unsharp mask

JD
Posted By
Joel_Dames
May 24, 2004
Views
1517
Replies
81
Status
Closed
At what stage in processing an image from raw to print would you use the unsharp mask?

How to Improve Photoshop Performance

Learn how to optimize Photoshop for maximum speed, troubleshoot common issues, and keep your projects organized so that you can work faster than ever before!

T
Todie
May 24, 2004
After re sizing/rotating, just prior to printing.
PC
Pierre_Courtejoie
May 24, 2004
Unless you apply a three step sharpening method: <http://www.creativepro.com/story/feature/20357.html>
MO
Mike_Ornellas
May 24, 2004
Capture sharpening is probably the most important one of the three.

Digital shots ALWAYS need sharpening done to them at this time.
T
Todie
May 24, 2004
Capture sharpening would be important if the capture was done at the final print size. Digital camera capture is usually small and scanning is (traditionally) large. Some shops scan very large files for archiving and resize for final output when needed.
VL
Venicia_L_2
May 24, 2004
Joel,

I disagree with the "capture" sharpening advice and agree with Todie that sharpening should be done just prior to printing. I can’t agree with the work flows that advise multiple sharpening steps of "partial strength" at several stages in the image handling.

Proper sharpening is absolutely essential for the best-quality image. But sharpening also introduces its own artifacts and multiple sharpening steps exaggerate the artifacts introduced by the earlier steps.

I (and others) have looked at this very carefully and have found that, whether destined for the computer screen (in which case the sharpening can be of the most primitive type – often involving global USM only) or for very high-quality halftone reproduction at 200 linescreen and above (which requires sophisticated sharpening methods), the best method is to capture a RAW image to which NO sharpening has been allowed to occur and keep that image absolutely unsharpened through all its other editing and color balancing needs until the very last step when sharpening for the intended use is applied.

One work flow to absolutely avoid is to prepare an image for print, including sharpening, then down-interpolate that (large) final image to appear on the Web and apply some "touch-up" sharpening to compensate for sharpness lost in the downsizing. The Web image needs to be generated in its own workflow from the RAW image all the way to final size, then sharpened as the last step for the needs of monitor display.

VL
BF
Bruce_Fraser
May 24, 2004
I have to disagree strenuously with the above.

There is simply no way, in one hit of sharpening, to account for

The softness introduced by digitization, whether camera or scan

The specific needs of the image (close subjects with soft detail need wider sharpening haloes than busy high-frequency images).

The softness introduced by the printing process.

In a properly-designed sharpening workflow, no artifacts are introduced by the capture sharpening, hence they don’t get exaggerated by any creative or output sharpening. If you just hit the image with global unsharp mask two or three times, you will, of course, get a mess…

And output sharpening (whether for print or web) absolutely must be done at final output size and resolution.
GB
g_ballard
May 24, 2004
Ya, I’ll plug ur Complete Sharpening Workflow for Photoshop¨ 🙂 <http://www.pixelgenius.com/sharpener/index.html>
BF
Bruce_Fraser
May 24, 2004
G, thanks for the plug for the plugin, but do realize that I’m not just talking about the plug-in. It’s only one of many possible implementations of a sharpening workflow. Whether or not you buy into that particular implementation (which is continually being refined), the broader issues still remain.

Each capture medium has its resolution limits and noise signature. It’s pointless to try to sharpen detail that falls below a digicam’s antialiasing filter, and we want to sharpen the detail without sharpening the noise.

Each image has its own sharpening requirements. High-frequency images need different handling from low-frequency ones. A sharpen that works well on a low-frequency image may actually obscure detail on a high-frequency one, and a sharpen that works well on a forest full of trees will make a head shot look nasty by overemphasizing skin texture without giving the necessary sharpness to eyes and hair.

Each output process has its own way of translating pixels into dots. You can only sharpen the pixels, and you need to do so differently for an 85-lpi screen and a 150-lpi one. You also need to do it differently for a 300-ppi image destined for 150-lpi screen and a 225-ppi image destined for the same screen, because the pixels wind up being reproduced at different physical sizes.

More often than not, each of these three issues needs a different, often conflicting, answer. If someone can show me a way to address all three with one hit of sharpening, I’d be very happy. But as yet, no-one has come close.
VL
Venicia_L_2
May 24, 2004
Bruce,

I’d better qualify my remarks as I well know that offering advice contrary to yours is not well-received on this forum (I mean that as a compliment – I have benefitted much from your wisdom – I was one of the first to buy your book).

I stand by what I said about "capture" sharpening and handling the file unsharpened until the last. But I admit that the sharpening methods I use, applied immediately before final use of the file, do indeed involve several (up to 4) sharpening "passes" of the USM matrix. They also involve several softening passes. These passes are made on various masks of the image depending on image detail, edge configuration, frequency distribution, final resolution, linescreen and color information. This process is carried out as one continuous "macro" set of actions.

As for images destined for the Web, especially digital SLR camera files which tend to be very noise-free, a single pass of USM (whose settings are dependent on each image at hand) seems to achieve all the quality that monitors are capable of displaying. More sophisticated methods just seem like overkill. I’ll accept that, in theory, this is wrong, but I don’t seem to be capable of seeing much difference on random samples of various methods, nor do any of the photographers or printers with whom I work.

VL
T
Todie
May 24, 2004
Bruce gives good tips to many, but I cannot advise other photographers to sharpen. I’m greedy this way: I want to do it all (that’s my workflow: )
BF
Bruce_Fraser
May 24, 2004
Venecia,

We all do what works for us, but let me humbly submit the following.

If you want your image to serve multiple purposes, I suspect that a good deal of the complex sharpening you currently do as a final step could be better done only once, on the master file. The rest would be done at output time on the final-rez file.

Your current approach seems akin to waiting until the final output is known before you attempt any tone or color correction. It’ll work, but it’s not terribly efficient…

BTW, if you shoot a DSLR at 800 or 1600 ISO, you’ll get plenty of noise!
VL
Venicia_L_2
May 24, 2004
Bruce,

I gratefully accept your humble submission, honoring it with its place among other wisdom you have provided.

In return, this unworthy student offers this further explanation of her work flow and its rationale:

I work mostly with RAW files in the Fuji S2 format, also a few Nikon and Canon Raw format files. As has been the style with which to refer to such files of late, these are my "digital color negatives."

Only my S2 seems to truly turn off all sharpening in-camera. I convert the RAW files into Adobe 98 RGB space at 16 bits with sharpening off in the conversion. I try to do the lion’s share of tone compression/expansion in the conversion so that the 16 bit TIFF that results is as "fat" as possible. The RAW file is then archived.

All color corections, tone corrections and any other edits are done to the TIFF and it is then saved in 16 bit form at its maximum size (70 Mb in the case of the S2) as my working image. It is a relatively simple process, then, to resize the working image, as needed, for each of its several uses (if there are several), convert to 8 bits and apply the correct sharpening workflow for the particular output requirement. Efficient enough.

As to noise in DSLR files. Sure, there is noise at ISO 800 and 1600. But amazingly little in the S2 files, (monster noise in 1Ds files), especially if they are properly exposed – underexposure is BAD. But the real perspective I think about this subject – and I very humbly submit this opinion – is that we have so quickly come to expect so MUCH.

It was only a very short time ago when even ISO 100 from film gave more noise (grain, etc) than digital SLRs produce at ISO 200 and even 400. As far as color work was concerned, for maximum quality from film, I NEVER would have entertained anything exceeding ISO 100, NEVER.

At 35 mm, Velvia at ISO 40-50 or Kodachrome at 64 was my standard for grainless quality.

Many DSLRs now produce stunning results at ISO 400.

Now, the Kodak 14n and the 1Ds are criticized because their performance at high ISO is a big departure from their ISO 100 quality. I am very glad I can get such incredible quality from high ISO DSLR files.

I am continually amazed at ISO 800 performance. I used to be satisfied with 8 x 10 prints from Ektachrome 100. Now the S2 and other DSLRs can produce 11 x 14 prints at ISO 800 that are as good or better.

As always, receptive to knowledge,

VL
L
Larryr544
May 25, 2004
Veronica – How do you get a 70 Mbyte file from 3024 by 2016 pixel CCD sensor? Even at 16 bits and up scaling it to RGB the mathe gives a 36Mbyte file. And that would still be up scaling the RAW file.
JS
Jeff_Schewe
May 25, 2004
I would submit that you haven’t tried the demo or you would understand that with all camera/raw sharpening off, your images are soft while working on them. Doing an appropriate amount of edge sharpening-without driving the edges to white or black halos-will provide a superior working file. Additionally, localized sharpening and smoothing is useful in the creative stage of image enhancement.

For output sharpening after final sizing, there is indeed an optimum degree of sharpening to be applied based upon the image resolution and the output media and resolution. These are knowable amounts, based on testing. And the amounts and degree of sharpening can only be determined by visual inspection of a print-not on screen.

Again, no single sharpening can EVER be the "correct amount" to address the three main phases of sharpening requirements. You are welcome to your opinion, but ours is based on a lot of testing. . .and a whole lot of prints & output.
VL
Venicia_L_2
May 25, 2004
Jeff,

What demo? Did I miss something?

"you would understand that with all camera/raw sharpening off, your images are soft while working on them" What are you referring to. Yes, I am quite aware that my images are soft until the very last step of my workflow. So are the images from my drum scanners.

Is this a tag team effort? I’ve already agreed that my sharpening process does NOT use a "single sharpening." I apologized to Bruce for mis-stating the process. I’m using programs that do a pretty thorough analysis of the needs of the sharpening given the parameters I’ve already listed.

There is no proof whatever that addressing the various requirements at several different temporal stages in the handling of the file vs addressing them all in one multistage process, done as a "last step," makes any difference in the final quality. If the parameters are handled, they’re handled. My workflow is quite good, actually. And not at all unique.

And what makes you think you have a monopoly on testing? prints? output?

I like talking to Bruce better. You don’t sound like you’re any fun!

VL
VL
Venicia_L_2
May 25, 2004
Larry,

The S2 produces files as though the sensor were a 12 megapixel device. At 16 bits, that’s a 70 Mb file.

The S2 sensor has 6 million photosites (give or take). The sensor is arranged in a hexagonal configuration called "Super CCD" by Fujifilm. The diagonal orientation allows mathmatical treatment which gives increased image resolution in the vertical and horizontal axes while resolution in the diagonal axes remains equivalent to a 6mp sensor. Since the eye responds more to vertical and horizontal image information than that at other orientations, the S2’s increased resolution over other 6mp sensors is readily apparent. It’s also been confirmed in tests.

The S2 does not, however, deliver the resolution of a true 12 mp sensor. The increased resolution is about equal that which an 8 or 9 mp sensor of "standard" construction (not Super CCD hexagonal) could be expected to give.

The increased resolution is not just simple up-interpolation of 6 mp data. To date only Fuji’s software conatins the proprietary code that can do the required interpolation. It’s slow, clumsy and doesn’t even use colorsync to display a color-correct display image. PS CS Camera Raw has recently greatly improved the Fuji RAF Raw file handling, but it’s still only the Fuji Converter EX that delivers the extra resolution at "12 mp."

VL
PC
Pierre_Courtejoie
May 25, 2004
Venicia, you can find a demo of the plug-in at: <http://store6.esellerate.net/store/catalog.aspx>

<http://www.pixelgenius.com/sharpener/workflows.html> for two possible Workflows

I have to say that this thread is very interesting, both by the gentle tone of the exchanges, and also by its technical background.

Venicia, I suppose that your sharpening action is proprietary, but could you please enlighten us more about its concept?
MO
Mike_Ornellas
May 25, 2004
V,

I’ve done some of the output for PG and am very impressed with the routines and methodology that they are applying. I’ve seen the paper choices, I’ve seen the various experiments with different settings during the beta period. I’ve seen the line screens and angles.

I’ve been doing prepress for way to long and know what is crappola, and what is not.
JS
Jeff_Schewe
May 25, 2004
Venicia,

Unless you understand the theory as written by Bruce in his Creative Pro article <http://www.creativepro.com:80/story/feature/20357-1.html>

And the practice as one can understand by using the demo of PhotoKit Sharpener, then you are talking from the point of view of inexperience.

Yes, the "old way" of waiting till the very end of the image processing workflow for sharpening application has been the typical way that people have used to apply sharpening. But that is NOT optimal and there is no way of knowing exactly how much to sharpen for the various output requirments nor for the content of the image. There is also no way to soft proof, on screen, to tell you if the sharpening being applied is too much or not enough. A monitor screen simply can not show you a reliable simulation of the needs of print output. The only output sharpening that CAN be determined from the monitor view is for the web-where the monitor is the final output.

You stated disagreement with the principal of a multi-pass sharpening workflow. However, you have not backed up your point of view with any substantial proofs that a 3 phase sharpening workflow is NOT better than what you’ve been doing. . .and when I talk about making prints to evaluate, I’m talking about assembling a series of images into a sharpening target-one that incorporates a variety of image sources and image types-and using a scientific approach to evaluatie the sharpening based upon both the media and the resolution of both the file and the output device. I’ve not run into many people who have done that.

As for the tag team approach, well yes, Bruce and I worked on the concept, the workflow as well as the sharpening taget and the sharpening routines. Bruce and I are members of PixelGenius.
AW
Allen_Wicks
May 25, 2004
This is a great thread. Keep up the dialogue! I am particularly interested in what sharpening/smoothing edits digicam images are most sensitive to; i.e. what must be extremely precise and what is less critical. [Yes, I am studying all the links provided.]
L
Larryr544
May 25, 2004
Venicia – A 12 Mpixel camer will produce a 12 Mbyte file at 8 bits and a 24 Mbyte file at 16 bits unless there is some up sampling. Each pixel is only one color.
P
PShock
May 25, 2004
A 12 Mpixel camer will produce a 12 Mbyte file at 8 bits …

Not quite, Larry. That’s 12 mb PER CHANNEL. Three RGB channels: 12 mb x 3 = 36mb @ 8 bits.

In any event, color it any way you want – the Fuji is using interpolation.

-phil
B
Buko
May 25, 2004
I’m using the S2 and I sharpen my RAW files when I open them with ACR2.2. sharpening is off when I shoot.

You have the ability see how much sharpning you are adding and it gives you a nice image when the sharpening is done when converting the image. same with changing the color temp of the pic.

I never save over the original RAW thats my neg.
VL
Venicia_L_2
May 25, 2004
Whoa people!

How did this discussion get to be about a specific product? It didn’t start out that way at all. The original question was:

"At what stage in processing an image from raw to print would you use the unsharp mask?" That’s all.

I replied that it should be done as the last step. I also said that multiple passes of USM exaggerated artifacts of the process. I stand by that statement.

Bruce then, correctly, advised that a multiple sharpening plan of attack done at several stages in the imaging handling process can avoid any enhancement of artifacts and can address the various needs of each particular image and its output environment. I AGREED with that advice. I admitted that my original advice seemed to advocate a single application of USM. I didn’t mean to say that and described the method(s) I use to involve MULTIPLE, sophisticated edge-enhancement/smoothing passes, tailored to several critical image requirements, done as the last step prior to output.

Mike, I don’t doubt that the product for which you did testing does a good job and that the "routines and methodology" are sound. I accept that you have a lot of pre-press experience. But you cannot state, and no one else can state, that the product is superior to every other method that exists. It simply is not. If you’ve found other products that are "crappola," I am not at all surprised. I’ll accept your assertion that the product you tested is not. This thread is NOT about some particular product.

Don’t assume that others here don’t have experience in this field. I can remember pulling color separation negatives off a huge drum scanner (Dupont?), processing them and preparing Cromalyn (sp?) proofs and checking for sharpness, dot gain and a myriad of other factors, then sending them for press runs and further charting performance vs paper type, ink type, press type, etc. etc. We charted sharpness vs linescreen (110 through 250), sharpness vs paper type. Nothing was easy then. It took days of hard labor to get simple data. There was no such thing as desktop systems. There was no Photoshop. The scanner was a huge beast with "COLOR COMPUTERS" inside which we were told came directly from space research. (I believe they were analog computing systems). The sharpening technology was well-developed then and in some ways was more sophisticated than that which we now use in Photoshop, because it not only did edge enhancement of the image data, but varied the size and shape of the halftone dot which it burned to the film separations.

(Of course I was just a sweet young thing then, awed that someone from graphics was allowed into the manly world of the scanning department, and pressroom. Little did they know one day an art director would be in charge!)

Jeff, I find your manner a bit offensive. You imply that anyone not familiar with your product is inexperienced and that any way other than yours is the "old way."

Back off Buckwheat! There is nothing "old" about my techniques. You haven’t read anything I’ve actually written, but continue to pound away with your "superiority."

Then you jump to soft proofing (where did that topic come from) and I really cannot tell what your point is here. Are you for it or agin’ it? You seem to object to keeping the image soft until the last step because there is no way to evaluate the sharpening that has been incrementally applied, then you say that soft proofing is not reliable.

I don’t use the monitor to soft proof sharpening, other than to prepare images for the Web. (and then it’s unreliable because my monitor is a high-quality 21" and I have no idea what the image looks like on a smaller, lesser monitor).

The sharpening I apply is based on hard data and just plain experience (there is no way around the human factor – there is still a good deal of "art" in this field). The hard data includes image size and resolution, printing linescreen, paper type (coated [gloss or matte], uncoated). The software also determines needs based on global and local contrast and edges. Last comes the "experience factor" to select more or less sharpening depending on a whole gamut of factors that range from knowing a customer’s preferences to the phase of the moon.

But I find it most offensive that you have missed no opportunity to impose yourself here and unabashedly advertise your product. Just a few weeks ago, you demanded that a contributor’s message be removed because you insisted he was "driving traffic to his Website." All he did was have the nerve to offer some tutorials for which he charged nothing. (of course the ensuing pissing contest revealed much about both of your personalities). I have no trouble at all with someone else recommending your product, But I don’t think you should be doing any self-promotion here. At least not the way you have.

Pierre, a product which applies very flexible multi-sharpening/smoothing passes (as the final series of operations on an image otherwise completely ready for output), either as an automated sequence, or fully configurable, completely under the control of the operator, is "FlexSharp." It’s distributed by its author. I don’t have his email address with me, but I’ll post it later.

I can’t recommend FlexSharp without some reservations, however. I believe it was developed specifically for low-noise, digital camera files, especially the S2, and recent 6 mp Canon and Nikon cameras. It especially likes files on which NO sharpening has been applied by the camera or by the RAW to TIFF conversion. It works beautifully for those files. It also works well with scanned images, although it is a bit aggressive on those unless you really no how to use its many "settings."

It is currently at version 1.4. I have version 1.2 and the 1.4 upgrade (there was no 1.3). I have found that v1.4 produces a lot of noise in the blue channel of my images. I don’t know yet if I’m doing something wrong or if it’s the software, so I am doing most of my work with the 1.2. If you get it, I recommend asking the author for BOTH 1.2 and the current version. It was $25 when I got it last year. FlexSharp is not a plug-in, it’s a collection of Photoshop actions.

I’ve "tested" it against some other methods including the famous product mentioned above, NIK sharpener and some "public domain" smart sharpening routines available around the Web. My bench mark for evaluation is output to Epson 2200 and 7600 printers. Several of us regularly try to do "double-blind" evaluations of various images. I have not had the luxury of extensive evaluation on press between it and other products. I also have proprietary software from Agfa and Scitex which is REALLY old and serves mostly for benchmarking.

The most meaningful thing I can say is that Flexsharp does as good a job as anything I have used costing many thousands of dollars. So does PG Sharpener. Flexsharp and methods like it suit my workflow. The PG sharpener appeals to a slightly different workflow. I didn’t like the NIK product, but can’t remember if we were getting excessive noise or what the problem was. At a certain point with these techniques (because they work so well), it comes down to just splitting hairs. They do the job, but like everything else, the operator needs to know what he or she is using them for.

VL
B
Buko
May 25, 2004
the Fuji is using interpolation.

well yes if you shoot jpegs.

No if you shoot RAW. if you shoot RAW, interpolation is done by photoshop if you use ACR.
MO
Mike_Ornellas
May 25, 2004
But you cannot state, and no one else can state, that the product is superior to every other method that exists. It simply is not.

Did I?

Color is subjective along with the rest of the industry, but quality is quantitative and can be assessed.

most people can’t tell the difference or care to.

btw. what size skirt do you ware?
VL
Venicia_L_2
May 25, 2004
Buko,

Or by the Fuji EX converter. And the two interpolations are different. The ACR interpolation did not utilize the unique arrangement of the Super CCD sensor in the earlier version. It treated the data as a standard 6 mp format. In fact, Photoshop engineers expressed their opinion that the S2 sensor did not offer greater resolution. The new version is much better, but who knows how it works. Resolution is still better with Fuji EX.

Those who are supposed to know about these matters say Fuji won’t divulge its proprietary information about the sensor to anyone, including Adobe. Mere mortals such as you and I are kept in the dark about such things.

VL
VL
Venicia_L_2
May 25, 2004
Mike,

"Did I?"

Did you?

VL
MO
Mike_Ornellas
May 25, 2004
no.
T
Todie
May 25, 2004
Venicia,
That must’ve been a Crosfield scanner and Cromalin proofs. I’m on your side, but Jeff didn’t post the first link to Pixel Genius and he and Bruce have payed enough dues to have preferred status on this forum.

I apply sharpening on layers, so it doesn’t matter when I do it. Most of my work is scanned in-house on Scitex, Hell or an old Isomet. We have a Tango that was considered "too sharp" and sits in a corner, looking like a water cooler : )
I evaluate sharpening by viewing images at 50% on Barco, Artisan and LaCie monitors, print to Epson 10600, Kodak Approval, Lambda, Theta, Frontiere and an old LVT.

I’d like to discourage photographers from sharpening their digital captures, because too often they ask for 48×90" prints from the Lambda and all the artifacts show.
MO
Mike_Ornellas
May 25, 2004
Todie,

I’m interested in that water cooler if you arn’t.
T
Todie
May 25, 2004
I’ll tell the owner of the company. He’s trying to return it to a manufacturer that doesn’t… exist anymore.
err… it’s brand new (if that makes sense).
MO
Mike_Ornellas
May 25, 2004
you can contact me off list at

VL
Venicia_L_2
May 25, 2004
Todie,

Can you compare the image quality of the Lamda vs Epson 2200/7600/9600?

Yes, evaluating sharpening at 50 % gives a reasonable "preview" and is probably pretty accurate for an experienced worker. So many people insist on viewing at 100%. Many images that look awful on screen at that size print very well in certain printing environments. The sharpening requirements for halftone reproduction are awfully hard to evaluate on screen

It’s too subjective for me and I find that after several straight hours, my ability to judge images by their monitor appearance really drops off.

VL
T
Todie
May 25, 2004
Venicia, Lambda is… special, in that it uses photo paper and wet processing. The Epsons can have greater contrast and possibly range, but photo paper is pretty in itself.
JS
Jeff_Schewe
May 25, 2004
Venicia,

Yes, I’ve read the entire thread. . .particularly where you state: "I disagree with the "capture" sharpening advice and agree with Todie that sharpening should be done just prior to printing. I can’t agree with the work flows that advise multiple sharpening steps of "partial strength" at several stages in the image handling."

From my point of view, that is an old style of sharpening workflow that can easily be shown to be less than optimal. It’s very easy to show that various input sources require different sharpening-capture size, ISO, with or without JPG artifacts and on the film side, different scan sources also require different sharpening. However, it’s not a stretch to develop a source based edge sharpening that regains the original sharpening lost by the digitalization of an image. High frequency images require different sharpening than low frequency images. Edge width sensitive sharpening with proper blending to avoid noise sharpening and driving hilight detail to white pretty much means that "capture sharpening" is a vital approach to image sharpening.

If one wants to work with images while editing that are soft, one can certainly "wait till the end" to apply sharpening. However, that also effects the creative stage of sharpening where a combination of sharpening and smoothing can substantially improve an image.

The final stage, output sharpening can certainly be determined by empirical testing and does not impact the requirements of the image content nor creative sharpening requirements. It can be shown that a 300PPI image needs X amount of sharpening for 150 LPI halftone repro on uncoated paper. Same with inkjet-a 360PPI image printed on watercolor paper at 1440DPI will require a specific amount of sharpening. This is knowable.

However, to wait till the very end. . .even if you are using exotic sharpening routines comprised of multiple passes of USM, you will NOT experience an optimal per image sharpening workflow. You will be working with images that have lost initial sharpness due to digitizing. You will not employ a creative stage for image enhancements-something that is critical for optimal image processing. If you wait till the very of of the image processing workflow, you lose flexibility, creativity and also the ability to apply final image rez and output specificity. Additionally, there are a LOT of optional sharpening one can employ that does NOT use Photoshop’s USM at all. High Pass sharpening set to overlay or soft light produce a more controlable output sharpening than USM ever could.

As far as the product in question, I was not the first to mention it. However, the way you were posting indicated that you had not used the demo so did not understand the nature of the workflow. It was also unclear whether you had read Bruce’s article so as to comprehend the concept of a sharpening workflow.

As far as "softproofing", Bruce and I do compare a sharpening workflow to a color management workflow-with the possible exception that there is no way to reliably softproof or predict, on a monitor, exactly how much and what kind of sharpening an image needs for a specific output. Softproofing for color however is quite mature and reliable.

So, on your "final sharpen" approach, do you have specific routines for for various types of input? Do you ever use localized creative sharpening? Do you have routines for 180, 240, 300, 360 and 480 PPI image resolution for both matte and glossy inkjet paper? Do you have special routines for contone and halftone output?

It seems that YOU have no problem casting doubts on a multi-phase sharpening workflow and yet you’ve yet to indicate precisely how you have arrived at your opinion that waiting till the end is superior. As for the "Buckwheat" comment . . .no comment.
L
Larryr544
May 25, 2004
I realize I’m mixing in here but it is on topic. A pixel on a digital camera is one color. A raw file from the Canon digital Rebel is 6 Mbytes from a 6Mpixel sensor. So if the S2 is interpolating (which they state they are in their literature) from 6 Mbytes to 70 Mbytes then that is what is happening. There must be some sharpening as well happening there. That’s a lot of interpolation. Glad to hear that the images are good!
AW
Allen_Wicks
May 25, 2004
Chill, everyone, this is just a discussion…
————————————-

My understanding was that _all_ SLR digicam images are "interpolated" [sic] (more accurate may be to say manipulated) to some degree by typically very proprietary in-camera software algorithms prior to being written to the camera mass storage (typically a CF card). That would include RAW pix.

Someone please correct me if I am wrong. Note that I am not referencing sharpening, just computer code applied to the sensor data prior to writing to the card.
B
Buko
May 25, 2004
A RAW fuji S2 file is 12.5 MB. On the RAW file, the interpolation is not done till you open the image with Photoshop or the fuji software. and you can choose whether or not to sharpen at this point. you can also turn sharpening on in the camera. I choose not to.
R
Ram
May 25, 2004
Allen,

RAW is supposed to mean just that, uncooked. Unmanipulated. However, what goes on in there is as mysterious as to what goes into a (raw) sausage.
AW
Allen_Wicks
May 25, 2004
The reason I ask is because I disagree with the implication of what Buko states.

My point is that AFAIK what we get in every RAW file is data that has been heavily manipulated by proprietary software algorithms, e.g. before opening in PS. I believe that is the case in every DSLR and with all in-camera options like sharpening turned off.

The point may be a small one, but I think we make a mistake when we visualize RAW data as raw data, because it is not.
L
Larryr544
May 25, 2004
Thank you Buko. I was getting worried that Fuji had invented magic. The diagonal sensor placement isn’t magic it’s still just the best interpolation that they have been able to come up with.
MO
Mike_Ornellas
May 26, 2004
AHHHHHHH…….

process control!

What a lovely subject that seems to elude us with every breath I take…….

Every move I make…..

Every cake I bake….

————————–

ACR, NEED’s Process control, and guess what?

It does, but the simple fact that screws this whole thing is, is the lack of standards in a RAW format, along with mind control, that I’ve developed picking the caucus of some very fine folks.

Adobe needs to create it’s own RAW format. It may be easier said than done, but its gotta get a fair chance to be implemented as a standard, which buy the way, and yes, buy,……… ALL OF US WOULD…… drop our pants for….

To be or not to be…

phew..
MO
Mike_Ornellas
May 26, 2004
And yes, Sharpening is a very important topic and needs to, as well as, be grown from seeds of passion…

I’d like to publicly thank the Pixel genius team for an education.

That doesn’t mean that I use or agree with everything, but it’s food for thought guys, and that’s what all this stuff is about.

Basically play with yourself until competition or until satisfied with desired end results.

The workflow is very important and when you have establish completion, You understand why, and make edits to your logic based upon what you currently understand in your environment. It creates thought and this is what we need.

think people, think and compare….

The way in which the renderings get duped to a new layer and is fully editable. That’s good logic and it will be assimilated into the collective…

Think more, and you see architecture.

Think even harder and you have Zen…

a standard to live by….
MO
Mike_Ornellas
May 26, 2004
Become a legally ordained minister within 48 hours…..

As a minister, you will be authorized to perform the rites and ceremonies of the church!

Perform Weddings, Funerals, Perform Baptisms, Forgiveness of Sins Visit Correctional Facilities

Want to start your own church?

Press here to find out how ….

O
MO
Mike_Ornellas
May 26, 2004
Help me, I can’t escape my job!!!!!!!!!!!!!!!!

I’m having a melt down……

pleas….pleas, pleas… someone ass me, as a minister of the church!
MO
Mike_Ornellas
May 26, 2004
note-to-self,

buy real estate now…..

for my F&%$ing cell!

mo
VL
Venicia_L_2
May 26, 2004
Raymond, Buko, Larry, Allen,

My knowledge of the real nature of the data in the various "RAW" digital camera formats is as shaky as anyone’s here. FWIW, my understanding of the Fuji RAF format is that, other than the Bayer interpolation to derive "3 channels" of color information, no other interpolation (that is, spatial manipulation) is done in-camera. And whether Fuji does NO in-camera sharpening or does very LITTLE in-camera sharpening, The S2 appears to have the softest RAW files of the current crop of DSLRs. Although I understand that the Canon 1D Mark 2 can also turn off "all" sharpening.

The "up sampling" interpolation of the S2 file to the "12 mp" size all takes place in the conversion software. Who knows what magic Fuji put in there. However, I have it on good authority that eye of newt is not employed. I’m not sure about bat wings, though.

Jeff,

The thread’s all yours, big boy. You just like to beat people up and be confrontational. You don’t listen and you twist words to suit your purpose. There is no sense in trying to have a discussion with you.

Put your big club down. I’ll make this easy for you. You are the greatest photographer around. Your program is the bomb and every one should buy it and stop using anything else. Your workflow and process sequence are the only logical ways to accomplish the task. If we had all just admitted your superiority, life would have been ever so much better.

Larry,

FLEXSharp (Photoshop actions) is available from:

Ferenc Harmat,

VL
MO
Mike_Ornellas
May 26, 2004
The bomb?

What you say then?

what makes your logic better V?
AW
Allen_Wicks
May 26, 2004
Venicia-

I disagree. It seems to me that modern DSLRs do much more than just apply a Bayer filter between the sensor and the file on the CF card. I believe the (very proprietary) demosaicing and other algorithms applied to every image file before we see them as RAW files are very extensive. The significance of that, if true, would be to suggest that files from different DSLRs may optimize with different sharpening edits.
T
Todie
May 26, 2004
A RAW standard may prevent innovation.
R
Ram
May 26, 2004
A RAW standard would be an oxymoron, it would seem. If it has to be converted to a standard, the image capture is no longer RAW.
T
Todie
May 26, 2004
My take, Ramón, is that if the standard uses the Bayer grid and can help adjust exposures, color profile and sharpness, it’s not dumb, but may prevent innovation in cases like Fuji’s hexagonal geometry and Foveon’s chip technology.
JS
Jeff_Schewe
May 26, 2004
Todie,

A raw standard COULD include STANDARD white point balance, tags and locations. It COULD include standard EXIF metadata for things such as lens, ISO, camera owner, etc. It could ALSO include a standard file wrappers that would allow for, among other things, standard locations for IPTC metadata, sensor spectral data, calibration data and other goodies that do NOT need to be proprietary.

A standard raw format could also guarantee that long-term viability of reading the raw file-even after a camera is long discontinued. There are many cases right now where digital camera’s raw files are not able to be opened because the software required has not been rev’ed to current OS’s.

Then you have the situation where Canon, in a move that was idiotic, decied that a .tif file extension would be a good idea. . .even though only the file wrapper conformed to tiff specs. They have changed that to a CR2 file extension now because it caused so many problems for Canon 1D and 1Ds users.

There would be no reason that proprietary info could NOT be included in even a standardized raw file format. Look at ICC profile to see where standardization combined with the ability to include proprietary data (the secret sauce) have made ICC profiles interchangeable cross platform and across application.

If all the camera makers continue to do “whatever they feel like” we will have a potential major problem down the road. . .

For more info regarding raw formats, there’s a new blog type web site at <http://www.rawformat.com>
T
Todie
May 26, 2004
Jeff, Your list is longer than mine, but we’re saying about the same thing about the benefits of RAW.
I think that Chris Cox considers the S2 a 6MP camera, while those who use Fuji’s software argue about better.
Can ACR deal with Foveon’s RAW format (if there’s such a thing)?
MO
Mike_Ornellas
May 26, 2004
You want to know what is idiotic?

Having to create RAW support for every digital camera.

No…… I think that’s insane.
VL
Venicia_L_2
May 26, 2004
Mike,

I NEVER said my logic was better. Equal, yes. That was my point.

Allen,

What do you disagree with? My admission that I am as much in the dark as anyone here regarding what goes on inside the cameras before they make their RAW files available to the Firewire port?

We’re all just speculating about this issue. I know I am. I said that I understood (meaning assumed, based on my understanding of hints, rumors, whatever)) that no up-sizing interpolation was being done in-camera. Do you disagree with that?

Of course different images need "different sharpening edits." Two consecutive images from the same camera can need different edits, let alone different images from different cameras. Have I said otherwise?

VL
MO
Mike_Ornellas
May 26, 2004
V,

Look at it this way.

Input and output sharpening that is fully editable and non-linear, is the future.

The same ideology with source and destination color spaces.

Creative sharpening, is, well, moot to me.

What you say has total merit, but the simple fact that you can adjust the amount of sharpness for input and output, is killer!

Hell, the PG team, let alone Adobe hasn’t realized where the F&*% I’m coming from.

There is a bigger picture……
JS
Jeff_Schewe
May 26, 2004
Todie

Camera Raw 2.2 will handle the Foveon chip for the current cameras. The one that just shipped (or will ship) will require a new rev of Camera Raw.

The S2 _IS_ a 6MP camera by pixel count.

From the Fuji spec:
CCD sensor – 23.0 x 15.5mm Super CCDNumber of total pixels: 6.49million pixels in an interwoven pattern

Number of recorded pixels 3,024 x 2,016 pixels for an actual usable pixel count of 6,096,384

The diagonal bias of the photosites makes for a more compact arrangement but also produces a bit of waste so the actual recorded pixel count is a bit less than the actual pixel count. The 12MP count is a non-standard calculation based upon post-capture interpolation.
VL
Venicia_L_2
May 26, 2004
Mike,

OK. I can be educated. If something works better, I’ll adopt in in a second.

But I’m still having difficulty understanding how you are benefitting from "input" (capture?) sharpening. Maybe it’s me. I don’t need to have an on-screen suggestion of the (partially done) sharpness of my images to do page layout. It’s enough that I know the images will be correctly processed in the end.

VL
MO
Mike_Ornellas
May 26, 2004
We’re all learning here.

Everyone is right and wrong at any given moment. Right now were all right until a brighter idea comes along and we can get it implemented to the powers that be.

That, in itself, is a feat of wonderment.

now,

put yourself in a position such as mine. I work with images day in and day out. I see all walks of life and death that comes cross my screen. When I do get a chance to process RAW files or good captures without sharpness, I need to ad some sharpness due to the simple fact that it’s distracting to view an image for long periods of time that is soft along with the migraines that come into play gazing into the monitor for hours on end retouching. There should be some initial sharpness applied to the image. The working methodology of being able to turn the sharpness on or off and/or change it, is the key to the final outcome. So you can look at capture sharpening as output sharpening if you want and change it as you see fit.

I’m not suggesting that everyone fall in line with the PG team. I’m suggesting that you use the tools as you see fit, but there are "options" to work from within the ideology that has been created.

If you don’t care about capture sharpening for you workflow, than don’t use it as capture. Call it output.

It’s just a vague as color mgmt. anyways.

It’s quite simple actually.

You choose.
MO
Mike_Ornellas
May 26, 2004
Today we mourn the passing of an old friend by the name of Common Sense. Common Sense lived a long life but died from heart failure at the brink of the millennium. No one really knows how old he was since his birth records were long ago lost in bureaucratic red tape. He selflessly devoted his life to service in schools, hospitals, homes, factories and offices, helping folks get jobs done without fanfare and foolishness. For decades, petty rules, silly laws and frivolous lawsuits held no power over Common Sense. He was credited with cultivating such valued lessons as to know when to come in out of the rain, the early bird gets the worm, and life isn’t always fair.

He watched in pain as good people became ruled by self-seeking lawyers and enlightened auditors. His health rapidly deteriorated when schools endlessly implemented zero tolerance policies, reports of six-year-old boys charged with sexual harassment for kissing a classmate, a teen suspended for taking a swig of mouthwash after lunch, and a teacher fired for reprimanding an unruly student. It declined even further when schools had to get parental consent to administer aspirin to a student but cannot inform the parent when the female student is pregnant or wants an abortion.

Finally, Common Sense lost his will to live as the Ten Commandments became contraband, churches became businesses, criminals received better treatment than victims, and federal judges stuck their noses in everything from Boy Scouts to professional sports.

As the end neared, Common Sense drifted in and out of logic but was kept informed of developments, regarding questionable regulations for asbestos, low flow toilets, "smart" guns, the nurturing of Prohibition Laws and mandatory air bags.

Common Sense was preceded in death by his parents

Truth and Trust;

his wife, Discretion;

his daughter, Responsibility;

and his son, Reason.

He is survived by three stepbrothers: Rights, Tolerance and Whiner.

Not many attended his funeral because so few realized he was gone.

Apparentelly, common sense is not too common.
L
LRK
May 26, 2004
Excellent written piece on the Death of Common Sense!

Wow Mike, that is sobering and sadly true. I got goose bumps but not the happy kind. Sad, isn’t it.
T
Todie
May 26, 2004
Thanks, Jeff!

I like Fuji’s idea of not using a simple horizontal/vertical grid (whenever I can, I use the 45° halftone angle for fleshtones), and the small sensor should help increase shadow detail. (if it doesn’t yet, it might in the future)
MO
Mike_Ornellas
May 26, 2004
Today, I have too much to say. Tomorrow, I’ll follow the sun.
AW
Allen_Wicks
May 26, 2004
Personally I reject many of the arguments for a RAW file standard (e.g. “Otherwise I cannot archive my pix”) but do accept other arguments like Jeff Schewe’s in post #54. It seems to me that a compromise could be reached.

What is the argument against a standardization of what “RAW” means? That’s easy. Digital photography is still in a growth phase. Manufacturers’ engineering departments are creating constantly, trying to think outside the box (e.g. hexagonal pixels, which may or may not be a good thing) in attempts to build better digital imagery. To limit that creativity at this stage of digicam evolution would IMO be a mistake.

A compromise that requires inclusion of parameters like the ones Jeff mentions but allows other parameters to remain free form should be doable, however. Such a standard would be helpful, but could not address all issues (individual brands would still require separate SDKs, for instance). IMO facilitating unfettered creativity at this stage of digicam development needs to be primary.

Regarding folks concerned about archiving I say get over it. Use the RAW file to create the best file you can and then Save As in a file format that will open in your next few OSs and PS versions. Then make redundant copies including off site, recopying every 5-10 years to new media as necessary. Such a workflow provides infinite secure archiving, unlike film which requires an air conditioned vault for true archiving.

And who knows? Maybe the Nikons and Canons of the world will have the foresight to provide software upgrade paths for their RAW files forever. Or maybe not.

————————-

Venicia-

What I disagreed with was your statement "…other than the Bayer interpolation to derive "3 channels" of color information, no other interpolation (that is, spatial manipulation) is done in-camera." IMO it is a mistake to view RAW files as basically raw data, because – despite the nomenclature – they are not. You may well be correct that spatial manipulation does not occur; it is the concept that RAW is raw that I object to, because IMO we must realize the RAW files are already heavily manipulated.

Much, much more very complex proprietary processing other than Bayer interpolation is applied to the raw data before it becomes a RAW file. That is a big part of the challenge of creating a RAW file standard I think.
JS
Jeff_Schewe
May 27, 2004
Allen,

You’re wrong about the long term viability of raws. . .it _IS_ both critical and crucial for one reason; software for converting those files are improving. Never before in photography have we had the opportunity to have currently shot images improve by virtue of the software made to convert them. Case in point, Camera Raw 1.0 allowed for improved raw conversions from the Canon FVU but now with CS, the exact same file could be converted while taking advantage of chromatic aberrations and substantially improved color noise reduction.

If I had done as you suggest, convert to 16 bit tiff. I could not now go back and reprocess previous images for improved conversions. In point of fact, I have about 8 gigs of processed tiffs from files that I failed to archive. . .so those raw file original are now gone forever. Yes, I have 16 bit tiffs, but I do not now have the chance to reprocess them.

Nope, the camera companies are playing loose and fast cutthroat games of market share and lens mount wars on the backs of photographers. While I’m all for innovation, I don’t want it at the cost of orphaned technologies and unreadable files a few years from now. We can take a Curtis neg from the 1800’s and make a print from an original. What if only his prints (the processed tiffs) were available now?

The raw files ARE the originals, the processed tiffs-even in 16 bit-are second generation. And, to tell you the truth, I simply don’t trust the camera companies to work in the best interests of photographers unless they are forced to. Case in point, Sony started selling a new camera last fall that captured raw. Problem was, they encrypted the raw format to keep 3rd parties from being able to process the raws. Guess what? Photographers saw that as a SERIOUS flaw in design and didn’t buy a lot of the cameras for that exact reason-no 3rd party raw support. Course, it only took a couple of months before the encryption was hacked and even now, Camera Raw is able to support the Sony raws.

Do NOT think for a moment that Nikon and Canon are keeping their raw files proprietary for any noble altruistic reasons. . .they are doing so for completely self-serving reasons. They will NOT play nice with others until they are forced to by their customers.
VL
Venicia_L_2
May 27, 2004
Allen,

"it is the concept that RAW is raw that I object to"

Um. Ok. I guess. (??)

This thread has been hijacked so many different directions, we might well be discussing the sound of one hand clapping for all I know, now.

RAW is a data form from which it’s actually possible to derive raster images. What’s the difference if the manufacturer has to apply a lot of sophisticated control of the light sensitive element in his device and the low-level voltages and responses that the electronics generate. All THAT stuff is not an image, or even its precursor. The resultant RAW file is. Do you want converters that work on the "data" prior to its becoming the RAW file?

Following your logic, gasoline suppliers should provide us with a detailed description of the distilling process and the complex organic reactions that are needed before 87, 89 and 92 octane fuel spurts out of fuel pumps. We all shouldn’t be content to just select one of those three numbers in order that our cars can run.

Would it change anything you do as a photographer if you were privvy to the way that any or all of the current DSLRs get the fact that light has fallen on their sensors into the code that comprises their RAW file format?

VL
B
Buko
May 27, 2004
This thread has been hijacked so many different directions,

thats the way it works here. always has, always will.
R
Ram
May 27, 2004
Right, Buko. Sometimes, when the arguments get rather heated here, they remind me of a remark made by the Argetinian writer Jorge Luis Borges at the time the war over the Malvinas (the Falkland Islands) was about to begin: “Dos calvos peleándose por un peine.” (Two bald-headed men fighting over a comb.)
AW
Allen_Wicks
May 27, 2004
Jeff-

Sounds like your lost RAW files are because you did not archive them, which is not my recommendation. I include RAWs in the archiving described. The saving of a non proprietary format like TIFF is just in the event that the manufacturers pull a Microsoft and make their old RAW files unreadable by newer software.

I more than fully agree with "Do NOT think for a moment that Nikon and Canon are keeping their raw files proprietary for any noble altruistic reasons. . .they are doing so for completely self-serving reasons. They will NOT play nice with others until they are forced to by their customers."

—————

Venicia-

Actually I am a strong proponent of not caring about how a manufacturer works the digital alchemy. I agree that what matters is the resultant image. In fact I do generally agree with your last post.

However I think it may be more complex than you envision. Different manufacturers I believe have different parameters they consider important in defining what a RAW file is – for technical as well as <sigh> marketing reasons. In the gasoline analogy, the fact is that octane can be achieved in multiple ways (e.g. by adding lead) that do _not_ work the same on all vehicles. Government fiat (“no more leaded gas”) made the standard more standard. The digicam tech is young enough that IMO we should let it shake out without constraining creativity too much (yet).

————–

P.S This is a discussion; for an argument go to the war thread in the Lounge. <g>
MO
Mike_Ornellas
May 27, 2004
They will NOT play nice with others until they are forced to by their customers.

cats and dogs living together……
MO
Mike_Ornellas
May 27, 2004
Good people.

RAW has to be the original.

A standard RAW file format has to be extensiable to support future developments and has to be a common language.

Think of it this way.

Do you know how difficult it is to communicate to others in a train station whereas, for example, 6 people are having a conversation, for which they all speak a common language, but they all have unique, heavy regional accents, that creates a barrier?

That’s what you are asking images to do as of now.

Alan,

I don’t want to stifle the development of technology by any means. But I’m telling ya, It’s a total mess here in the real world. Anarchy doesn’t even describe it.

We have to provide a common language that is flexible for change.

Just like it’s been since the dawn of time.

Adobe needs to create that standard because of their standing in the industry. They need to give themselves to the world without limits.

Um,

just like PDF.

um

yea,
PH
Paul_Hokanson
May 27, 2004
I’m patiently waiting for the innovation juggernaut of Microsoft and Sony to step up the plate and write a RAW standard so advanced in its capabilities that future updates will not be necessary and every manufacturer, including Adobe, will flock to implement its code into their products. It will do away with the need for color profiles and hexigonal pixel hocus pocus and automatically deliver gigabit files without interpolation no matter what size the image sensor, in an as-yet-unannounced bit depth that can actually capture the forth dimension! People of the world will embrace this RAW standard as the technological holy grail that it is and peace and harmony will ensue. It will be Bill Gates’ (and whomever happens to be CEO of Sony at the time) shining moment!

Oh yeah, it’ll be waterproof too.
MO
Mike_Ornellas
May 27, 2004
If that’s how you feel, than I can honistelly say, with confidence/ that color mgmt. is a total failure.

If anyone want’s to go head to head with this, ………lets go!
R
Ram
May 27, 2004
Sony is an absolute, total, dismal failure at attempting to write software.
PH
Paul_Hokanson
May 27, 2004
It’s not a failure at all… it just needs an innovator like Microsoft to rescue it.

::: tongue planted firmly in cheek :::
MO
Mike_Ornellas
May 27, 2004
mbad.
PH
Paul_Hokanson
May 27, 2004
psst… Ramon…

I was joking. I don’t believe Microsoft is an innovator any more than I believe that Sony can write software without issues.

🙂
R
Ram
May 27, 2004
I know, Paul. I just can’t pass up an opportunity to dump on Sony.
MO
Mike_Ornellas
May 27, 2004
Sony is too busy down sizing.

MacBook Pro 16” Mockups 🔥

– in 4 materials (clay versions included)

– 12 scenes

– 48 MacBook Pro 16″ mockups

– 6000 x 4500 px

Related Discussion Topics

Nice and short text about related topics in discussion sections