How I feel after about a year of using PL4 as my go-to image editor

Hi Mike

Well done on your latest shot but, yes, I am going to suggest a few things :nerd_face:

You have certainly taken a “minimalist” approach, which is always a good idea, with only a grad filter over the sky and the addition of ClearView Plus overall.

Let me start by suggesting that global ClearView Plus wasn’t really necessary and, IMHO, has added too much contrast in places where it wasn’t really needed, like the reflections on the water.

The problem with the grad filter is that it has affected, not only the sky but, also, the trees that “intrude” into the filtered area.

I summary, what I have done instead is to remove both those adjustments and, for the sky, use Control Points to select the sky, with negative control points to stop it affecting the rest of the image, especially the reflections.

For the settings for the sky control points, I have reduced the exposure (like you did), added a tad of micro-contrast and boosted the vibrance

In terms of global adjustments, this is all I did

So, here is a jpeg export of my version and the DOP file, including my version

L1002943 | 2021-07-25-Walkabout.dng.dop (25,2 Ko)

Of course, it may not be what you wanted to achieve but it’s intended more to give you ideas of other ways of working.

Oh, and i installed the lens module and applied the four essential lens corrections of vignetting, lens sharpness, chromatic aberration and distortion - all at their “automatic” levels

I wish I knew enough to know what I wanted - I am much better at recognizing what I do NOT want. It’s obvious to me once again that I could have achieved the effect I was after using control points, which would only work on those parts of the image that needed it (sky) and not mess with other parts of the image (tree) that were fine to begin with. The biggest thing on my mind was not over-doing things. As a test, I tried out some of what you wrote, and yes, I prefer the results your way.

I am confused about you installing the lens module and applied the four essential lens corrections. When I first use this camera and lens, I thought all of that was installed after the first time I got a notification. So, if I go into my settings, I should be able to see where they are NOT applied, and check on something so from now on, all photos with this lens would be applied…

But there is a catch. These are all ancient Leica lenses, without the newer markings that tell the camera which lens is in use. I may very well have had a lens on the camera, and not manually set the camera so it knew which lens I was using. I never thought of that. I guess from now on, I need to verify that my camera “knows” which lens I am using, so the EXIF data is correct, and so PL4 will know, and use the appropriate settings for the lens I was really using.

I think you suggested these things long ago, but it’s so easy to just make a global adjustment - but I agree with you, using control points lets me make changes in the areas I am interested in, and not with the whole image.

Since them, I have bought a Plustek OpticFilm 8200i Ai scanner, and started shooting B&W film again in my old Leica M3 (now off to DAG Camera for making it like new) and my equally old Leica M2 (will also go off to DAG when I get my M3 back). I’ve been posting my images in this forum:
https://www.l-camera-forum.com/topic/324032-scanning-leica-bw-film-plustek-optifilm-vuescan/

First I scanned new b&w images I’ve been taking with my old Leica film cameras, just like I did in the 1960’s. Then I started going through my old negative collection. Some of my negatives are in very poor condition, but most are fine. I set the scanner to the same resolution mentioned in the scanner specifications, 7200 dpi.

My original question was that when I switch back and forth between the “hand” tool, and the “correction” tool, the image jumps around right or left, rather than staying put like it did at first. Is there a number of corrections one can make on a single negative, before PL4 gets “overloaded”?

I tried to upload the original TIF image here, but got this error message:

Screen Shot 2021-09-03 at 13.11.49

The difference between this image, and the others I’ve been editing recently, is that I’ve made what. feels like several hundred “dust corrections”.

Maybe the file size is the reason why the image seems to jump around between using the hand tool, and making a dust correction?

I tried another image from the same roll, which needed far fewer corrections. I scanned all these at a resolution setting of 7200 dpi. Maybe I should switch to 3600 dpi? That should make a huge difference in the file size.

Does PL4 have a maximum file size? Am I just creating the problem by using 7200 dpi on a 35mm negative?

I found dust removal in PL4 to be… challenging. Which is why I clean up my negatives in Affinity Photo now. It, too, can sometimes lag a little, but it always catches up and does a fine job 99% of the time with the Inpainting brush.

BTW, 7200dpi seems rather high. When I was researching prior to scanning my father’s negatives I found a general consensus that 35mm negatives should be scanned at ~4000dpi for “archival scans”. I therefore scan negatives at the nearest figure my scanner can do, 4800dpi.

Thanks for the advice. I went from 7200dpi to 3600, and can’t see any difference. The file sizes are now much more reasonable. I think I’ll stick with 3600.

About PL4 and Affinity, I have both, but I’ve been trying to do everything in PL4. Dust removal is …challenging, as you wrote, but PL4 does an excellent job of correcting the issue and not messing other things up. It all just comes down to the time it takes, doing every speck of dust individually. The Silverfast software is supposed to do dust removal very well with the Plustek, but only with color negatives, not B&W.

Here’s an example, with my corrections:
2021-09-03-0003.tif (32.5 MB)

2021-09-03-0003.tif.dop (65.8 KB)

…and a reduced size result:

2 Likes

Looks good! Might I suggest you give your watermark a small offset from the right and bottom edges? The final ‘y’ gets a bit lost in the corner. Great font for this sort of imagery.

What resolution you scan at depends on what use you are going to make of the scanned file.

For printing and viewing at arm’s length, the output file should be 240ppi.

When I scan my 5" x 4" film, I scan at 2400ppi, which then gives me a 10x enlargement to 50" x 40" - adequate for most exhibition spaces.

So, if you were to scan at that resolution, a 35mm film image is (approx) 1" x 1½", which would enlarge to a 10" x 15" print.

Taking into account the grain size of the film, 2400ppi tends to be a sweet spot where you don’t get too much grain, which can look as ugly as digital noise, or too little detail and lose the grain structure completely.

As @zkarj says, dust removal is not PL’s forte. Because, as you noticed, things start to get graunchy, the more you repair, I tend to do so much, then export the file to the same format and resolution (in your case TIFF) and start again.

Your image

I noticed a couple of problems with what you have done in this example:

  1. you are not always moving the source of the repair to somewhere that doesn’t contain “rogue” information…

Notice the bits of wire in mid-air.

It looks like you may have made geometric corrections before de-spotting, otherwise, I wouldn’t expect to see so many dust corrections outside of the finished frame size…

It is well known that geometric corrections can displace repairs; and it also saves you a lot of time and effort not having to de-spot unnecessary areas.

This next screenshot shows your horizon correction…

You have had to make a guesstimate for a horizontal line, whereas you could have made a vertical line on the telegraph pole much easier.

Apart from the de-spotting, I reworked your image, removing some adjustments and adding others. Most usefully, I used the Fuji Acros 100 film rendering and the fine contrast sliders for midtones and shadows, to bring out detail.

Here’s a quick export…

And here’s the combined DOP file…

2021-09-03-0003.tif.dop (127,4 Ko)

As someone who is notoriously prone to ‘off-kilter’ shooting I use the horizon tool a lot and while some poles can make a good vertical reference, I find many are not vertical at all! Sometimes it’s obvious. Other times the whole picture doesn’t look right after using one as a reference. In general, wooden poles are less straight than concrete which are less straight than steel. I often judge suitability by context. A steel light pole at an airport is far more likely to be vertical than a wooden signpost, or indeed a wooden telephone pole.

This is why I’ve asked/voted for a drag-rotate ability in the horizon tool. Sometimes, ‘what looks right’ is the best answer and being able to drag to achieve that is ideal.

1 Like

Indeed. Then you’ve got the receding verticals syndrome where you straighten them out and now the building looks like it has diverging verticals. In that case, the advice is to always leave some convergence because it “looks right”. The same applies to “horizontality” you do it until it looks right, regardless of what you see when you place horizon line.

In this case, Mike seems to have got it as he wanted it using a horizontal and, in my rework, I used the pole - it looks like we were both right :nerd_face:

Dear Joanna,
would you suggest scanning old slides with lower resolution if it’s not so good material.
So if the company offer 3 settings waht to choose. I need them only for small prints in booklets, and the slides where made with Fuiji material and sometimes with 1660 Iso films

  • Eco-Qualität 2.400 dpi (ca. 7 Mega-Pixel) ab 0,08 €

  • Standard-Qualität 3.900 dpi (ca. 19 Mega-Pixel) ab 0,12 €

  • Top-Qualität 5.000 dpi (ca. 32 Mega-Pixel) ab 0,18 €

Thanks in advance

Guenter

If the maximum printed size is less than 24cm x 36cm, there is no point in scanning at anything more than 2400ppi. I know 7Mpx seems like a low resolution file but, don’t forget, a 24 Mpx image can be printed up to 63cm long without having to use interpolation.

Dear Joanna,

tahnks for the first info :smiley:

What do you think. is scanning with higher resolution also a pro/contra for the visibility of grain?

As often, I propose to have a test batch run by your provider and decide after you’ve seen the results. Maybe a few images want a crop and some rotation and then, a few extra pixels can come in handy.

1 Like

So much to reply to, and so many questions… In order…

1 - I deliberately put the watermark in this location - nobody else does it this way, so what I did is “more unique”. I rather like it now, maybe because I’ve gotten used to it. It also leaves more of the image as-is, not being over-written. I’ve been doing it like this for about a year, and have never found anyone else doing it this way, which is a good thing.

2 - Resolution - I don’t know what the eventual purpose for the image might be, so I decided to treat the negative as if I want to make a digital copy, and later I would decide what might be done to/with it. The scanner specs noted 7200dpi, so that’s what I started using, but the files were too big, so I went to 3600. My choices from the drop-down menu are Auto, Custom, 7200dpi, 3600dpi, 1800dpi, and 900dpi. Are there any good reasons not to set it at 3600 and forget it? You are suggesting 2400dpi. Maybe I should just select 2400 until/unless I might have a reason to do something unusual. Again, right now, I’d just like to make a digital copy of the negative, so I won’t need to go back to the film and do it again. The most likely use is to post the image on my SmugMug gallery, and people can select any size at which they want to view the image.

3 - How do I get to view the source of a repair, as you show in your image? That’s one of the questions I forgot to ask. With the information you show, it’s obvious. I undid mistakes when I noticed them, but obviously I missed some. Can I turn “on” those before/after circles?

4 - In the original image the station at the left was leaning one way, and the pole at the right was leaning the other way. I tried using the “parallel” tools to make both of them vertical, and I used the horizon tool at some point to make the engine look “level”. The pole must have been at an angle, and what I paid the most attention to was the engine and freight cars. The final cropping at the bottom came much later, when I decided that area I cut away was too distracting.

5 - “It is well known that geometric corrections can displace repairs; and it also saves you a lot of time and effort not having to de-spot unnecessary areas.” ----- I had no idea. I’ll remember this for the future. Yes, I made a guesstimate of the horizon, but I didn’t trust the telegraph pole - I tried the telegraph pole, and the whole image was tilted as a result. Everything came down to a guesstimate. …but I was very wrong about this - without the telegraph pole, fixing the tilt from my camera would have been trivial. Had I realized then that the pole was tilted, that would have saved a lot of needless work. I’ll definitely remember this for the future, and maybe re-do the image from the beginning.

6 - Your “quick export” is a rather pleasing image, but it looks too “shiny”. Knowing the engines were anything but shiny, and especially the tank car, it looks “wrong” to me. On the other hand, the tilted telephone pole looks just like what it was, tilted. For most people, your “quick export” is a better image, but the “shiny-ness” bothers me. I dunno. For most people, what you did makes for a better image. How did you do this? …the tank car in my image looks realistic to me, and in the quick export it doesn’t, unless maybe it had been raining. Maybe I will split the difference.

1 Like

There is a drag rotate in the horizon tool. The slider right in the Horizon palette. A way to make the slider more sensitive to small adjustments when holding down option or shift. I’d like to have a modifier key which would work on almost every slider to making small adjustments much easier. I’m often forced to type in a value for the final adjustments.

6 - Your “quick export” is a rather pleasing image, but it looks too “shiny”. Knowing the engines were anything but shiny, and especially the tank car, it looks “wrong” to me. On the other hand, the tilted telephone pole looks just like what it was, tilted. For most people, your “quick export” is a better image, but the “shiny-ness” bothers me. I dunno. For most people, what you did makes for a better image. How did you do this? …the tank car in my image looks realistic to me, and in the quick export it doesn’t, unless maybe it had been raining. Maybe I will split the difference.
[/quote]

I find the detail and realism in your version much more persuasive Mike (@mikemyers).

This would be a huge improvement!!! Instead of constantly guessing, I could visibly adjust the image until it looked best. I find this to be very annoying in PL4, as I’m not that great at always having the camera perfectly horizontal!!!

1 Like

If it’s only for viewing on a screen, then the resolution really only needs to be about 120ppi but a scan at that resolution is going to be rubbish. 2400ppi will be fine.

Check the “show masks” box

Then, depending on the size and position of the dust/hair, I tend to use Clone rather than Repair in tight spots, well zoomed in.

Heres a hair at the front of the loco (zoomed in to 200%)

Capture d’écran 2021-09-04 à 16.15.49

The problem with the repair that you did was that you ended up with a repeat pattern

Capture d’écran 2021-09-04 à 16.11.56

Here’s the source of that repeat

Capture d’écran 2021-09-04 à 16.10.12

I moved the source up and left a tad

Capture d’écran 2021-09-04 à 16.12.52

Then I deleted your repair and working close in, I used the Clone tool instead of the Repair tool and did this

Capture d’écran 2021-09-04 à 16.15.19

I would have used the 8 points tool

Capture d’écran 2021-09-04 à 16.33.15

… to take care of both the verticals and the horizontals with one tool

We’ve had this conversation before :wink: My adjustments were only ideas but you might like to investigate the selective fine contrast tools a bit more.

I will turn on the “show masks”, and hopefully it will show me the information you’ve posted.

I never used the “8 points tool” yet - sounds like a better way than what I’m now doing.

I need to read up on the difference between “clone” and “repair”. I thought repair meant the software would “see” the white spec, and make it more like the surrounding area. I guess I was wrong.

I need to find the “selective fine contrast tools” - how do I access them?

Sort of. it’s like an “intelligent” version of cloning that can blend in the repair better if it is complex or highly textured. Cloning is good old-fashioned stick this on top of that.

Click on the little + sign

Capture d’écran 2021-09-04 à 16.59.00

Capture d’écran 2021-09-04 à 16.59.11

Basic difference is replacement strenght.
You can test this your zelf.
Find a pole in the sky or other background and try repair with a blob /circle acces show mask and try to replace the pole with sky. You will see that by repair it stayes glansing through whatever slider you move.
Cloning has the ability to replace 100% the original pixels.

You need Filmpack for that.