Eljae
Jr. Member
Posts: 68
|
|
« on: August 24, 2009, 01:45:59 PM » |
|
Hi everyone, I'm Eljae (pronounced Lj) am new to the forum.
I have a question about the Qimage output sharpening for printing. I currently use a selective sharpening technique in Photoshop. I photograph people so there are parts of photos that I intentionally leave minimally sharpened, and other parts that need to be very sharp. I ncrease the view to 50-100% depending on how close up I have shot, to make sure I have not introduced sharpening artifacts.
At this point this is as viewed on my Samsung SyncMaster monitor, so for LCD viewing it is very sharp where I want it.
Should I turn off Qimage sharpening at the time of printing, or does Qimage still need to apply output sharpening?
Thanks.
Eljae
|
|
|
Logged
|
|
|
|
Jeff
|
|
« Reply #1 on: August 24, 2009, 04:14:27 PM » |
|
I use Lab printers and always leave Qimage at default 5 additional sharpening and prints always look ok.
If you do your own printing, print a couple of tests one without additional sharpening and one with and let us know what is best in your opinion.
Jeff
|
|
|
Logged
|
Grumpy
|
|
|
Terry-M
|
|
« Reply #2 on: August 24, 2009, 04:14:55 PM » |
|
Hi Eljae, welcome to the forum. Should I turn off Qimage sharpening at the time of printing, or does Qimage still need to apply output sharpening? Qimage Help says: Qimage offers the ability to apply different levels of unsharp mask to the final print. This unsharp mask is applied after all interpolation has been performed (after your print has been resampled to the DPI of your printer). Since this setting affects the sharpness of your prints, a setting should be used that makes your final prints match what is displayed on your monitor with respect to sharpness. The type of printer used and the print driver version can affect apparent sharpness of prints. It is best to leave this setting at the default slider position unless you prefer softer or sharper printed images. I think the short answer to your question is no, leave Smart sharpening on but you may need to adjust the slider, up or down to suit your taste. Once set, you probably won't need to adjust it again. You will need to experiment but start with the default value of 5. Terry.
|
|
|
Logged
|
|
|
|
Terry-M
|
|
« Reply #3 on: August 24, 2009, 07:25:26 PM » |
|
ElJay, excuse me coming back again, but on re-reading your post I have questions. But first, some ancient history. When, as a teenager, and learning about physics and photography, my first bellows-front 120 film camera had no focussing aid other than a distance scale and a depth of field scale. Thus, from an early age, focus point and depth of field were known to be critical and important as part of a photographic composition. So, when I read: I currently use a selective sharpening technique in Photoshop. I photograph people so there are parts of photos that I intentionally leave minimally sharpened, and other parts that need to be very sharp. This puzzled me because unless you specifically chose an aperture that throws parts of your shot out of the Depth Of Field for that lens, aperture, and distance to subject, how do you leave some parts minimally sharp, and other parts very sharp? Even landscapes, the eye see what the eye sees. It never sees a bird in a tree at 5000 metres when the initial focus was at 250 metres. I reasonably understand what digital sharpening does and in some senses it creates an illusion, but I would ask: if the image leaves the lens and is not sharp when it is recorded due to various reasons, can we really put that image back to the sharpness, or make it “very sharp” where it should have been if the focus was correct and the camera movement removed, etc? Sorry to be a bother, but as long as you brought it up... Terry.
|
|
|
Logged
|
|
|
|
Eljae
Jr. Member
Posts: 68
|
|
« Reply #4 on: August 27, 2009, 12:47:08 PM » |
|
Terry, sorry for the lag getting back to you. I am not a very technical person, so anyone here please feel free to correct me if I am in error, but I'll do my best to explain what I was talking about regarding partial sharpening.
A Bayer pattern image is a product of the imaging device of digital cameras, and the final image from it is basically the RAW output. Each pixel of an image is filtered to record only one of three colors, i.e. RGB, and a full-color image is produced by in-camera demosaicing algorithms. It is the demosaicing process which introduces blurr into the image even when using the methods you described would normally produce a tack sharp image on film.
Digital post processing is an artistic tool, so the final image can be manipulated to reflect your own vision. I shoot only digital in manual, and only in camera RAW. Also, I photograph people, and most often it is very close up with a smaller aperture, so features are pretty sharp, but, on a close examination you would notice slight to moderate blurr at 100% and progressively more as you increase the magnification. Sometimes this blurr is fine as it is for my purpose of smoothing skin, or I may need to introduce more, but there are image parts that can not be blurred such as the eyes and lips, or the detail on a wedding dress.
In Photoshop you can apply techniques that allow masking of the parts of an image that you do not want to show a particular action, such as sharpening, sort of in the same way you would mask off an area when spray painting. So with this mask partially removed from the image, i.e. areas of the eyes, lips, hair, and jewelry such as ear rings or the lace of a wedding gown and so forth, a sharpening action can now be applied selectively. It is during this manual post-processing that I can remove the demosaicing blurr and return the image to tack sharp, however I am now in complete control of how much and where it is applied and not limited by my in-camera entire image sharpening process as in jpeg or automatic.
A mask on another layer could allow you to introduce another action, maybe gaussian blurr to smooth over pores or blemishes, or introducing blurr that mimics a shallow depth-of-field, etc.
Thanks for asking this question, I hope my explanation helps.
Lj
|
|
« Last Edit: August 27, 2009, 01:27:09 PM by Eljae »
|
Logged
|
|
|
|
Seth
|
|
« Reply #5 on: August 27, 2009, 01:56:14 PM » |
|
This puzzled me because unless you specifically chose an aperture that throws parts of your shot out of the Depth Of Field for that lens, aperture, and distance to subject, how do you leave some parts minimally sharp, and other parts very sharp? Depth of Field CAN be used to do that. On anything less than 4x5 (with a magnifier) it becomes progressively tougher to accomplish. Photoshop, and other actions, adds the ability to "alter" DOF by select focus. Kind of like the old cigarette cellophane trick in the darkroom.
|
|
|
Logged
|
Seth <CWO4 (FMF) USN, Ret.>
|
|
|
BrianPrice
|
|
« Reply #6 on: August 28, 2009, 07:37:48 AM » |
|
I've used the Photoshop Duplicate Layer/Layer Mask technique in the past to blur backgrounds - here is a quick one I've done as an example: Before: After: You can get more realistic effects using gradient fills for the mask. Brian
|
|
|
Logged
|
|
|
|
Terry-M
|
|
« Reply #7 on: August 28, 2009, 09:53:10 AM » |
|
to blur backgrounds I've done the same too, see attached. This seems a reasonable technique to create a differential focus effect but I still can't see how localised sharpening can make a blurred area sharper. It may locally increase contrast at a micro level to fool the eye but it's not actually sharper. A Bayer pattern image is a product of the imaging device of digital cameras, and the final image from it is basically the RAW output. Each pixel of an image is filtered to record only one of three colors, i.e. RGB, and a full-color image is produced by in-camera demosaicing algorithms. It is the demosaicing process which introduces blurr into the image even when using the methods you described would normally produce a tack sharp image on film. What also needs to be understood is that because there is a difference in the number of sensor sites for R, G & B each colour will have a different resolution, so greens, (most sensor sites) appears sharper than other colours. Qimage USM in both raw preferences and the batch filter can take care of this with the Equaliser. Terry.
|
|
|
Logged
|
|
|
|
Fred A
|
|
« Reply #8 on: August 28, 2009, 10:18:16 AM » |
|
and the batch filter can take care of this with the Equaliser. I use the Equalizer quite often in Qimage (using fairly high values for radius and strength; 3/350 or 4/100 depending on what needs to be sharpened and what needs to be not oversharpened.) A 3/350 might call for a 90% equalizer, but that will enhance the apparent sharpness of the softer segments without over sharpening the already sharp segments. It does take a little trial and error practice. Also bear in mind that we are selectively increasing contrast on edges, not really sharpening. What is out of focus cannot be brought back from the dead, only made to fool the eye a little with some makeup on the corpse! :-) Just for fun, let's try something interesting in Qimage. If you do not have Qimage, download the free demo please. Find some portrait type shot you took that was really sharp, but perhaps too sharp showing facial wrinkles and blemishes which would make the subject happier to be less emphasized. Place the image into the batch filter screen and set the Unsharp mask to a 4 radius and a MINUS 100 strength. Now move the equalizer through its range until you find the spot that makes your face look just right. BTW, that was a -100, not a plus 100. Perhaps, saying it another way... The best way to put it is that sharpening is like a salt shaker to the cook. Just the right amount and you enhanced your dish. Too much and you ruined it. Too little, and well, it could have been better but each to his own taste. What the equalizer really does is to (need a high setting 3/350) reduce the contrast of the edges as you slide it, and allow that strong setting to apply itself to not so edgey stuff like a face or a soft feathered bird without allowing the sharp edged grass and tree foliage to get over sharpened. That equalizes the applied sharpness. Simple... Fred
|
|
« Last Edit: August 28, 2009, 11:01:21 AM by Fred A »
|
Logged
|
|
|
|
Eljae
Jr. Member
Posts: 68
|
|
« Reply #9 on: August 28, 2009, 05:49:16 PM » |
|
to blur backgrounds I've done the same too, see attached. This seems a reasonable technique to create a differential focus effect but I still can't see how localised sharpening can make a blurred area sharper. It may locally increase contrast at a micro level to fool the eye but it's not actually sharper. A Bayer pattern image is a product of the imaging device of digital cameras, and the final image from it is basically the RAW output. Each pixel of an image is filtered to record only one of three colors, i.e. RGB, and a full-color image is produced by in-camera demosaicing algorithms. It is the demosaicing process which introduces blurr into the image even when using the methods you described would normally produce a tack sharp image on film. What also needs to be understood is that because there is a difference in the number of sensor sites for R, G & B each colour will have a different resolution, so greens, (most sensor sites) appears sharper than other colours. Qimage USM in both raw preferences and the batch filter can take care of this with the Equaliser. Terry. Sorry Terry, I can see that you are very advanced in the technical side, so I won't even try to explain the relationship of the anti-aliasing filter to the image. I am not as technical as I wish I were. I am not sure of those technical aspects as to how selective sharpening works in Photoshop, it just does. When these Photoshop techniques are applied the photos appear significantly sharper in the areas areas we want them to be, and smoother in the areas that we choose. Maybe the eye is just tricked, but the final print is better, and a more pleasing appearence of the photo is our objective. There is no doubt that I can learn much from everyone here, thank you all for your posts. Wish I could be more helpful.
|
|
« Last Edit: August 28, 2009, 05:55:32 PM by Eljae »
|
Logged
|
|
|
|
Terry-M
|
|
« Reply #10 on: August 28, 2009, 07:06:15 PM » |
|
Maybe the eye is just tricked, but the final print is better, and a more pleasing appearence of the photo is our objective I think that is the answer, our eyes can be tricked and what Fred said: Also bear in mind that we are selectively increasing contrast on edges, not really sharpening is how it's done. Thanks for starting this interesting thread. Terry.
|
|
|
Logged
|
|
|
|
Seth
|
|
« Reply #11 on: August 28, 2009, 07:08:57 PM » |
|
[quote} author=Terry-M link=topic=299.msg1887#msg1887 date=1251453190 This seems a reasonable technique to create a differential focus effect but I still can't see how localised sharpening can make a blurred area sharper.[/quote] Yep, you're right. But that brought to mind a little trick I learned a long time ago. (Well, digital long time ) When you have something that isn't quite as sharp as you like, you can trick the eye into an illusion of sharp. Not exactly the same thing as here. What you do is use the sharpen "brush" in PS. I usually just set it to 15-30% so it isn't artifacting, etc. Certain speculars that the eye expects sharp can be hit a little at a time. E.g., jewelry, little flecks in the eye, the rim of eyeglasses. Even works a little on the edges of lips, fingers, etc. I try to stay to stuff that is almost there but could be a little more tack. At normal viewing distance for the print size is usually works. Just as we have said before, less is more. You have to think ahead to what the final print sharpening will add/intesify/artifact.
|
|
|
Logged
|
Seth <CWO4 (FMF) USN, Ret.>
|
|
|
Eljae
Jr. Member
Posts: 68
|
|
« Reply #12 on: August 29, 2009, 12:59:00 AM » |
|
Maybe the eye is just tricked, but the final print is better, and a more pleasing appearence of the photo is our objective I think that is the answer, our eyes can be tricked and what Fred said: Also bear in mind that we are selectively increasing contrast on edges, not really sharpening is how it's done. Thanks for starting this interesting thread. Terry. Yes, Terry, thank you for pointing that out...it is just a contrast adjustment. But the gaussian blurr is different, it appears to actually blur the image, and by masking the subject and applying a gradient blurr from the top towards the bottom of the picture it can mimic bokeh. But still it is not as nice as the real thing, and overdoing it looks more like a soft filter. Still, I find it useful when needed to increase the bokeh that is already present, and smooth out skin too. You are most welcome Terry, and thank you Fred, Jeff, and Seth for your posts.
|
|
|
Logged
|
|
|
|
photodude
Newbie
Posts: 4
|
|
« Reply #13 on: August 30, 2009, 04:36:20 PM » |
|
Eljae,
I also use some selective sharpening techniques, mostly stuff from the "real world image sharpening" book by Bruce Fraser.
The advantages of the mask I like are for example, keeping the sky masked to keep clouds soft, to keep long exposure water shots soft, etc.
So my workflow has been Source sharpening, selective (creative) sharpening, and then I resize and final (print) sharpen in photo shop. I will still use qimage to print, but I shut off all the sharpening and resizing features of qimage..
Or - my second option is Source sharpening, selective sharpening in photoshop and the let qimage resize (But I will turn the print sharpening down to 2 or 3) After some experimenting I have found that the default sharpening setting of 5 produces an oversharpened image, being that it is subsequent to two previous rounds of sharpening.
My question/comment to you is: do you use the mask tech. from Bruce Fraser? If so have you set up the automation (actions)? I recently did, wow what a breeze it is to make some quick sharpening actions. I did modify his process a little bit, in order to have some more manual control over what gets masked out.
I have been really happy with the quality of prints using these techniques.
todd
|
|
|
Logged
|
|
|
|
photodude
Newbie
Posts: 4
|
|
« Reply #14 on: August 30, 2009, 04:43:05 PM » |
|
Just thought of another thing.
When doing the initial sharpening passes - I protect the highlights from being sharpened - using the "blend if" controls from the layer menu. It has made a huge difference in the quality of prints (sometimes just looking at a print from a DSLR - screams digital) protecting the highlights, and only sharpening the midtones, has really turned the prints into a much more film like appearance.
Then regardless of final print sharpening pS or Qimage, the final results have been really pleasing.
Todd
|
|
|
Logged
|
|
|
|
|