Three Types Of Image Sharpening
ฝัง
- เผยแพร่เมื่อ 24 มิ.ย. 2024
- I learned from some people with a lot more experience and knowledge than me that image sharpening can be broken down into three different categories: input or raw sharpening, creative sharpening and output sharpening. In this video, I explain what each one is and how I address them in my workflow.
Mac Holbert: artauthority.net/about/
John Paul Caponigro: www.johnpaulcaponigro.com/sto...
The Digital Print by Jeff Schewe: luminous-landscape.com/the-di...
Using the TK Panel Web-sharpener:
• Sizing And Sharpening ...
• Get Instagram To Stop ...
Test printing for sharpening:
• Hard Proofing for Opti...
Using Topaz Gigapixel AI:
• My Workflow For REALLY...
• Which Software For The...
Killer video Sean, this is why your Sean freaking Bagshaw... thorough and articulate
Also your video skills are getting better all the time!
Thanks Nick! Credit goes to you and the other people pushing the video production quality standards. Before that I was content being a disembodied voice with crappy sound and almost no editing. 🤣
As is often the case, nicely done Sean. Thanks for taking the time to organize and present the information.
Hi Sean: Thank you for preparing this video. I must admit that I only do Input Sharpening so with this additional knowledge I am sure that my end-goal for my images will be enhanced. Cheers, Keith (Canada)
Excellent topic and video. Very informative and incredibly well explained.
Excellent overview Sean; thank you !
Super helpful as always!
As usual Sean, well done - great info, and well articulated/presented -- THANKS so much!
You bet Paul!
Very insightful Sean. I had been wondering what you meant in previous videos by smart sharpening for printing. I myself need to dive a little more into becoming better at printing and sharpening for that process itself. The Tk panel web sharpening feature is fantastic. Thanks for sharing.
Extremely helpful!
So nice tutorial. 👍
yeah Sean! helpful most indeed!
Well done. Good information.
Thanks so much Sean!
Hi Sean, great tutorial. I bet there is a strong demand on a class on this topic. I would highly appreciate one.
Really really helpful Sean. A+
Nice job!
Thanks Sean for reducing my confusion with the various types of sharpening. I'm off to watrch some of your other videos in my quest to learn how to prepare images for perfecrt lab printing - the trickiest part of my photography journey.
I'm in agreement with you. Getting the printed image to look how we envision it is a challenge for sure. My recent video on my print workflow for big prints might be a good one. Good luck.
@@SeanBagshaw Thanks for the reply. After doing some research I realised my monitor wasn't up to the job so got myself a decent photography specific monitor a couple of weeks ago, the BenQ 270C. Had a couple of test prints back from a lab and things are looking much better already, I feel I've overcome the 'prints too dark' obstacle. However I reckon it's 55% design and 45% luck for the rest. Lots of really useful information in videos from yourself and others but between hardware profiles, icc profiles, resizing, sharpening, proofing etc, information overload can be a problem. Every day I understand a bit more and I'm determined to get to grips with it all.
Then I'm going to order up a REALLY big print, framed and ready to hang.
Appreciate all your work.
@@richardwood6121 It can be information overload, but you are doing the learning process right. Complex things take time and practice and that's all part of becoming skillful or even "masterful".
Great video Sean some great information there man😀👍✌
Good WorkMan
Thank a lot شكرا جزيلا
That was a great, explanation on those sharpening categories. I have not experimented much with LR (or PS) sharpening; I just let the software defaults stay. But I have not yet printed much of my work either.
I always learn something from your videos. Thanks for your effort Sean.
Right on Steve. The reality is that images can look great with the defaults. At some point these little things give ever diminishing returns. But on the other hand, it's often that last 1 percent that rally makes a difference.
Sharpening my photos specifically for instagram made a huge improvement on how they look. Definitely play around with it a little bit to see for yourself.
My step number 1 is to start out with my sharpest lens set to f8 or f11. DXO Mark's lens reviews will help figuring out which one that is. In my case it is the 24-70 f4 L.
Hi Sean, very useful 7:32 of my time. While not a tutorial in the strict sense (as you told us early on), breaking the sharpening into three categories did give me some aha moments. It has only been a couple of years since I've moved from an "if I can't do it in Lr, it doesn't need to get done" philosophy to understanding the full power of Ps. So now I am working at developing the skills to make full use of the subscription I have. So rethinking the three categories pushed me way down the track. Many thanks!
Great to hear that Gord!
Good stuff Sean 🤙🏻
Thanks Joseph!
Thanks Sean . Learned a lot. Love and support from Pakistan. 👍👏👌
Right on Atif!
Another first-class video, Sean. Thanks! When you are output sharpening for a large print, do you allow Ps to resize larger with upsampling, or do you always do that in another app?
Photoshop can do a pretty good job, but Topaz Gigapixel AI is even better, so for really big prints that's what I use these days. I did some videos on it. th-cam.com/video/2khOr_FqykA/w-d-xo.html th-cam.com/video/-m07uDBheT0/w-d-xo.html
Nice video! Do you have any tutorial of how resize Raw images for small sizes (300 kb) and have a HIGH quality for internet even when you zoom in? I never saw any tutorial explainig how to do that. Would be a great content for all of us. Thank you. Keep going the amazing work!
Sean, to explain "sharpening" we need to visualize the problem of "digital" and pixels. If you continue with a next episode on the subject, I suggest you look at "TrueType", "ClearType" and the history of depicting fonts on monitor displays. This also touches on anti-aliasing. It helps to use the narrative of "What are TrueType fonts?" on computer.howstuffworks.com/question460.htm (one page on howstuffworks) as a backbone to the explanation. In the case of our digital raw images, we essentially have a bitmap. In order to depict a 24MP bitmap to a, say, 33,000 * 22,000 (726MP) A2 size print (1,440 DPI), "sharpening" tries to determine the vector representation of our photograph, run it through a raster image process, and in this, determine the optimal representation of the image at that level of raster detail. In the case of a TrueType font, the vector representation is what we start with, so that is relatively easy. If, in upsampling 30.25 times from 24 to 726, we just replace one pixel by 30.25 pixels that are all the same as the single original, then we get very ugly depictions. The reverse applies when we downsample our 6,000 * 4,000 (24MP) image 3.4 times in order to depict it at 3,240 * 2,160 on our 4K monitor.
The part of the problem that involves rasterization of the vector image (object) as in a TrueType font, is how to avoid jagged edges of curved lines on a regular raster like a chess board. Lines exactly vertical or horizontal are simple, as long as they align with the display raster. All other lines, curved and diagonal or slanted, could be treated in an on/off way and cause jagged edges in the process. Alternatively, rastering could calculate how much to switch a display pixel on as well as in what RGB values. Thus presenting the illusion of a smooth edge.
The other part of the problem is how to get to the vectors. In TrueType fonts, the design was actually represented as vectors (mathematical formulae). Computer games with user definable level of anti-aliasing, generally also have a vector representation of objects to start with. That means, to get a beautiful image on a specific raster size and resolution we just have to run the vectors through the rasterization process.
In digital imaging, vector analysis (cf. edge detection and convolution) is repeated time and again. It is applied when the sensor image gets depicted in the eVF, in the generation of JPEGs, in the "size" or "detail" reductions, likely in AF and, I guess at least when contrast based, and also when inventing the RGB raw ("fake") file from the R-G-G-B monochrome Bayer sensor readout (with the addition of Moiré suppression). In fact, vector analysis happens so often that optimization of when, where and how this happens in the camera could reduce a lot of computational overhead (that is, if/when the computation takes longer than reading back from a cached result). And it is used so much that we have to wonder why the vector analysis data of the camera should not be stored in the raw image file, along the "raw" pixel data so the computational effort does not have to be repeated all the time.
The teaching problem with this story is, when people understand these words, they already understand the whole problem. The others may need a couple illustrations to build the understanding.
For the cognoscenti it helps to visually and visibly explore the rationale of parameter settings in sharpening (as in amount, radius, detail and masking).
A2 size: 420mm x 594mm, or 0.42m x 0.59m=0.25m^2, or 16.54" x 23.39"=2.7sqft, at 1,440DPI is 23,811pixels x 33,676pixels= 801.8MP (but if we use a 3:2 ratio as in the camera, we get to 756MP, using 396mm*594mm - so the example above implies a print with a white border)
This is as I see it solid advice all the way through with good depth (blurring is part of creative sharpening etc).
I would note that there is nothing that can be done in an accessory software that cannot be done in Photoshop alone, and learning to do it in Photoshop adds fundamental understanding while maximizing your value in the cost of renting that Adobe app.
Some thoughts, perhaps obvious (sharpening is indeed a very detailed process):
View at 100% to sharpen the archive file and and at "print size" to sharpen print output (presuming you are doing either manually not via an app, in which case it does not matter).
Set your Photoshop preferences to display the (magnification tool) option "print size" at real print size! So for instance when your Photoshop border ruler (cntrl/r) is set to "inches" or "millimeters" rather than "pixels," a real ruler held up to the screen will match it exactly! You do this via numerical entry at: Edit/Preferences/screen resolution/units & rulers. The preferred number will be about your your monitor width in pixels divided by it's width in inches. Ballpark that, then check & adjust by using a a real ruler held against your on screen Photoshop ruler, to make sure the Photoshop ruler (set to inches, mm, cm) matches it exactly.
My preference setting is 94 px/in for my 1920px wide Dell 24", and a steel ruler held up against my screen ruler (set to mm, cm, or inches) matches exactly WHEN (an only when) the image magnification is at "print size" via the magnifier tool.
Experts will know this already, but for non-experts it's quite a bit of fun to associate the real world (a real ruler) to your Photoshop environment via setting correct print size. PS you may find it's very close already and wonder why you are bothering, and you will also see "print resolution" in the preferences. The print resolution is a separate issue and can be adjusted for any print individually during a printing process.
Thanks for sharing those helpful tips John!!
@@SeanBagshaw - You are welcomed, I tried to keep my text as consice as your presentation, out of respect.
Excellent video Sean. In past tutorials I have noticed that you turn off sharpening in the LR detail tab before making edits and exporting to Photoshop as a smart object. Is this recommended, or do you recommend adding input/RAW sharpening in LR? I usually add input raw sharpening as a smart object filter in PS after making all edits- now I’m second guessing 🤔
Hi Chris. I don’t think I ever turn off the LR sharpening that I can think of. I always apply some amount of input sharpening to my raw files in LR be Florence going to PS.
Sean seem like a nice guy. ✌✌
Hi Sean, thanks for the sharpening tips. If I want to use orten effect or other way to give a dreamy look of the image, that should be before output sharpening or after output sharpening?
Before. Output sharpening is done at the end when the file is otherwise ready to output.
@@SeanBagshawTks Sean!
Sean Great video! I use a Nikon D750 (24mp) and I'm new to printing. Primarily I print on metals, acrylics, and paper prints. I would like to be able to print sizes like 30x40, 40x60. Is there any reason the files from a D750 couldn't do this? I understand how I process plays into that as well. Thanks
A lot comes down to viewing distance and personal expectations too, but in my experience 24 MP is good for 24x36 for sure. With the right image or if using Topaz Gigapixel AI I think you could go to 30x45 and maybe even 40x60. I take 30 MP images to 40x60.
Nice tutorial, can you tell me how i can reduce noise with the TK7 panel?
The TK panel doesn’t have its own noise reduction function. However you can use either PS or LR noise reduction in combination with luminosity masks to target noise reduction to the parts of an image that need it most...usually shadows. This video might give you some ideas on that. th-cam.com/video/ARlwXNeXxbg/w-d-xo.html
@@SeanBagshaw Thank you for the link!
Upsampling instead of sharpening. I generally did no sharpening until I accidentally slid the sharpening slider on one of my shots. Details became visible that I had not noticed before., Still, I generally find LR's default input sharpening good enough. A panorama shot I tested a new camera and lens with seemed to lack detail. It was good but not great. Printing this to 24" (60cm) also looked good but not great. My printer set to 1,440x1,440 dpi, at the print size, needs more pixels than my camera produces. Without most of us noticing this, somewhere between LR or PS and printer, pixels are invented that weren't there originally. This I call upsampling.
At the larger print size or at a larger display than my 28" 4K, it matters how the upsampling is done. My position is it starts best with a non-sharpened image.
At 23.6 inches (60cm) and 1,440 dpi, we need 34,000 pixels in the printer, at the print head. The question is of course, where do we do that. And what intelligence was applied. Here I compared a simple app where my assumption is that the upsampling happens in the driver layer and the printer, versus printing from Lightroom and printing from Photoshop. In all cases using a picture upsampled to the 32,015 pixels needed for the full 60cm. Here I got the best - best detail resolution - result with Topaz Gigapixel. This application was not very mature in stability but when it produced a result I asked it to do, that was excellent. Better than Photoshop's built in algorithm. I am sort of disappointed that Adobe cannot give me this level of upsampling yet, and think Topaz has a couple things to address still.
The experience makes considering what to sharpen, when, why very relevant. It also indicates the relevance of the tool to use and in what order you do one thing or another.
Would you care to comment on Lightroom’s Print Module’s output sharpening? In the past I did what JP and Jeff taught but for the past few years relied more on Lightroom. Aside from sharpening, the Lightroom print controls are less refined than using Photoshop layers to tune contrast and correct color.
Hi Ed. I think LR output sharpening does an OK job, but the control is limited. Your choices are Low, Medium or High and Glossy or Matte. I think the sharpening results in LR are perfectly acceptable especially for people who don’t want to mess around a lot in PS or other software. My bigger concern with creating output files from LR is the sizing algorithm. In this video I learned just how inferior images enlarged from LR are. th-cam.com/video/ARlwXNeXxbg/w-d-xo.html
Great video as always Sean, thank you for your fantastic education! I have a question...I have my website through Smugmug which allows people to purchase my images for print through their 3rd party printer. Smugmug's platform makes it easy for clients to get prints and for me because I don't have to manage that process. I upload my images to my website at the image's native resolution (I use a Nikon D850) so that the high quality, full resolution is available for whatever size print (or digital version) the client may want. However, I'm caught between how I should sharpen my images for this output. I want my images to look sharp on my screen (my website), yet if I sharpen to optimize for the web, it limits the size of print or digital copy a client can order. Yet if I sharpen the image for print output at the image's native resolution, it doesn't look as sharp/crisp on my website because it wasn't sharpened for web optimization. Like you, I use the TK panel to output sharpen my images for web. I can proof delay the images so that if someone wants the full, hi-res image for a bigger print, I can make that available to them, but unfortunately, Smugmug doesn't allow proof delay for digital purchases. I'd like to be able to provide my images in a format that suits a one-size-fits-all. I hope this makes and appreciate your feedback!
I fully sympathize with that situation Ben. Unfortunately, I don't know of a solution. SmugMug provides some great convenience...but until they allow you to upload separate files for screen viewing and print output you are stuck choosing one or the other.
@@SeanBagshaw No worries, I know that was a verbose question, so thank you for your time and input. Much appreciated!
Is it me or sound and video are not synced
Did you watch on phone or computer? The audio and video sync in the editor and when I watch on laptop, but are out of sync on my phone. Not sure what is going on with that.
@@SeanBagshaw I have problems on laptop. Will check later on mobile to see is problem also there
It aint you. It's out of sync.
Very well said. Also audio ≠ video.
Thanks. Did you watch on phone or computer? The audio and video sync in the editor and when I watched on laptop, but did notice out of sync on my phone. Strange.
Hi Sean. Thanks for the video. I am a real estate photographer, 90% of the images are interior. I use a 20mp camera. How would you translate the landscape sharpening settings for interior photography. Thanks again.
Hi Kirt. If it were me I would add a small amount of input/raw sharpening as I do for all my photos...maybe something like amount 50, radius 1 and detail 80. I probably would do much creative sharpening for real estate. Output sharpening would depend on the output. For the web I would use the TK web sharpen action which takes into account the size of the screen image. For printing it would be different depending on the size of the print and the type of print material.
4:29 Your monitor is moving... you can use image stabilisation in adobe premiere but not to much.. :p thanx for the tips!!
Love the content...but your audio appears out of sync to me? Only channel this happens on...
Thanks! On all videos or just this one. Not sure what’s happening. Sometimes it’s out of sync and other times not. I think YT is messing with me.
@@SeanBagshaw it appears on this one, and the milky way one. I haven't noticed it on the Polarizer Sky video however.
You aren't describing 3 "types" of sharpening, but 3 different times where you might choose to sharpen in your workflow.
*mind blown*
Fair enough, but "3 Different Times Where You Might Choose To Sharpen In Your Workflow" is too long for the title and wouldn't fit on the thumbnail. Blame it on the Internet?
I rarely subscribe to channels. Not for any particular reason, I just almost never do it. Your videos are too good not to subscribe to your channel.
Why does it seem to me that the words I hear from you aren't in sync with your lip movement? I don't intend to be critical but maybe it's something you can look for in your next video upload.
TH-cam is doing something strange to the audio track on this video. Sometimes it is in sync when I watch it and sometimes it isn't. Not sure why, but it is what it is.
You prefer smart sharpen over Nik.
Hi Chris. I used NIK to sharpen for quite a while and it does a good job. But I feel I can get comparable results with Smart Sharpen. The problem with NIK for me was that every time there was an update or it switched ownership I would need to re-install and get it all set up again. Eventually I just quit messing with it. And I'm not a fan of external apps anyway because doing the roundtrip away from PS and back gets annoying. If I can find a way to get the same results while staying inside the PS environment I opt for that. :-)