[phat_bog]

I am sharpening for print, but I don't know how much sharpening is best for print because I haven't done any hard proofing with various amounts of sharpening. C prints are softer than display, but it's not necessarily unwanted, once on print the surface looks more natural, not as sharp as display but not soft... I don't know , it's different, I'm afraid of sharpening too much for prints but every when image look crazy sharp on display they just look really good on paper prints. And all the large size prints I've done are mat finish, I have no idea if they'd look sharper on glossy, I make a test image with various amounts of sharpening printed on mat and glossy.

---------- Post added 09-12-21 at 18:39 ----------


I printed 12 x A1. Two prints aren't very good due to depth of field and autofocus. One print I don't know what happened, the camera took the shot without having locked focus (in AFS) I don't know how this was possible but the lens wasn't focused on the subject. For the other print I shot hand held trying to preserve noise, so I used a too wide lens aperture but it looked sharp on the back of the camera display, I spent lots of time to sharpen it until acceptable on display, but didn't come out quite as good on the print.

---------- Post added 09-12-21 at 18:44 ----------


Other prints were shot on tripod, long exposures or distant subject, or with more light, the lens aperture was set to f8 - f16, everything was in focus. Subject distance plays a role. For example I shot the distant dolomites with a 200 mm lens f8, I made ~300 Mpixels stitch, everything is tack sharp down to the pixel, all over the frame.

OK thanks. i understand much better, i thought it was more complex than this.
All i can recommend, thus you maybe don't need this reco, for printing large prints with sharpening in post process: "cleaning" the image is an important step before sharpening. It's sometimes fairly long to do. Topaz's "AI" will not give those great results many photographers dream of, if a proper image cleaning is not done before. Very well known 3rd party filters are nice on screens but for large prints they rarely shine.

[biz-engineer]

My rule of thumb is to double the amount of sharpening.

What does it mean to "double" the amount of sharpening? Are you referring to unsharp-mask type sharpening with radius + amount settings? Problem is, I now use Topaz AI , and I'm not sure what the sliders represent exactly. I'd be interested is any one has done MTF testing of printers and how to compensate for it.

[pschlute]

What does it mean to "double" the amount of sharpening? Are you referring to unsharp-mask type sharpening with radius + amount settings? Problem is, I now use Topaz AI , and I'm not sure what the sliders represent exactly. I'd be interested is any one has done MTF testing of printers and how to compensate for it.

Exactly that. Generally for screen with my K-1, I use USM settings of Threshold 1 Radius 0.6 amount 35%. If printing, I use 2/1.2/70%

This is a rule of thumb as some images require more or less, or even just selective sharpening.

[biz-engineer]

Exactly that. Generally for screen with my K-1, I use USM settings of Threshold 1 Radius 0.6 amount 35%. If printing, I use 2/1.2/70%

That's kinda the output sharpening presets of the Silkypix exports. There is also the difference between c prints and inkjet prints, inkjet prints are sharper. I think i'll have to always have one print batch in advance so that I can use one print out of current batch to hard proof print samples from the next planned batch.

[pschlute]

That's kinda the output sharpening presets of the Silkypix exports. There is also the difference between c prints and inkjet prints, inkjet prints are sharper. I think i'll have to always have one print batch in advance so that I can use one print out of current batch to hard proof print samples from the next planned batch.

Hard proofing is the best way to know what you will get. After a few goes you generally know what to expect. My commercial printer uses Hahnemuhle and Permajet for their fine art giclee prints. i dont use c-type.

[biz-engineer]

My commercial printer uses Hahnemuhle and Permajet for their fine art giclee prints. i dont use c-type.

Inkjet is sharper, as I could notice when I've printed the same picture (same file) on both Fuji Crystal Archive and Canon inkjet. C-type doesn't show as crisp but all tiny details are there, it looks really different from display. Now I've got a 4K monitor, it shows some kind of artifacts, aliasing lines at edges, and noise is visible. On paper, the noise is subdued, and the aliasing lines are gone, but all fine details are there, which give a different look, perhaps closer to how film photos looked like.

[Dartmoor Dave]

C-type doesn't show as crisp but all tiny details are there, it looks really different from display. Now I've got a 4K monitor, it shows some kind of artifacts, aliasing lines at edges, and noise is visible. On paper, the noise is subdued, and the aliasing lines are gone, but all fine details are there, which give a different look, perhaps closer to how film photos looked like.

I agree. I think chromogenic prints have a much more film style look -- perhaps inevitably because of the way they're made -- and that's why I prefer them. I've tried inkjet giclee prints from reputable labs, but the results always look too digital for my taste.

Relevant to this thread, I'm wondering how much the different printing technologies will affect the apparent depth of field. Surely whether a point of light will appear as a point or a blur disc must be affected by whether it's printed using dots of ink or grains of chemicals, even when the resolution of the digital original used is the same?

[biz-engineer]

I'm wondering how much the different printing technologies will affect the apparent depth of field. Surely whether a point of light will appear as a point or a blur disc must be affected by whether it's printed using dots of ink or grains of chemicals, even when the resolution of the digital original used is the same?

It's complicated. When I sharpened the image before uploading to the lab for prints, I sharpened selectively using a mask because I wanted to recover some sharpness in out of focus areas but without sharpening what was obviously out of focus, so the transition from in focus to out of focus is a bit weird. Basic sharpening works well if the de-focus blur is similar to softness from the lens optics or softness from diffraction, but as soon as the de-focus blur is significantly softer than other factors, sharpening it doesn't look good. Then the whole thing get softened by the print process and the final result lacks consistency because selective sharpening broke the OOF transition. So there is really something like a perceived depth of field, and it's all relative to other factors that cause softness.

[Ian Stuart Forsyth]

It's complicated. When I sharpened the image before uploading to the lab for prints, I sharpened selectively using a mask because I wanted to recover some sharpness in out of focus areas but without sharpening what was obviously out of focus, so the transition from in focus to out of focus is a bit weird.

You also have to be careful of sharpening spurious detail. you will often see this in people who sharpen feather detail to make it look sharper and in the process they also apply sharpening to the spurious detail making it look sharp while it should look more like out of focus transition in the feathers.

The very same can happen in water repels and other repeating patterns as they fall out of focus

[biz-engineer]

You also have to be careful of sharpening spurious detail. you will often see this in people who sharpen feather detail to make it look sharper and in the process they also apply sharpening to the spurious detail making it look sharp while it should look more like out of focus transition in the feathers.

I understand that. And I now beginning to think it's better to sharpen the whole frames with various amounts and then use gradual layer masking to blend sharp and OOF areas. Masking transitions in Topaz Sharpen AI aren't soft enough to maintain a natural look between sharp and OOF areas.

[Dartmoor Dave]

I now beginning to think it's better to sharpen the whole frames with various amounts and then use gradual layer masking to blend sharp and OOF areas.

That's the way I do it when I need selective sharpening. I make a duplicate Photoshop layer and sharpen the entire layer, then I can use a mask to control which parts of the photo it applies to, and I can adjust the opacity of the layer to reduce the amount of sharpening if needed.

Using layers and masks would allow you to sharpen some areas with Topaz AI and others with a different sharpening method, depending on what suits different parts of the photo.

[biz-engineer]

Using layers and masks would allow you to sharpen some areas with Topaz AI and others with a different sharpening method, depending on what suits different parts of the photo.

You are correct, I can get the best of both worlds, best sharpening algorithms from Topaz and most advanced layer masking & soft brushes from layering tool (I'll use GIMP, not as good as PS, but the masking is a lot better and a lot faster than in Topaz).

[bobbotron]

So, in practice, there isn't only one stop difference between FF and APSC , but it's two stops. Taking another example, from FF image to Phase One image printed 2x larger than the FF one (crop factor 0.5x), for the Phase One to match the FF depth of field its lens should be stopped down by as much as 4 stops, that's huge. And from FF to 4x5 (crop factor = 0.27), for making 3x larger prints out of the LF camera, I would have to stop the LF lens by 8 stops from FF aperture values (e.g FF lens aperture f/5.6 => LF lens aperture ~ f/100), leading to super slow shutter speeds in the LF camera. Did I understand correctly?

This seems like you're really twisting yourself into loops here. Are we doing math or taking photos?

[biz-engineer]

This seems like you're really twisting yourself into loops here. Are we doing math or taking photos?

You'r right, I went to far with the theory. Yet, there is something valuable there, it comes down to the nature of geometry:

- small prints of large things: very easy

- larger prints of larger things: easy,

- smaller prints of smaller things: easy,

- larger prints of smaller things: difficult

[StiffLegged]

You'r right, I went to far with the theory. Yet, there is something valuable there, it comes down to the nature of geometry:

- small prints of large things: very easy

- larger prints of larger things: easy,

- smaller prints of smaller things: easy,

- larger prints of smaller things: difficult

Professor Northrup just imploded - you stole his thunder for the next video series. Ah well.