Canon slide copying set-up

[wiskerke]

15 hours ago, Chuck Nacke said:

 

BTDT, but luckily the photographer made great 11 X 14s.  I have also copied several of my B & W 11 X 14s using my D800 and was happy with the results, but for 35mm chromes I still prefer to scan.  Your Layer idea is good, I do something like that as well.  The LS 5000 is a really good machine and they have gotten silly expensive.  The Canon 4000 has a tad lower DIN but it does produce a really sharp scan. I have also heard that the 9000 is a wonderful scanner, but again $$$.

 

A decade ago when I had a Windows XP machine with a PCMCIA slot I only used the CanoScan with a 16bit ADAPTEC PCMCIA card and I liked it a lot better.  I am still trying to fully understand VueScan?

 

Is Image ID: AM5KH5  from a film scan?

 

Chuck

 

Better: most are 16x20 !  😁

The negatives are quite a pile, but the volume is nothing compared to the prints.

 

AM5KH5 is not a scan. Interesting question though, so after some searching, this was the last scan I have uploaded here:

It was a TIFF. A tiff? That's unexpected. I had totally forgotten about having to send tiffs. So I went looking for my first JPEG.

This seems to be it:

 

wim

 

[Chuck Nacke]

6 hours ago, wiskerke said:

 

Better: most are 16x20 !  😁

The negatives are quite a pile, but the volume is nothing compared to the prints.

 

AM5KH5 is not a scan. Interesting question though, so after some searching, this was the last scan I have uploaded here:

It was a TIFF. A tiff? That's unexpected. I had totally forgotten about having to send tiffs. So I went looking for my first JPEG.

This seems to be it:

 

wim

 

Wim, 

 

I copied Image ID: AM5KH5 directly off of your images and it was a wide shot of the Smith Tower in Seattle with a jet flying by?

Do not understand what is going on?

 

Chuck

[spacecadet]

4 hours ago, Chuck Nacke said:

Wim, 

 

I copied Image ID: AM5KH5 directly off of your images and it was a wide shot of the Smith Tower in Seattle with a jet flying by?

Do not understand what is going on?

 

Chuck

The light fitting is actually  A4NRC5 .

[wiskerke]

5 hours ago, Chuck Nacke said:

Wim, 

 

I copied Image ID: AM5KH5 directly off of your images and it was a wide shot of the Smith Tower in Seattle with a jet flying by?

Do not understand what is going on?

 

Chuck

 

The light fixture (A4NRC5) was the last scan that I uploaded to Alamy.

 

If you look at my portfolio, everything after that is a digital photo. Including AM5KH5:

which is a digital from a 1DS2.

But everything that comes before the light fixture is a scan.

 

I soon gave up on my back catalog. Not because the images weren't good, but because I found scanning a pain. Like you I managed about 1 or 2 images a day at a quality that would pass my own QC. Direct digital quality was far better from the 16 megapixel full frame dslrs onwards.

Before with the Nikon scanner I had just scanned whole rolls directly from the lab, which was far easier and faster. With the many flatbed and film scanners that came before the Nikon I had just done less demanding jobs (some) or jobs that would pay for the extra time (mostly). Printing large format prints from slides was done from drum scans, which I eventually learned to do also (but just barely).

To put it into perspective a little more: the Nikon scans were used in print, but for annual reports and such they were re-done on a drum scanner even for thumbnail sized images in print. However 4 stories high banners with portraits on the outside of buildings were printed straight from my 5 megapixel Olympus E-20 with it's tiny thumbnail sized sensor.

 

wim

[MDM]

 

14 hours ago, Chuck Nacke said:

Michael,

 

Arguing about machines is pointless, there are many scanners that will do a good job for editorial submissions, all it takes is learning how to use them. 

 

Chuck

 

 

It's not about mine is better than yours or whatever. However, some methods are superior to others in terms of image quality and speed - they are not all the same.  I am simply posting my opinions and findings based on direct experience - stuff I have done myself - and the aim is to help answer the original question that started the thread, as well as maybe to help others decide. From my own findings of almost 20 years ago, I would not recommend the Canon 4000 over the Nikon 4000 scanners. 

 

And things have come a long way since these Canon and Nikon scanners were introduced almost 20 years ago. Having used Canon and Nikon 4000 slide scanners as well as the newer methods discussed at length in this thread, I  am fully convinced that modern methods of digitisation in camera are now significantly superior to prosumer slide scanners in terms of image quality and speed of the basic copying process. That does not mean that the scanner methods are bad,  just that they have been superseded by advances in terms of camera sensor technology (in my opinion of course). 

 

So my recommendation, based entirely on experience, to anyone considering investing in a method of digitising 35 mm slides would be to go for in camera copying although exactly which approach to take would depend on one's existing equipment.

 

 

Edited February 19, 2020 by MDM

[Harry Harrison]

My experience is the same as MDM's, but it's fair to say that most photographers these days don't own an old but good quality dedicated film scanner from the likes of Canon, Nikon or Minolta. These may be 20 years old and with no service or spare parts available so buying any of these secondhand is risky, especially if you have to part with a lot of cash. Then the choice is between the DSLR scanning method described and a new and relatively expensive consumer film scanner from Plustek or Pacific, or a flatbed. That's a much easier decision to make I think.

Edited February 19, 2020 by Harry Harrison

[Chuck Nacke]

3 hours ago, wiskerke said:

 

The light fixture (A4NRC5) was the last scan that I uploaded to Alamy.

 

If you look at my portfolio, everything after that is a digital photo. Including AM5KH5:

which is a digital from a 1DS2.

But everything that comes before the light fixture is a scan.

 

I soon gave up on my back catalog. Not because the images weren't good, but because I found scanning a pain. Like you I managed about 1 or 2 images a day at a quality that would pass my own QC. Direct digital quality was far better from the 16 megapixel full frame dslrs onwards.

Before with the Nikon scanner I had just scanned whole rolls directly from the lab, which was far easier and faster. With the many flatbed and film scanners that came before the Nikon I had just done less demanding jobs (some) or jobs that would pay for the extra time (mostly). Printing large format prints from slides was done from drum scans, which I eventually learned to do also (but just barely).

To put it into perspective a little more: the Nikon scans were used in print, but for annual reports and such they were re-done on a drum scanner even for thumbnail sized images in print. However 4 stories high banners with portraits on the outside of buildings were printed straight from my 5 megapixel Olympus E-20 with it's tiny thumbnail sized sensor.

 

wim

Wim,

 

Looks like we share a lot of past.  I use to work with a corporate art director that could color check a 4 or 6 color annual report, on press, two colors at a time....

A4NRC5 looked more like a film scan to me?  Check out Image ID: CYT07C that was shot with a Kodak / Nikon DCS-620 at 1600 ISO originally at $30,000+

DSLR (Ah the "Bad Old Days.)

 

Best,

 

Chuck

2 hours ago, MDM said:

 

 

 

It's not about mine is better than yours or whatever. However, some methods are superior to others in terms of image quality and speed - they are not all the same.  I am simply posting my opinions and findings based on direct experience - stuff I have done myself - and the aim is to help answer the original question that started the thread, as well as maybe to help others decide. From my own findings of almost 20 years ago, I would not recommend the Canon 4000 over the Nikon 4000 scanners. 

 

And things have come a long way since these Canon and Nikon scanners were introduced almost 20 years ago. Having used Canon and Nikon 4000 slide scanners as well as the newer methods discussed at length in this thread, I  am fully convinced that modern methods of digitisation in camera are now significantly superior to prosumer slide scanners in terms of image quality and speed of the basic copying process. That does not mean that the scanner methods are bad,  just that they have been superseded by advances in terms of camera sensor technology (in my opinion of course). 

 

So my recommendation, based entirely on experience, to anyone considering investing in a method of digitising 35 mm slides would be to go for in camera copying although exactly which approach to take would depend on one's existing equipment.

 

 

Michael,

 

The best film scanner is the one you have and it works.  I do disagree about "digitising (sp)"  At a certain level, I still prefer film scans to photographing chromes.  I is important here,

"I PREFER" does not mean that one method is better than another.  I never intended to suggest that anyone invest in old outdated equipment.  I do wish that I still had the old

Kodak / Nikon DSC 460, modified by NASA.  The CCD sensor in the 460 was marvelous and the sensor alone cost Kodak over $5,000 per sensor to make.  While I do like modern

36mp + DSLRs, I also wish I still had my black Nikon SP as well but...   Now I have to get back to film scanning. 

 

I would say to everyone, "The most important accessory for digital photography is a very comfortable chair."

 

Chuck

[wiskerke]

2 hours ago, Chuck Nacke said:

I would say to everyone, "The most important accessory for digital photography is a very comfortable chair."

 

Chuck

 

Amen to that! Photography has become a sedentary occupation. May I recommend Aeron chairs? Lots cheaper in the US than here I'm afraid. Even second hand.

 

wim

[Betty LaRue]

47 minutes ago, wiskerke said:

 

Amen to that! Photography has become a sedentary occupation. May I recommend Aeron chairs? Lots cheaper in the US than here I'm afraid. Even second hand.

 

wim

Wow! Looks like great chairs. Sale going on. Not in my budget, yet with my spine problems, I need one!

[wiskerke]

6 hours ago, Betty LaRue said:

Wow! Looks like great chairs. Sale going on. Not in my budget, yet with my spine problems, I need one!

Mine are all pre-owned; used; second hand. Not even refurbished. Well most of them now are. And I do have a box full of spare parts 🤣. Now. 

Sorry OT again..

 

wim

 

[geogphotos]

Maybe a very uncomfortable chair is the best accessory? 😏

[MDM]

 

 

18 hours ago, Chuck Nacke said:

The best film scanner is the one you have and it works.  

 

 

 

This is not worth arguing with. It is patent nonsense applied to scanners. A soundbite worthy of the best newsfakers. The idea here is to discuss the best methods of digitising film not to dismiss all as equal or to say it doesn't matter. Of course it matters. That is what the fundamental theme of the thread has been all along.

 

19 hours ago, Chuck Nacke said:

I do disagree about "digitising (sp)"  

 

Presumably this refers to using a digital camera rather than a scanner. This is the stuck record bit - no evidence has ever been provided for this opinion after 18 pages of this thread.

 

 

18 hours ago, Chuck Nacke said:

 

I would say to everyone, "The most important accessory for digital photography is a very comfortable chair."

 

 

 

Or a chair with a special filter to prevent people from talking out of the wrong orifice 😎.  Maybe one that shakes just before one hits the submit button. I can feel mine shaking now but I can't help it - here goes 😁.

 

 

 

[MDM]

 

Now coming back to the original subject of the thread, here is  a summary of my findings to date in relation to the Nikon ES-1 slide copier.

 

Having tested the Tamron 90 and Micro-Nikkor 55 on a Nikon D810, both lenses give pretty equivalent results in terms of sharpness, including edge-edge and corner-corner sharpness on the ES-1. The AF on the Tamron works very well and makes the whole process easier than manually focusing with the Micro-Nikkor 55 but very careful focusing seems to give results that are just as good as the Tamron. 

 

The Micro-Nikkor 55 needs a PK-13 (27mm or thereabouts) normal extension tube behind the lens as it only goes to half life size. The Tamron, which goes to full life size, needs about 80-90 mm of extenders in front of the lens in order to get it to focus close enough. These are inexpensive and can be bought on eBay from China. A 58-52mm step down ring is also necessary.

 

Getting this to work on other makes of camera and other macro lenses should be no problem as far as I can see. Experimentation would be required to determine the amount of extension required in front of the lens for other lenses to enable the lens to focus and to get near to lifesize on the image. 

 

Finally edge-edge and corner-corner sharpness are very important considerations in choosing a lens for camera copying of slides. 

 

Edited February 20, 2020 by MDM

[M.Chapman]

8 hours ago, MDM said:

Presumably this refers to using a digital camera rather than a scanner. This is the stuck record bit - no evidence has ever been provided for this opinion after 18 pages of this thread.

 

Bemusing isn't it...  You and I have both posted 100% images whilst contributing to this thread to demonstrate the best we have achieved so far with the DSLR copying technique.

 

A respected contributor (Chuck) with extensive experience states that their preference/personal opinion is that scanning is better (which is fine). But he also states "I do think that a large part of the problem is that people think they can do many finished scans in a day?  On my best day I can finish three images in one day, but usually it is one, but I have finished 16bit TIFFs in aRGB color of every image that I invest my time scanning."

 

It leaves me intrigued. It takes me between 10-30 minutes digitise and fully post post process a slide using the DSLR + RAW file technique. Am I missing something? I want to be enlightened...Is the level of post-processing used by Chuck required to enhance the original photo (e.g. remove distracting background from a picture of Steve Jobs) then fair enough (I don't include that in my 10-30 minute estimate). But if that level of post-processing time is required to fix problems arising from the scanning process, then what are those problems? and does a DSLR + RAW file copy have the same problems? Maybe it's time needed to combine manually multiple scans taken at different exposures to get the best shadow detail?

 

I'll post an update on my further experience with Topaz Denoise shortly as I've now processed around 50 different slides (all landscape pictures) all using Topaz.

 

Mark

 

Edited February 20, 2020 by M.Chapman

[David Olsen]

Thank you for the info MDM, I have the Tamron 90mm.  I just need the motivation now, maybe when I retire later this year.

[Chuck Nacke]

4 hours ago, M.Chapman said:

 

Bemusing isn't it...  You and I have both posted 100% images whilst contributing to this thread to demonstrate the best we have achieved so far with the DSLR copying technique.

 

A respected contributor (Chuck) with extensive experience states that their preference/personal opinion is that scanning is better (which is fine). But he also states "I do think that a large part of the problem is that people think they can do many finished scans in a day?  On my best day I can finish three images in one day, but usually it is one, but I have finished 16bit TIFFs in aRGB color of every image that I invest my time scanning."

 

It leaves me intrigued. It takes me between 10-30 minutes digitise and fully post post process a slide using the DSLR + RAW file technique. Am I missing something? I want to be enlightened...Is the level of post-processing used by Chuck required to enhance the original photo (e.g. remove distracting background from a picture of Steve Jobs) then fair enough (I don't include that in my 10-30 minute estimate). But if that level of post-processing time is required to fix problems arising from the scanning process, then what are those problems? and does a DSLR + RAW file copy have the same problems? Maybe it's time needed to combine manually multiple scans taken at different exposures to get the best shadow detail?

 

I'll post an update on my further experience with Topaz Denoise shortly as I've now processed around 50 different slides (all landscape pictures) all using Topaz.

 

Mark

 

Mark,

 

It is what it is.  Just for the record, as a photojournalist, I do not make large or context changes to finished images.  I come from the school of Gene Smith,

on non breaking news pictures I prefer to work on my images until I am happy with them or exhausted.  As per posting original files, very few reading this

thread would understand my reasons for not doing that.

 

To Michael,

 

You claim to have a higher degree, but you can not even spell "DIGITIZING"  you wrote "method of digitising 35 mm slides"  

 

Chuck

Edited February 20, 2020 by Chuck Nacke
deletion

[M.Chapman]

21 minutes ago, Chuck Nacke said:

 

You claim to have a higher degree, but you can not even spell "DIGITIZING"  you wrote "method of digitising 35 mm slides

Digitising is the correct UK spelling. Digitizing is the US variant.

 

Mark

[MDM]

9 minutes ago, Chuck Nacke said:

Mark,

 

It is what it is.  Just for the record, as a photojournalist, I do not make large or context changes to finished images.  I come from the school of Gene Smith,

on non breaking news pictures I prefer to work on my images until I am happy with them or exhausted.  As per posting original files, very few reading this

thread would understand my reasons for not doing that.

 

To Michael,

 

You claim to have a higher degree, but you can not even spell "DIGITIZING"  you wrote "method of digitising 35 mm slides"  

 

Chuck

 

 

It's not just a claim Chuck. I do have an actual PhD in geology (specialised in volcanology and geochemistry) from the University of Bristol (England, UK). My first degree was a 1st in geology with minors in chemistry and physics from University College Dublin (Ireland). Happy to show you my degree certificates if you want. Not sure what the relevance is to the current topic though. I did mention that my scientific training makes me want to see evidence to back up assertions so I guess that is where this came from. 

 

As for the spelling, digitise is the correct way to spell it on this side of the Atlantic. They call it British English although it is also used in Ireland where I come from. This applies to a lot of words that in US English end in ize - here they are spelt ise. Other differences include words such as color - colour, labor - labour. I actually prefer the American myself as it tends to more phonetic but I live here so I conform. 

 

 

 

 

[MDM]

4 hours ago, David Olsen said:

Thank you for the info MDM, I have the Tamron 90mm.  I just need the motivation now, maybe when I retire later this year.

 

Don't mention it David. I have had the ES-1 copier for a few years but hardly used it much until this thread got me going to see what would actually work. The Tamron 90 is now going to be my lens of choice for copying, as the AF works well and makes life a lot easier than trying to manually focus. There have been several versions of this lens. I still have a really old manual focus Adaptall one which is optically very good as well. I am keeping it in case I ever get my one surviving Nikon film camera out again as the newer lenses without aperture ring won't work.

 

Great shots of Hawaii by the way. Looks like a fantastic place. 

Edited February 20, 2020 by MDM

[MDM]

5 hours ago, M.Chapman said:

 

 

 

It leaves me intrigued. It takes me between 10-30 minutes digitise and fully post post process a slide using the DSLR + RAW file technique. Am I missing something? I

Maybe it's time needed to combine manually multiple scans taken at different exposures to get the best shadow detail?

 

 

I am taking around 30 mins (max 1 hour) I guess to process in total. I have changed my method a bit from the discussions we have had and am spending a bit more time on each image but, even with spotting and careful localised noise reduction, I can't imagine this taking more than an hour. I don't see any need myself in general for multiple scans as the dynamic range of the D810 is astoundingly good but it might be a good idea for some really contrasty images. However, It all comes down to what detail there is in the slide itself. If there is no detail in the highlights or shadows of the slide, then there may not be a lot there to recover. Negative film may have a much higher dynamic range than slides depending on the scene, the film and how it was exposed and  processed so taking several images might work better.. 

 

Edited February 20, 2020 by MDM

[spacecadet]

12 hours ago, MDM said:

Negative film may have a much higher dynamic range than slides depending on the scene, the film and how it was exposed and  processed so taking several images might work better.. 

Actually neg has a lower density range. It's not going to be viewed directly so deep blacks and clear highlights aren't needed.

Have a look at some characteristic curves- straight line portion of Portra is under 2logD as against north of 2.5logD for something like E100D. Maybe 2 stopsor a bit more  in exposure terms.

Edited February 21, 2020 by spacecadet

[M.Chapman]

Topaz De-noise AI update

 

As posted previously I've been trying Topaz De-noise AI (TDN) to remove film grain. I've also tried using every combination of sliders and selections in PS/LR to achieve the same.

 

So far I'm unable to find any combination of sliders in LR/PS that does as good a job of removing grain as TDN does, whilst also retaining detail. But, TDN is not without its problems. It's not fast (about 1 - 2 minutes to process a 13MP 16 bit image on my 2012 MacBook Pro). But more significantly, if there are any high contrast edges in the image, before applying TDN, then artefacts can be introduced. There are 2 main forms of artefact I've seen.

 

1) Halo (sometime double) around high contrast edges in the image (they appear over-sharpened)

2) Strange "whisps" that may appear at the edge of the image in low contrast areas (especially skies) and extend maybe over 30 pixels.

 

I've changed my workflow to tackle these problems. Previously I was doing most of my processing, then downsizing, and then applying TDN as an almost final operation. This made TDN faster (smaller image), but the image was already sharpened (by the downsize and my initial RAW conversion settings). So my new workflow is:-

 

1) Open image in ACR or LR with NO sharpening at all (i.e. 0 setting sharpening amount sliders) and some luminance noise reduction (25, 50, 0)

2) Level horizon (if needed) and crop rebate from image (so histogram becomes meaningful)

3) Apply any levels, tone and WB adjustments

4) Apply Topaz De-noise filter

    - Remove noise 15 (default)

    - Sharpen 15 (default)

    - Original detail 5 (to retain a tiny amount of grain texture)

5) Downsize using Bicubic (sharper) or Bilinear (less sharp) to 6MP (just above Alamy minimum size)

6) Inspect at 100%

    - heal or clone out dust

    - inspect edge of frame for "whisps" and use large healing brush to remove by brushing along edge of image

7) Save as jpg quality level 12

 

This image was produced this way. Feel free to download and inspect at 100%.

Here's a 200% crop showing the "whisp" effect before I removed it from the bottom right corner of that image.

 

 

Mark

Edited February 22, 2020 by M.Chapman
Updated time to process an image to 1 - 2 mins

[MDM]

6 hours ago, M.Chapman said:

Topaz De-noise AI update

 

As posted previously I've been trying Topaz De-noise AI (TDN) to remove film grain. I've also tried using every combination of sliders and selections in PS/LR to achieve the same.

 

So far I'm unable to find any combination of sliders in LR/PS that does as good a job of removing grain as TDN does, whilst also retaining detail. But, TDN is not without its problems. It's not fast (about 1 minute to process a 13MP 16 bit image on my 2012 MacBook Pro). But more significantly, if there are any high contrast edges in the image, before applying TDN, then artefacts can be introduced. There are 2 main forms of artefact I've seen.

 

1) Halo (sometime double) around high contrast edges in the image (they appear over-sharpened)

2) Strange "whisps" that may appear at the edge of the image in low contrast areas (especially skies) and extend maybe over 30 pixels.

 

I've changed my workflow to tackle these problems. Previously I was doing most of my processing, then downsizing, and then applying TDN as an almost final operation. This made TDN faster (smaller image), but the image was already sharpened (by the downsize and my initial RAW conversion settings). So my new workflow is:-

 

1) Open image in ACR or LR with NO sharpening at all (i.e. 0 setting sharpening amount sliders) and some luminance noise reduction (25, 50, 0)

2) Level horizon (if needed) and crop rebate from image (so histogram becomes meaningful)

3) Apply any levels, tone and WB adjustments

4) Apply Topaz De-noise filter

    - Remove noise 15 (default)

    - Sharpen 15 (default)

    - Original detail 5 (to retain a tiny amount of grain texture)

5) Downsize using Bicubic (sharper) or Bilinear (less sharp) to 6MP (just above Alamy minimum size)

6) Inspect at 100%

    - heal or clone out dust

    - inspect edge of frame for "whisps" and use large healing brush to remove by brushing along edge of image

7) Save as jpg quality level 12

 

This image was produced this way. Feel free to download and inspect at 100%.

Here's a 200% crop showing the "whisp" effect before I removed it from the bottom right corner of that image.

 

 

Mark

 

 

Presumably you are opening the image into Photoshop (i.e. converting the raw) before step 4 and doing the TDN stuff from within Photoshop. If it was me, I would be inclined to remove the dust and do everything else (including removing the wisps as you call them) before downsizing and then save it (PSD for me) as you then have a larger master copy in case you want to do anything else. The artifacts (the wisps) are one thing that put me off using TDN as well as the slowness on my machine but it is good you are finding it useful. It certainly gives a very clean background. It is difficult to really judge the effects on that falcon image as it is not clear what should be in sharp focus. A landscape with sky would be interesting. 

 

 

 

[M.Chapman]

8 hours ago, MDM said:

Presumably you are opening the image into Photoshop (i.e. converting the raw) before step 4 and doing the TDN stuff from within Photoshop.

 

Yes I am, but if you have the TDN plugin in LR I assume TDN could be launched from within LR and the data transfers automatically?

 

8 hours ago, MDM said:

If it was me, I would be inclined to remove the dust and do everything else (including removing the wisps as you call them) before downsizing and then save it (PSD for me) as you then have a larger master copy in case you want to do anything else.

 

Good point. I downsized first to save time on the 100% inspection stage. Also, in my case, I'm not sure my originals warrant retouching before downsizing as the quality of the lenses I was using at the time (in general) perhaps don't justify saving at 13MP.

 

8 hours ago, MDM said:

It is difficult to really judge the effects on that falcon image as it is not clear what should be in sharp focus. A landscape with sky would be interesting. 

 

OK. Please find a two layer TIFF here.

 

The base layer is simply produced from the RAW file using ACR default levels of sharpening (40, 1, 25, 0) and some noise reduction (25, 50, 0, 25, 50, 50), followed by downsizing (Bicubic sharper) from ~13MP to ~6MP. So this layer shows the grain I am removing. (Although it's not the worst example of grain I have - I maybe should have picked a worse image).

 

The upper layer was produced from the RAW file using 0 sharpening and the same level of noise reduction as the base layer. TDN was then applied (15, 15, 5, 0.00). The result was then downsized from ~13MP to ~6MP (Bicubic sharper). I've left all the "edge whisps" in place so you can see them (e.g. just above mountain on left hand border, top edge just right of centre) although these are quite minor.

 

Both layers also have the same crop, level, WB and levels/tone adjustments applied.

 

Downloading and opening in PS and toggling the upper layer on an off shows the grain removal (partly from NOT applying any sharpening in ACR, the rest is from TDN) whilst detail is retained (IMHO). 

 

Footnote: It seems that NOT sharpening the image before TDN can have a massive effect. If grain is above a certain threshold, TDN may decide it's not noise, but is a "feature" to be retained and enhanced, rather than removed. Leading to patches of grain remaining in the image.

 

Mark

Edited February 22, 2020 by M.Chapman

[geogphotos]

Maybe we have exhausted the topic of making digital images from photographic slides?

 

So how about we start discussing the reverse? Making photographic slides from digital images.

 

Curiosity led me to try a Google search and yes it exists as a thing!

 

https://www.digitalslides.co.uk/

Edited February 22, 2020 by geogphotos