Luminous Landscape Forum
Raw & Post Processing, Printing => Printing: Printers, Papers and Inks => Topic started by: FrankG on January 29, 2016, 08:22:03 pm
-
No doubt this is a resolved issue for most but not me.
I need to make a 20 inch (long dimension) print at 360 ppi from a M43 Raw file (16MP) that is 4592px x 3448px = 12.75 x 9.57 in (4:3).
* at the highest possible, close up viewing, pixel peeping, quality.
My options are:
1. to enlarge (interpolate) it in ACR (Raw) and save as a tif.
2. Save as tif and use Image size>Bicubic smoother to enlarge (interpolate)
3. Leave it at it’s original/native size and scale it up to the paper size in the Epson dialog
I have Nik Sharpener and would apply some Raw capture sharpening to the tif converted from the Raw after it is re-sized to 20 inches. Or, to the PS (Bicubic Smoother) enlarged file. And then again some Output sharpening prior to the printing. Option 3 doesn’t facilitate Output sharpening at the final size.
What is the guru established recommended workflow steps for a scenario like this?
Thanks
Thanks
-
Last time I tested this, door #1 (do it in ACR/LR) with good capture sharpening produced the best appearing prints.
-
Thank you. Would you duplicate the Raw file first in order to keep one at the original size. Or is it infinitely adjustable, up or down, with re-sizing as it is with the other image corrections?
-
Would you duplicate the Raw file first in order to keep one at the original size. Or is it infinitely adjustable, up or down, with re-sizing as it is with the other image corrections?
The raw is raw, isn't ever changed. The TIFF you create from the rendering is of course new and unique. So you'll have the upsized (past native resolution) after you render the raw. You could then render it at any size thereafter.
-
Perfect. Thanks !
Would you consider "good" Capture sharpening to be judged as "good" when it pleases the eye, viewed at 100% or at the screen image sized to approximate the final print, or is it "good" because it follows a particular, or a formulaic procedure?
I have Nik Sharpener and it has a slider which moves across to 100% towards Edge sharpening. So I guess that depending on the nature of the subject matter - further towards Edge for wide flat areas (c/u portrait, flower, sky?), and away from Edge for smaller details (textures)? Perhaps set at 50% (halfway)is satisfactory for most images with mixed content (sky, walls, people...). And it has a strength % slider.
-
Would you consider "good" Capture sharpening to be judged as "good" when it pleases the eye, viewed at 100% or at the screen image sized to approximate the final print, or is it "good" because it follows a particular, or a formulaic procedure?
In ACR/LR, that's about the only way to go. Output sharpening is a different story; can look ugly on-screen and make a beautiful print.
DO test all this!
-
Do the sharpening in ACR ?
And, apply to the strength that is eye-pleasing at 50%, 100% view?
-
Do the sharpening in ACR ?
And, apply to the strength that is eye-pleasing at 50%, 100% view?
Yes, capture sharpening prior to resizing up and rendering that raw data, view at 100%. I think that's key to why LR/ACR was slightly better than using BiCubic Smoother in Photoshop in my tests AFTER the rendering to native resolution of the raw. LR/ACR's resizing algorithms are not identical, slightly superior too. Again, test this on your end to output (even a small 8x10).
-
I will definitely be doing several tests.
This help puts me on a track.
Thanks.
But why the sharpening at the original size prior to sizing up?
Since both happen in the same ACR dialog, don't they get applied to the file at the same time when you click Save & Done?
-
But why the sharpening at the original size prior to sizing up?
Since both happen in the same ACR dialog, don't they get applied to the file at the same time when you click Save & Done?
There really is no 'original' size per se unless you set ACR/LR for the native resolution and you're going to exceed that. But the point is, the sharpening in both products is based on the resolution you ask for when you render the image. So if you ask for a 1MB or a 100MB render from the raw, the product is smart enough to sharpen based on what you ask for, and both will be different. IOW, the sharpening is resolution dependent.
-
I was confused by the Sharpening in the Deatail tab and the sharpening (output) in the resizing window (attached above).
So the order is to sharpen (amount, radius, detail, masking sliders) in the Detail tab, and then go to the resize window
-
I was confused by the Sharpening in the Deatail tab and the sharpening (output) in the resizing window (attached above).
So the order is to sharpen (amount, radius, detail, masking sliders) in the Detail tab, and then go to the resize window
Yes. IF you decided to interpolate UP from the get-go; from the raw to the TIFF. I as said, I worked best for the testing I did for a Webinar a year or so back. But you should trust but verify.
-
I need to make a 20 inch (long dimension) print at 360 ppi from a M43 Raw file (16MP) that is 4592px x 3448px = 12.75 x 9.57 in (4:3).
* at the highest possible, close up viewing, pixel peeping, quality.
My options are:
1. to enlarge (interpolate) it in ACR (Raw) and save as a tif.
2. Save as tif and use Image size>Bicubic smoother to enlarge (interpolate)
3. Leave it at it’s original/native size and scale it up to the paper size in the Epson dialog
I have Nik Sharpener and would apply some Raw capture sharpening to the tif converted from the Raw after it is re-sized to 20 inches. Or, to the PS (Bicubic Smoother) enlarged file. And then again some Output sharpening prior to the printing. Option 3 doesn’t facilitate Output sharpening at the final size.
What is the guru established recommended workflow steps for a scenario like this?
A good option for this task is the software Photozoom Pro (http://www.benvista.com/photozoompro). I have ver 5 and have found it good at preserving edge detail when upsizing. It is highly recommended by Bart van der Wolf, among others.
Drawbacks are it is relatively slow and expensive. Version upgrades are also relatively expensive.
Regards,
Bill
-
Thanks, but cost is an obstacle, so Unless it was head and shoulders above the options that are already included within ACR or PS.
Are there any Raw (ACR) Sharpening pre-sets available that I can use as a capture sharpening starting point, and then adjust from there ?
-
Thanks, but cost is an obstacle. Unless it was head and shoulders above options that are already included within ACR or PS
Hi,
Well, it is. PhotoZoom Pro (http://www.benvista.com/photozoompro) adds resolution to edge detail. Thin lines stay sharp and relatively thin, and sharp edges stay sharp and don't become blurry, despite the magnification, and lower contrast features are upsampled without common artifacts that most other programs produce. I agree it costs a lot for only that purpose (I do not like its downsamplling quality), but for upsampling it is better than the rest.
Another good option is 'Perfect Resize (https://www.on1.com/apps/resize9/)' which is also available without the full OnOne Perfect Photo Suite.
Another good option is Qimage Ultimate (http://www.ddisoftware.com/qimage-u/index.html) which is more of a print management application but with very good resampling, for a very reasonable price and an attractive update and upgrade scheme (currently for a reduced sales price). Although it's a Windows program it also runs under e.g. Parallels on the Mac OS. I does the resampling automatically, and offers to add Smart output sharpening fully automatic, regardless of the output size. It even has a decent Rawconverter built in, so one can print directly from Raw files. It also has cataloging functionality.
Cheers,
Bart
-
Another good option is 'Perfect Resize (https://www.on1.com/apps/resize9/)' which is also available without the full OnOne Perfect Photo Suite.
To each his own. In my testing and video mentioned, I tried that product and it produced lesser quality to actual printed output than either Photoshop or ACR/LR. But then there's a tremendous amount of options in the product so I admit there may be some specific settings that would have produced better results. But why? In the end, the output from all the printed examples looked virtually identical at proper viewing distance. Up close, under a loupe (if you're inclined to do that), yes, there were subtle differences. Not enough that I'd spend a time on additional software.
-
Bart's recommended programs may well be superior and easier, but unless the difference is marked, I can't justify the additional cost (that's just where my finances are at right now).
I'll try to use the detail tab in ACR first (and then up-res the size before saving the raw file to tiff).
Do you know of a few basic pre-set settings for the amount/radius/detail/mask to launch a starting point ?
For example for an image without much fine detail, one with lots of detail, and perhaps one that is a 50/50 average ?
And of course, there is also the adjustment brush to do local capture sharpening
-
Bart's recommended programs may well be superior and easier, but unless the difference is marked, I can't justify the additional cost (that's just where my finances are at right now).
I'll try to use the detail tab in ACR first (and then up-res the size before saving the raw file to tiff).
Do you know of a few basic pre-set settings for the amount/radius/detail/mask to launch a starting point ?
For example for an image without much fine detail, one with lots of detail, and perhaps one that is a 50/50 average ?
And of course, there is also the adjustment brush to do local capture sharpening
Frank, for several years I've followed Jeff Schew's advice in turning the ACR sharpening module from USM to de-convolution sharpening mode by pushing the detail slider all the way to 100. I apply this to my m4/3 files of 3456x4608 px with my output default set to 15x20" @ 300ppi, 16-bit PP-RGB TIF files.
My sharpening defaults are amount=40 and radius 1px, but I can push the amount to 60 or so, depending on contrast, without ringing artifacts. Also frequently add additional larger radius output sharpening (2.5-5px) in PS6 with the Smart Sharpening module (also de-convolutional) to enhance micro-contrast. I get a very detailed natural print appearance with this workflow.
Pete
-
That is interesting. Never heard the term before.Reading about it now
Pete, do you put the Detail slider at 100 for all images, regardless?
Please link me to some so I can just see what type of photography (if you dont mind)
I was just now testing the sliders on an image that has both a finely textured wall and some smooth blue sky (with white clouds).
I found that when the Detail slider was at 100% there were 'artifacts' of some kind visible around the edge between puffy clouds and blue sky, so I brought it back down to 50.
I can also leave Detail at 100 and bring Amount down to get rid of artifacts, but then I lose the fine texture on the wall too
-
To each his own. In my testing and video mentioned, I tried that product and it produced lesser quality to actual printed output than either Photoshop or ACR/LR. But then there's a tremendous amount of options in the product so I admit there may be some specific settings that would have produced better results. But why? In the end, the output from all the printed examples looked virtually identical at proper viewing distance. Up close, under a loupe (if you're inclined to do that), yes, there were subtle differences. Not enough that I'd spend a time on additional software.
I'm confused. When did ACR get a print output module? Shouldn't you be talking about LR and Photoshop?
LR module does do good interpolation and output sharpening appropriate to the output size. In my testing, and that done by Jim Kasson some time ago, Qimage and Perfect Resize (JK testing) easily improves upon what LR can do. With Qimage the interpolation, sharpening, and printing are done in one step and does not require a different tiff for each output size.
In PS, interpolation and sharpening are separate steps and may require sharpening experience or separately purchased products.
-
That is interesting. Never heard the term before.Reading about it now
Pete, do you put the Detail slider at 100 for all images, regardless?
Please link me to some so I can just see what type of photography (if you dont mind)
I was just now testing the sliders on an image that has both a finely textured wall and some smooth blue sky (with white clouds).
I found that when the Detail slider was at 100% there were 'artifacts' of some kind visible around the edge between puffy clouds and blue sky, so I brought it back down to 50.
I can also leave Detail at 100 and bring Amount down to get rid of artifacts, but then I lose the fine texture on the wall too
I forgot to include the very important "Masking" setting in ACR's sharpening module. It reduces the sharpening in less detailed/smooth areas, and my default setting is pretty high at 60, but higher as ISO/noise increases.
Also, looking at 100% after interpolation from 16MP to ~ 38MP (15x20 @ 360ppi) is not very useful IMO, and a 50% view gives a much more realistic appraisal of print IQ, which will be quite equivalent to the size at 25% on my 24" 1080HD monitor.
Pete
-
I'm not sure if it's bad workflow practice but I like to do the Masking slider first so that only the unmasked areas are affected by subsequently moving the other sliders.
Well, actually some Luminance noise (about 20) reduction first
Do you have the Detail slider at 100 for All images, regardless of content?
-
I'm not sure if it's bad workflow practice but I like to do the Masking slider first so that only the unmasked areas are affected by subsequently moving the other sliders.
Well, actually some Luminance noise (about 20) reduction first
Do you have the Detail slider at 100 for All images, regardless of content?
Yes, that shifts the gears to the De-convolutional mode, and I vary mainly the Amount slider and Masking.
Pete
-
Awesome.
All this info really kicks me off to a good start.
Thanks folks.
Just one last thing - I read that Jeff Schewe considers Luninance slider the 5th capture sharpening slider and says that All raw images need Some luminance noise reduction.
Firstly, do you concur? And secondly, would you say that somewhere between 10 & 30 would be an average starting point for M43 raw files shot at 400 or less iso ?
-
Just one last thing - I read that Jeff Schewe considers Luninance slider the 5th capture sharpening slider and says that All raw images need Some luminance noise reduction.
Firstly, do you concur? And secondly, would you say that somewhere between 10 & 30 would be an average starting point for M43 raw files shot at 400 or less iso ?
Yes, I concur :~)
You need to evaluate luminance noise reduction if you've done any lightening to the raw image because doing so makes the noise more visible. This is true even if you shoot at ISO 100. Also, if you are doing any additional sharpening over default, particularly the Detail slider, that will also increase the appearance of noise.
So, ask your self, are you willing to do a bit more work to get a better image? BTW, can't tell what numbers to use because it all depends on the sensor noise signature and the ISO. Evaluate the noise at least at a 200% zoom to see what the noise reduction is doing. (I often view at 400%) this isn't about how much the noise will look like in the final output, this is just to evaluate what the reduction is doing. If working on high ISO, I'll also often add back a bit of Grain effect to mitigate the synthetic look of noise reduction.
-
"additional sharpening over default, particularly the Detail slider"
Yes, because I am learning, from your writings, via a post in this thread, to go 100 on Detail
"are you willing to do a bit more work to get a better image?"
Yes, definitely.
Do you mean to spend more time correcting noise if I've lightened the image, and/or moved the Detail slider to 100?
I've always found that noise reduction makes the image kinda mushy. Counter to the sharpening. Would it not be better to bring down the Detail rather than adding Luminance noise slider?If not then what kind of grain slider numbers would bring it back to a 'normal' look - or is that just a personal taste call when viewing at 100%
-
I'm confused. When did ACR get a print output module? Shouldn't you be talking about LR and Photoshop?
Correct, no print module. First, we're talking about using either ACR or LR to interpolate above the native (max) resolution which both can do. In terms of sharpening, I'm referring only to capture sharpening which of course both provide. Prior to that upsized rendering.
-
Awesome.
All this info really kicks me off to a good start.
Thanks folks.
Just one last thing - I read that Jeff Schewe considers Luninance slider the 5th capture sharpening slider and says that All raw images need Some luminance noise reduction.
Firstly, do you concur? And secondly, would you say that somewhere between 10 & 30 would be an average starting point for M43 raw files shot at 400 or less iso ?
First, while I use lum. NR in ACR fairly often, I don't see it as necessary in most well-exposed low-ISO m4/3 16MP images resampled up to 28MP in ACR for my 15x20 (300PPI) prints on the Canon iPF5100. Even boosting shadows by 50% or so, with moderate use of the Claity slider to 25-30%.
Looking at 100% shows me all I need to see for sharpening and NR effect, and by just going down to 66% the difference is striking. Granted, the native 16MP RAW image can profit from 200% viewing during adjustments.
Pete
-
I have been testing this ACR sharpening procedure on many images since you all kindly provided an explanation and method. Usually around A 40, R 1, D 100. M 30 to 60.
But on many of them, albeit not all of them, I find that there is noise/artifacts (I don't know how to term the objectionable pattern/effect). And this is prior to up-sampling the size in ACR.
Particularly in dark areas, like a black person's skin.
Even though I have Not lightened the image, raised the Exposure or the Shadows, in the ACR sliders. Which leads me to think they're properly captured. Most images are made at 200 iso on a M4/3 sensor.
Is that just the nature of the beast or is my technique lacking.
When I use the Luminance NR slider (the col is at 25 by default) I have to take it up to around or over 25 to make the 'uglies' disappear. But I'm left with a soft pic so I tend to forget about NR ad live with it.
Can I improve my technique or am I dealing with the limitations of a small sensor (M4/3).
-
I have been testing this ACR sharpening procedure on many images since you all kindly provided an explanation and method. Usually around A 40, R 1, D 100. M 30 to 60.
But on many of them, albeit not all of them, I find that there is noise/artifacts (I don't know how to term the objectionable pattern/effect). And this is prior to up-sampling the size in ACR.
Particularly in dark areas, like a black person's skin.
Even though I have Not lightened the image, raised the Exposure or the Shadows, in the ACR sliders. Which leads me to think they're properly captured. Most images are made at 200 iso on a M4/3 sensor.
Is that just the nature of the beast or is my technique lacking.
When I use the Luminance NR slider (the col is at 25 by default) I have to take it up to around or over 25 to make the 'uglies' disappear. But I'm left with a soft pic so I tend to forget about NR ad live with it.
Can I improve my technique or am I dealing with the limitations of a small sensor (M4/3).
Frank, some screen captures @100% in ACR with settings included would be helpful for further insight. That said, I push Masking to higher levels, with my default set at 60. I will also use the NR Detail slider liberally to decrease softness. Which m4/3 camera body are you using?
Pete
-
Panasonic GX7 body.
Pushing Detail up to 100 appears to be the culprit?
I will do some scrrenshots with settings late tomorrow. I could even upload a raw file or two to dropbox
-
I'm not sure if it's bad workflow practice but I like to do the Masking slider first so that only the unmasked areas are affected by subsequently moving the other sliders.
Well, actually some Luminance noise (about 20) reduction first
The order in which you apply the various sliders is not necessarily the order in which LR/ACR applies the corrections. Various authorities have stated that the order in which you make corrections has little significance since the programs apply the corrections in the optimal order. Perhaps Jeff Schewe can confirm this.
Regards,
Bill
-
Pushing Detail up to 100 appears to be the culprit?
Probably. I've almost never been able to go beyond Detail 50-60 without generating horrible artifacts. LightRoom's and ACR's implementation of deconvolution is a bit rough (for reasons of speed I asume). It's miles behind dedicated tools like FocusMagic.
Cheers,
Bart