AM
I am still working on my web posting techniques, but might at least make some suggestions that you might try. I'm sure there are some of the experts out there who would offer better advice.
I ususally don't apply any USM before resizing. I'll flatten the image layers and then resize without using the bicubic sharpener. Bicubic sharpener works fine in some cases but sometimes it just doesn't match what I had hoped it to look like. Especially if I'm going from a pretty large size to much smaller. After resizing I apply USM. If the photo has a lot of fine details it seems like a smaller radius, say 0.3, works better than a larger radius. Then I convert to sRBG. And save for web. Since my original was probably at 300 dpi, I usually will have to slightly reduce the file size in the save for web window to match posting guidlelines. This causes a little bit more compression, so I will close the file and then re-open the saved JPG. There is sometimes some additional fine tuning of USM to get the effect I'd like.
Like a lot of PS things, it seems like there are a few paths to get to the same end. But this is what I'm doing and works OK, though there may be better ways. It's something you can play with a little to see how it works. I think there is a little bit of art and practice to the use of USM and the radius and amount are not the same among different images. I have a special "sharpen edges only" action from a Barry Haynes CD. With some images this helps improve sharpness without observable grain or halos, and can be sharpened further for fine tuning after it is applied.
For web sized images, I typically re-size as normal (bicubic or bicubic sharp - depending on the image), then use the sharpen tool (not the filter) set at 10% to 30% opacity with a feathered edge. Similar to dodging and burning, I just drag the brush over areas of high detail to pop them out.
Also, photoshop supports JPEG compression masks. Just like layer masks, you can create an alpha channel to define where you want more or less compression. But, I've found them only useful when you're pushing a size limit and need to squeeze a bit more detail while staying within a specific file size.
USM is best for full sized images, not so good for web sized versions. Personally, I like Photoshop's RAW settings for controlling sharpness and noise. Then if needed I selectively sharpen areas that are slightly out of focus.
EDIT: Oh, and if you're really pushing your size limit due to lots of detail then use "Save for Web" instead of "Save As". The web route will remove EXIF data, thumbnails and other metadata that cause file bloat.
Discussion
5 repliesThe part that does bother me is that the image has a huge amount of detail. The image has been printed to 12x18 and is in my office here at work and people are amazed that every single feather shows up in detail. The usual comment about the print is wow look at the detail in the feathers.
I usually just size the image down from my master file using using the image size dialogue and then go to save for web. Am I doing something wrong?
Thanks,
There are several articles in the NPN Digital Archives that may help you with workflow.
Thanks,
good luck!
I ususally don't apply any USM before resizing. I'll flatten the image layers and then resize without using the bicubic sharpener. Bicubic sharpener works fine in some cases but sometimes it just doesn't match what I had hoped it to look like. Especially if I'm going from a pretty large size to much smaller. After resizing I apply USM. If the photo has a lot of fine details it seems like a smaller radius, say 0.3, works better than a larger radius. Then I convert to sRBG. And save for web. Since my original was probably at 300 dpi, I usually will have to slightly reduce the file size in the save for web window to match posting guidlelines. This causes a little bit more compression, so I will close the file and then re-open the saved JPG. There is sometimes some additional fine tuning of USM to get the effect I'd like.
Like a lot of PS things, it seems like there are a few paths to get to the same end. But this is what I'm doing and works OK, though there may be better ways. It's something you can play with a little to see how it works. I think there is a little bit of art and practice to the use of USM and the radius and amount are not the same among different images. I have a special "sharpen edges only" action from a Barry Haynes CD. With some images this helps improve sharpness without observable grain or halos, and can be sharpened further for fine tuning after it is applied.
Also, photoshop supports JPEG compression masks. Just like layer masks, you can create an alpha channel to define where you want more or less compression. But, I've found them only useful when you're pushing a size limit and need to squeeze a bit more detail while staying within a specific file size.
USM is best for full sized images, not so good for web sized versions. Personally, I like Photoshop's RAW settings for controlling sharpness and noise. Then if needed I selectively sharpen areas that are slightly out of focus.
EDIT: Oh, and if you're really pushing your size limit due to lots of detail then use "Save for Web" instead of "Save As". The web route will remove EXIF data, thumbnails and other metadata that cause file bloat.
Cheers,
-Andy