sharpener workflow with genuine fractals

[tmbrbeest]

what is the proper workflow when using photokit sharpener with genuine fractals for upsizing and downsizing images?

do i (output) sharpen first and then convert to genuine fractals format or vice versa?

i normally convert from 16-bit RAW to 8-bit JPEG at the last minute so i can print on fuji crystal archive paper. i typically use contone 300. would i use the same setting or is there dependency on size of the print?

anyone have any suggestions?

[Wendell Rider]

I'm new here and I have recently been doing a lot of tests of upsampling and sharpening techniques. I'm interested in the same question but it has been my experience that GF is hard to sharpen because of the kind of file it produces and/or some smoothing that goes on in the process. GF recommends to not sharpen first. I am using a Canon 300D (6.3 mp) and trying to get to 16 inch wide prints at 240 dpi for my Epson 4000. (I have had great results at 150 and 180 dpi without upsampling but I want to go bigger.) So far I have found that using Photokit Sharpening with Bicubic Smoother for upsampling works better than GF in that range and I haven't even tried using the Capture at the original size (just saw it in the forums) before I upsample- which is what I am going to do right now.
Anyone else have thoughts on this?

[Bruce Fraser]

I find that the combination of my cameras, lenses, imagery, and taste is best served by uprezzing from an image acquired at native resolution and capture-sharpened at that resolution, but a good many photographers whose judgement I trust have expressed a preference for acquiring at the maximum resolution they can get from the raw converter, which is often considerably higher than the camera's native resolution. So I suggest testing both ways.

[Wendell Rider]

I just did a print on my Epson 4000 starting with a 240 dpi file (8.5 inches wide) from C1 LE. I did the Capture sharpening on that file this time before upsampling to 16 inches at the same resolution. I then continued with processing and it came out great. Somehow I missed the step of capture sharpening first before. What a difference. Now I may try the process of doing the upsample in C1, but do you have any thoughts about the difference betyween C1 and ACR. To me, I like the color better on C1 and it seems to have less shadow noise, even if noise reduction is turned down as far as it will go, which I understand is NOT all the way off. ACR seems to be more of a "brown" tint to me and I like the "bluer?" look of C1 better. Some people say that C1 is not as good as ACR in upsampling RAW images. Any thoughts.
And by the way, Bruce, I have a couple of your books and i have followed your advice on the Colorsync forum. I think this kind of feedback for a proiduct or subjects in general is just fantastic.
I have had a lot of luck with my landscape images, lots of fine detail, using a dual edge sharpening with USM, light and dark layers at about 500, .3 and 1 instead of the creative step and I like the surface texture and fine detail it gives me. The nice thing is, that if the picture has too much noise, I can do the supersharpener or edge sharpener instead. And of cours, all this stuff is adjustable later on so I can really tweak it. This is a great tool and i recommend it to everyone.
Thanks for your advice.

[tmbrbeest]

thanks for your comments but i'm not sure they've fully resolved my questions regarding sharpening, resolution & resizing workflow. let me clarify my question.

in the RAW conversion process, there is an opportunity to select the "best" or "correct" image size and resolution. i typically convert to 16-bits, 600 dpi, and the largest output size available from ACR. my thinking is that i should at least start out with the best image quality available to me before i do my editing so therefore "bigger is better", "more is better", down-sampling is better than up-sampling, and if i have to scale up, then i have sufficient dpi headroom to work with. so i definitely follow the camp bruce mentioned that goes for the maximum resolution at time of RAW conversion. i then do my capture sharpening after the RAW conversion and create a master file after all of my other edits (e.g., color correction, creative, etc.). this part of the workflow i got down.

with this master file, i can simply resize and (output) sharpen for different types and sizes of prints later. my questions really revolves around this latter part of the workflow. output sharpening and resizing are typically some of the last steps i do before i convert to an 8-bit JPEG to send to a lab.

Q1. would image quality be better if i were to output sharpen first and then resize or resize first and then output sharpen (regardless of whether i use genuine fractals or photoshop)? does the sequence of steps matter? my thinking is that the latter should produce better results(?).

Q2. the nik sharpener pro product factors in the final print size as a variable in setting the degree of sharpening. this sounds logical to me. why doesn't photokit sharpener factor in the final print size? i typically select contone 300 since i'm printing on fuji crystal archive paper with a noritsu printer.

Q3. are there any benefits to printing at a higher dpi than what the printer calls for (i.e., printing at 600 dpi on a noritsu printer that only asks for 300 dpi)?

Q4. are there any benefits to downsampling from a 600 dpi image vs downsampling from a 300 dpi image? does the extra dpi provide more information and make a better interpolation or am i just wasting space?

Q5. wendell, can you convert a .STN (genuine fractals) file to .PSD format and then sharpen it? would this resolve the issue of sharpening in genuine fractals or is it an inherent issue with the file once it's been converted to .STN?

[Bruce Fraser]

Q1. Output sharpen should always be done at final print rez.

Q2. The final print size isn't important. The final print resolution is. We can only sharpen the pixels?we can't control how the printer turns them into dots. We're aiming for a sharpening halo that's big enoug to be effective, but small enough to be indistinguishable as a separate feature. That depends on print resolution, but it's the same for a 4x6 at 300 ppi and a 20x30 at 300 ppi.

Q3. With Contone printers, no, there's no benefit to sending more data than they ask for.

Q4. There's no simple answer here. I've seen a benefit to scanning 35mm neg at 8000 ppi, applying some grain reduction and sharpening, then downsampling to 4000 ppi vs scanning at 4000 ppi. (Whether it's a big enough benefit to be worth the hassle is an open question.) I haven't seen any benefit to acquiring a digital raw at a size larger than native, then downsampling to native rez.

Q5. Haven't used GF for years, sorry.

[tmbrbeest]

ok, thanks bruce.

these were the explicit answers i was looking for. i can now refine my workflow.