Restore Old Photo

[cgleeson]

New to Affinity Photo - steep learning curve but enjoying the challenge and getting some excellent results. Beginning to restore old photos.

Any suggestions on how to remove the markings on the old B&W photo below? FFT Denoise does not seem to work.

Thanks.

[h_d]

Hi @cgleeson and welcome!

One way, probably not perfect:

Box Blur live filter + Brightness/Contrast live adjustment + Add Noise live filter (the latter for a bit of grit/grain, you may prefer to leave it out).

I've also converted to greyscale.

.afphoto format file attached.  You can adjust to taste.

Cheers,

H

oldphoto.afphoto

[GarryP]

In addition to the advice from h_d above, the FFT Denoise filter has been designed to help you to remove periodic noise.
Unfortunately the ‘noise’ in your photo isn’t particularly ‘periodic’ and seems to be semi-random.
One of the photo restoration experts in the forums may be able to help you in identifying the cause of the problem and, from that, you may be able to find the best solution.

[cgleeson]

52 minutes ago, h_d said:

Hi @cgleeson and welcome!

One way, probably not perfect:

Box Blur live filter + Brightness/Contrast live adjustment + Add Noise live filter (the latter for a bit of grit/grain, you may prefer to leave it out).

I've also converted to greyscale.

.afphoto format file attached.  You can adjust to taste.

Cheers,

H

oldphoto.afphoto 1.75 MB · 2 downloads

So much better than my attempts. The Box Blur live filter is quite effective. Thank you ever so much.

I'll have a play with the added noise when I get to print the photos. It intrigues me!

I've used a high pass filter to make it sharper.  Any other suggestions to sharpen it more??

Thank you again it is already much better than the original.

Chris

[cgleeson]

1 hour ago, GarryP said:

In addition to the advice from h_d above, the FFT Denoise filter has been designed to help you to remove periodic noise.
Unfortunately the ‘noise’ in your photo isn’t particularly ‘periodic’ and seems to be semi-random.
One of the photo restoration experts in the forums may be able to help you in identifying the cause of the problem and, from that, you may be able to find the best solution.

Thank you Garry, that makes sense.

I'm new to this kind of editing-never used Photoshop so there's plenty to learn.

Chris

[h_d]

30 minutes ago, cgleeson said:

Any other suggestions to sharpen it more??

To my eye High Pass is probably your best bet, but a lot of it is down to fiddling around until you get something you like (politely called "experimentation"). I tried Clarity without much luck. At the end of the day there isn't much in the original to sharpen. 

[cgleeson]

2 hours ago, h_d said:

Hi @cgleeson and welcome!

One way, probably not perfect:

Box Blur live filter + Brightness/Contrast live adjustment + Add Noise live filter (the latter for a bit of grit/grain, you may prefer to leave it out).

I've also converted to greyscale.

.afphoto format file attached.  You can adjust to taste.

Cheers,

H

oldphoto.afphoto 1.75 MB · 2 downloads

Thank you. 

 

24 minutes ago, h_d said:

To my eye High Pass is probably your best bet, but a lot of it is down to fiddling around until you get something you like (politely called "experimentation"). I tried Clarity without much luck. At the end of the day there isn't much in the original to sharpen. 

Yes, I think it's a not very good copy of the original that no longer exists.

Thank you once again. Much appreciated.

[cgleeson]

22 hours ago, h_d said:

Hi @cgleeson and welcome!

One way, probably not perfect:

Box Blur live filter + Brightness/Contrast live adjustment + Add Noise live filter (the latter for a bit of grit/grain, you may prefer to leave it out).

I've also converted to greyscale.

.afphoto format file attached.  You can adjust to taste.

Cheers,

H

oldphoto.afphoto 1.75 MB · 4 downloads

Just printed the whole photo with your suggestions. For anyone trying to improve photos in this type of condition, it is worth experimenting with the Noise Live filter. Doesn't look particularly good on the screen but does give the print what I would call extra zing.

Thanks once again H.

The complete photo has a lot of emotion attached to it and you have allowed me revitalize it.

C

[h_d]

Thanks @cgleeson, very pleased to be able to help.

[firstdefence]

I found this technique on Stack exchange a while ago and can sometimes work on random (non-FTT) patterns. It does however require the original so not much good in this instance but maybe good for future reference. Easily translated into Affinity

Quote

1. Scan the photo once as usual.
2. Rotate the photo 180% on the scanner and scan again.
3. In Photoshop, un-rotate the second scan.
4. Import it as a layer on top of the first scan.
5. Auto-Align Layers using Photoshop command.
6. Assign second scan 50% opacity to blend images together.
7. This technique comes from observing that the highlights and shadows of the photo paper texture are largely reversed when scanned from the opposite direction. Blending two such scans together cancels out most of the texture this way.

https://photo.stackexchange.com/questions/23445/what-is-the-best-way-to-remove-texture-from-a-scanned-textured-photo-paper

[John Rostron]

21 minutes ago, firstdefence said:

I found this technique on Stack exchange a while ago and can sometimes work on random (non-FTT) patterns. It does however require the original so not much good in this instance but maybe good for future reference.

https://photo.stackexchange.com/questions/23445/what-is-the-best-way-to-remove-texture-from-a-scanned-textured-photo-paper

I have found that this technique also works to some extent on scanned slides.

John

[cgleeson]

3 hours ago, firstdefence said:

I found this technique on Stack exchange a while ago and can sometimes work on random (non-FTT) patterns. It does however require the original so not much good in this instance but maybe good for future reference. Easily translated into Affinity

https://photo.stackexchange.com/questions/23445/what-is-the-best-way-to-remove-texture-from-a-scanned-textured-photo-paper

3 hours ago, firstdefence said:

I found this technique on Stack exchange a while ago and can sometimes work on random (non-FTT) patterns. It does however require the original so not much good in this instance but maybe good for future reference. Easily translated into Affinity

https://photo.stackexchange.com/questions/23445/what-is-the-best-way-to-remove-texture-from-a-scanned-textured-photo-paper

 

3 hours ago, firstdefence said:

I found this technique on Stack exchange a while ago and can sometimes work on random (non-FTT) patterns. It does however require the original so not much good in this instance but maybe good for future reference. Easily translated into Affinity

https://photo.stackexchange.com/questions/23445/what-is-the-best-way-to-remove-texture-from-a-scanned-textured-photo-paper

My copy of the photo is a hard copy, just not the original so I will experiment with this tomorrow and report how it goes. Thank you. (I am most impressed by the help given by this site.)

 

[cgleeson]

3 hours ago, firstdefence said:

I found this technique on Stack exchange a while ago and can sometimes work on random (non-FTT) patterns. It does however require the original so not much good in this instance but maybe good for future reference. Easily translated into Affinity

https://photo.stackexchange.com/questions/23445/what-is-the-best-way-to-remove-texture-from-a-scanned-textured-photo-paper

Thank you to first defence also.

[Paul Mudditt]

@cgleesonMy take on this:-

1. Luminescence adjustment noise reduction live filter layer, replicated 3 times. Use in-painting brush to correct minor blemishes.

2. Open in free Remini iPad app to sharpen using AI (Sorry don’t know of a PC solution)

3. Open in free Colorify iPad app to artificially colour. (Sorry don’t know of a PC solution)

 

 

SharpenBoy.afphoto

[cgleeson]

7 hours ago, Paul Mudditt said:

@cgleesonMy take on this:-

1. Luminescence adjustment noise reduction live filter layer, replicated 3 times. Use in-painting brush to correct minor blemishes.

2. Open in free Remini iPad app to sharpen using AI (Sorry don’t know of a PC solution)

3. Open in free Colorify iPad app to artificially colour. (Sorry don’t know of a PC solution)

 

 

SharpenBoy.afphoto

Thanks Paul. The improvement in skin texture is extraordinary. I'm now thinking of getting an iPad.

c

[cgleeson]

18 hours ago, firstdefence said:

I found this technique on Stack exchange a while ago and can sometimes work on random (non-FTT) patterns. It does however require the original so not much good in this instance but maybe good for future reference. Easily translated into Affinity

https://photo.stackexchange.com/questions/23445/what-is-the-best-way-to-remove-texture-from-a-scanned-textured-photo-paper

Attached is first fairly quick attempt at this technique. Looks quite promising.

C

[firstdefence]

Big improvement, definitely a workable solution.

@Paul Mudditt results are freaking me out lol!

[Paul Mudditt]

11 hours ago, cgleeson said:

Thanks Paul. The improvement in skin texture is extraordinary. I'm now thinking of getting an iPad.

c

 

11 hours ago, cgleeson said:

Thanks Paul. The improvement in skin texture is extraordinary. I'm now thinking of getting an iPad.

c

Affinity is fabulous on iPad with Apple Pencil. There are a few things missing so very occasionally still use the desktop but using iCloud it is easy to start working on PC/MAC then finish off on iPad and vice versa of course.

 

[GarryP]

While the generated face is interesting – and quite striking – I feel it is necessary to give a warning to anyone thinking of using this sort of ‘face generation’ software.

Since the face in the ‘enhanced’ image is generated as a ‘guess’ by the software it cannot be guaranteed to be a good match of the person’s actual face and, as such, may create mistakes/confusion/problems in the future.

For example, someone looking at the photo in a few years time may assume that it was the actual person’s face and not think that it was ‘generated’. This could lead them to mistaken conclusions about who is actually in the photo. Do they have their father’s eyes or those of Uncle Bob who we always thought was hanging around their mother a bit too often?

Or, to take that a bit further, imagine that the generated face closely matches that of someone who has committed a crime. The person who has been given the generated face in the ‘enhanced’ image could be accused of that crime even if they have no involvement in it. And all because someone had a bit of fun getting a computer to guess a face.

Obviously this is a stretch of the imagination, but it’s something that I think should be taken into consideration.

[Belifant]

9 hours ago, GarryP said:

While the generated face is interesting – and quite striking – I feel it is necessary to give a warning to anyone thinking of using this sort of ‘face generation’ software.

Since the face in the ‘enhanced’ image is generated as a ‘guess’ by the software it cannot be guaranteed to be a good match of the person’s actual face and, as such, may create mistakes/confusion/problems in the future.

For example, someone looking at the photo in a few years time may assume that it was the actual person’s face and not think that it was ‘generated’. This could lead them to mistaken conclusions about who is actually in the photo. Do they have their father’s eyes or those of Uncle Bob who we always thought was hanging around their mother a bit too often?

Or, to take that a bit further, imagine that the generated face closely matches that of someone who has committed a crime. The person who has been given the generated face in the ‘enhanced’ image could be accused of that crime even if they have no involvement in it. And all because someone had a bit of fun getting a computer to guess a face.

Obviously this is a stretch of the imagination, but it’s something that I think should be taken into consideration.

You are not wrong, but I tried it with a few of my own photos, and the results are remarkable!

I didn't notice any strange altering of facial features, except having nicer and smoothed skin

Very surprised how well it works

[Paul Mudditt]

On 10/8/2020 at 9:46 AM, GarryP said:

While the generated face is interesting – and quite striking – I feel it is necessary to give a warning to anyone thinking of using this sort of ‘face generation’ software.

Since the face in the ‘enhanced’ image is generated as a ‘guess’ by the software it cannot be guaranteed to be a good match of the person’s actual face and, as such, may create mistakes/confusion/problems in the future.

For example, someone looking at the photo in a few years time may assume that it was the actual person’s face and not think that it was ‘generated’. This could lead them to mistaken conclusions about who is actually in the photo. Do they have their father’s eyes or those of Uncle Bob who we always thought was hanging around their mother a bit too often?

Or, to take that a bit further, imagine that the generated face closely matches that of someone who has committed a crime. The person who has been given the generated face in the ‘enhanced’ image could be accused of that crime even if they have no involvement in it. And all because someone had a bit of fun getting a computer to guess a face.

Obviously this is a stretch of the imagination, but it’s something that I think should be taken into consideration.

Yes, that was my concern initially but results of using this free app for the past year on the openly restored Facebook group have shown the results to be good/excellent and fairly accurate according to family members, if used carefully.

It is important to repair the image first in Affinity and be prepared to help the AI by increasing the resolution of the image and for multi person images slicing the image first then recomposing afterwards. Sometimes I’ve found better results by colouring first before using the AI. Skipping these steps can give horrific results.

[firstdefence]

The only thing I would say is, these AI image restorations on children tend to look like mini adults,  to use the seasons analogy - more like people near the end of summer than early to mid spring.