Removing Halftone How Good is Affinity Photo

[Peter1418]

Hello,

I have just joined the forum as I am contemplating purchasing Affinity Photo to help me with quite a large project. Obviously, before I purchase I need to know if Affinity will fit my needs.

First of all, I should say that I am a complete novice when it comes to photographic restoration etc. and would be very grateful for any help or advice or preferably tutelage regarding the following.

I have over one hundred and fifty old halftone newspaper pictures (actually, they are photographs of the original pictures, see attached picture). All pictures are head and shoulder shots of individual people, and I will be printing these “pictures” out at a final size of 50mm x 80mm, portrait.

I need to learn how to “shop” the pictures so that they will look at their best when re-printed on plain white paper.

I need to try and get rid of the “halftone” as much as possible and this is what attracted me to Affinity as I believe it has FFT-based “noise” removal built-in, which if I understand things correctly this would be very useful. I will also need to enhance the pictures to some degree.

If anyone with the relevant experience would like to “take me under their wing” and teach me how to get the best results I would be extremely grateful.

Kind regards,

Pete.

 

 

 

[stefan 1967]

hi peter, Affinity photo is certainly good enough, the limitation might be the users capabilities : ).  My quick approach that does not lead to a final result would be at first step use the tonal correction tool and move the sliders quite narrow, the meeting point of the two sliders will define the overall brightness. luminocity layer mode helps to get rid of most of the colour.

black and white mode and further contrast work might help,

(it is my 3rd day in AF, so there are lots of more advanced users around : )

best

stefan

 

[PixelPest]

--- didn´t understand the question in first place.

 

[carl123]

48 minutes ago, Peter1418 said:

and this is what attracted me to Affinity as I believe it has FFT-based “noise” removal built-in

This is the result of just using FFT Denoise

 

[GarryP]

FFT Denoise Filter, then Dust & Scratches Filter, then Live Clarity.
Result from just a few minutes of playing around, so there will be better ways.
An interesting problem.

[stefan 1967]

wasn´t it "au contraire"? getting rid of halftones?

[PixelPest]

Yes - I believe you´re right. I took back my suggestion. 😷

[Peter1418]

Thanks Stefan 1967, PixelPest, carl123 and Garry P for the replies and some interesting results there, pretty impressive, thank you for taking the time, it is appreciated.

Pete.

 

[- S -]

On this particular image, the FFT Denoise filter is going to look something like this.

Before:

After:

 

[Peter1418]

42 minutes ago, GarryP said:

FFT Denoise Filter, then Dust & Scratches Filter, then Live Clarity.
Result from just a few minutes of playing around, so there will be better ways.
An interesting problem.

Hello Garry, this attempt looks really promising, could it be sharpened up a little without bringing back the halftone? it is still impressive though and when printed out at 50mm x 80mm may well be close to the mark, but may need brightening up a little. Thanks for taking the time.

Pete. 

[Peter1418]

4 minutes ago, - S - said:

On this particular image, the FFT Denoise filter is going to look something like this.

 

 

Thanks for that "Free" is the denoise done automatically or do you have to "blackout" all those parts manually?

 

[- S -]

10 hours ago, Peter1418 said:

Thanks for that "Free" is the denoise done automatically or do you have to "blackout" all those parts manually?

I've changed the 'before' image to make it a bit clearer.

You're looking for recurring patterns, which are the bright spots.  If you look at the 'before' image, you can see on this image they form a sort of diamond shape around the centre.  You then use the built-in brush to paint dots over the bright spots to mask them.  Adjust the brush size so that it's just large enough to cover the main part of the bright spot – if it's too big or too small you may not get the best results.  Never paint over the centre spot.

Each image will be different.  You need to be careful about filling in too many bright spots, as sometimes you can end up removing a pattern that's supposed to be in the image – like a striped object.  Also be careful painting over the bright spots further away from the centre (like those faint ones around the edges), as that can start to have a negative impact on the image.  It can be a bit trial and error and the results can vary from image to image.

[Peter1418]

1 hour ago, - S - said:

I've changed the 'before' image to be a bit clearer.

You're looking for recurring patterns, which are the bright spots.  If you look at the 'before' image, you can see on this image they form a sort of regular diamond shape around the centre.  You then use the built-in brush to paint dots over the bright spots to mask them.  Adjust the brush size so that it's just large enough to cover the main part of the bright spot — if it's too big or too small you may not get the best results.  Never paint over the centre spot.

Each image will be different.  You need to be careful about filling in too many bright spots, as sometimes you can end up removing a pattern that's supposed to be in the image – like a striped object.  Also be careful painting over the bright spots further away from the centre (like those faint ones around the edges), as that can start to have a negative impact on the image.  It can be a bit trial and error and the results can vary from image to image.

Thanks for the information mate its appreciated. I am now thinking how much work that would be for around 150 pictures, nothings ever easy (laughs)

regards,

Pete.

[WG48]

Try the Box Blur with a radius between 4 to 6.  This will eliminate most of the half-tone.  After that, use other tools to enhance tonal quality, exposure, etc and heal cracks and spots..

[eobet]

This is super cool, but I do not exactly understand how to use it on my scan result, and tutorials seem hard to come by...

Here's a 100% crop of my original image:

 

The best I've been able to get is this, but it introduces noise where there previously basically was none:

 

 

And then I have to use a gaussian blur to get rid of the final pattern... am I supposed to be able to do this in one go, or how does it work in my case?

 

I'm also confused on how this file, which is similar but which has a predominantly different halftone raster, can have such a similar pattern in that popup window...

[John Rostron]

Painting out all those  spots is not a horrendous as it appears at first. As you paint out one spot, it will automatically paint out the equivalent spots in each quadrant.

John

 

[NotMyFault]

On 12/29/2021 at 1:30 PM, eobet said:

I'm also confused on how this file, which is similar but which has a predominantly different halftone raster, can have such a similar pattern in that popup window...

These files show the limits of FFT denoise. It might be faster to re-create the simple shapes with pen tool, and sample the gradient colors with color sampler with high radius (65x65). It looks like the halftone raster has been made intentionally strong and visible.

As quick & dirty method: use threshold to isolate (1) black lines, (2) white letters, then fill with gradient / color the colored areas (filled simple curve layer put behind).

[Fixx]

While FFT denoise looks powerful it clearly cannot resolve all halftones. In PS I would use smart blur which is really nice for halftone removal but I guess here is nothing similar in AP. 

Bottomline is of course that halftoned images cannot be really recovered fully. In B&W it is sometimes smarter to use copydot technique.