Canon slide copying set-up

[geogphotos]

I haven't set up my scanner so haven't done a direct comparison between a scan and a camera shoot but my impression is that the camera shot is fine for straightforward images, less so if you need to open up shadows and add brightness. Just attempting to lighten dark areas even a little produces all sorts of nasty grain and noise. I suppose that this is as expected from the two different processes. The camera is recording just what it sees on the surface, whereas the scanner is recording all the data that it can find.

 

I'd like to know what the more technically knowledgeable think about my observation.

[Harry Harrison]

1 hour ago, geogphotos said:

The camera is recording just what it sees on the surface, whereas the scanner is recording all the data that it can find.

It's difficult to generalise so it would be worthwhile for your own peace of mind to directly compare the results from your scanner when you get a chance. Just as with a camera, scanners differ in their abilities to resolve shadow detail properly. Manufacturers liked to quote the DMAX together with the 'bit depth' but they often inflate the figure and there's no real way to hold them to account or directly compare them on figures alone.

 

However it's not that the two methods are different, they are both shining light through the slide and capturing the result on a sensor via a lens, and with a scanner there's a front-surfaced mirror also. Actually your phrase "the scanner is recording all the data it can find" is actually more relevant to a camera shooting RAW than it is for a scanner, and your camera sensor has the benefit of being 15 or so years younger than the one in the scanner. Posts further back in this thread have found the camera better than scanners at cleanly resolving shadow detail and that's what I've found myself. I use Canon also but in tests (DxOMark) Nikon consistently have a higher dynamic range than Canon so might be presumed to do better here, whether that can be seen in practice I don't know.

 

It's still just possible that there is something about your method that is degrading the image, flare from the light source, internal reflections etc. but only you can tell if that might be true.

 

 

Edited April 22, 2020 by Harry Harrison

[spacecadet]

1 hour ago, geogphotos said:

The camera is recording just what it sees on the surface, whereas the scanner is recording all the data that it can find.

You're photographing through the slide, by transmitted light from your light box or whatever, just as a scanner does, not the surface by reflected light. So everything's there, limited by the flare in the optics. But this works for you- the flare keeps the density range within the capability of the sensor, as long as it's not too bad.

If you can't open up the shadows without noise they are underexposed. So you need more without blowing out the highlights.

 

Edited April 22, 2020 by spacecadet

[geogphotos]

Thanks Harry and Mark for putting me straight.

[MDM]

1 hour ago, Harry Harrison said:

It's difficult to generalise so it would be worthwhile for your own peace of mind to directly compare the results from your scanner when you get a chance. Just as with a camera, scanners differ in their abilities to resolve shadow detail properly. Manufacturers liked to quote the DMAX together with the 'bit depth' but they often inflate the figure and there's no real way to hold them to account or directly compare them on figures alone.

 

However it's not that the two methods are different, they are both shining light through the slide and capturing the result on a sensor via a lens, and with a scanner there's a front-surfaced mirror also. Actually your phrase "the scanner is recording all the data it can find" is actually more relevant to a camera shooting RAW than it is for a scanner, and your camera sensor has the benefit of being 15 or so years younger than the one in the scanner. Posts further back in this thread have found the camera better than scanners at cleanly resolving shadow detail and that's what I've found myself. I use Canon also but in tests (DxOMark) Nikon consistently have a higher dynamic range than Canon so might be presumed to do better here, whether that can be seen in practice I don't know.

 

It's still just possible that there is something about your method that is degrading the image, flare from the light source, internal reflections etc. but only you can tell if that might be true.

 

 

 

There are far too many variables to make simple generalisations as you imply. I can only speak in relation to the kit I have used but comparing the Nikon LS4000 (very good prosumer film scanner) to a D810 with top quality lens (55 Micro Nikkor 2.8 or Tamron 90 2.8) with ES-1 then the camera wins hands down in every department. Firstly, shooting raw gives extraordinary dynamic range, not just for shadow detail recovery but also highlight detail, way beyond what one would expect from contrasty slides. Secondly, noise is difficult to judge as it is difficult to distinguish from film grain but again the camera setup allows for far better control of noise and grain through shooting raw. Thirdly, the camera setup is far better than the scanner in terms of sharpness. That may be different if using a high end scanner but the ability to shoot raw in camera is always going to a huge plus. 
 

The full frame Nikon cameras since the D800 came out in 2012 have sensors with incredible dynamic range and if the tests are to be believed then they will produce superior results to equivalent Canon cameras (all else being equal) for slide copying where high dynamic range is a vital parameter. Very importantly, it is essential to keep the camera and slide as perfectly aligned as possible, using the lowest native ISO available, an aperture of f8 to f11 for optimally sharpness and depth of field, as well as triggering using a remote or cable release with mirror up if using a DSLR. A diffuse light source with high CRI is advisable. 
 

 

[Harry Harrison]

23 minutes ago, MDM said:

I can only speak in relation to the kit I have used

Good to know your experience, I really am going to do some scanning soon, too busy digging for Victory in the garden right now.

[MDM]

1 hour ago, Harry Harrison said:

Good to know your experience, I really am going to do some scanning soon, too busy digging for Victory in the garden right now.

Good for you Harry 😀. I wish I had the energy to do some gardening as I bought loads of veggie seeds before the advent of Covid-19. I am over the worst by far but have  been experiencing fairly serious post-viral fatigue for several weeks now. I had loads of plans to do all sorts but mostly on the back burner for now. Slowly improving but it has been a long haul with numerous dips. I do recommend continuing to avoid Covid-19 as long as possible to anyone who has not encountered it yet. At least the weather is wonderful today. I might go and hang out among the dandelions in my wild garden and continue to appreciate being alive 😀. 

Edited April 22, 2020 by MDM

[Harry Harrison]

47 minutes ago, MDM said:

I am over the worst by far but have  been experiencing fairly serious post-viral fatigue for several weeks now

Very sorry to hear that, I had wondered. I'm certainly doing my best to avoid it but as the clamour to get back to normal gets louder who knows what will happen.

[MDM]

24 minutes ago, Harry Harrison said:

Very sorry to hear that, I had wondered. I'm certainly doing my best to avoid it but as the clamour to get back to normal gets louder who knows what will happen.

I am glad to have come out the other side but it is taking a long time to recover. However, I am optimistic really. Things will have to start opening up again at some point and those who have been isolating and avoiding will have to decide whether to continue to do so for another year or more before a vaccine becomes available. There is also the question about how long immunity lasts. We shall see.
 

I do hope to do some copying of slides before too long anyway. 

[M.Chapman]

23 hours ago, geogphotos said:

Just attempting to lighten dark areas even a little produces all sorts of nasty grain and noise. I suppose that this is as expected from the two different processes.

 

Five tips to avoid/reduce noise problems in the shadows, in decreasing order of importance.

 

1) Shoot at low ISO

2) Expose to the right (using the histogram or clipping indicators to almost, but not quite, clip the highlights)

3) Shoot in RAW to allow maximum adjustment range in Post

4) Use a camera with a sensor with a decent dynamic range

5) Use a decent bright light source (I can copy most slides at around 1/250th F/8 ISO 200)

 

Even following all the above tips (apart from 4 - I have a M4/3 Lumix G7), I've still occasionally had problems when trying to open up deep shadows in Velvia slides taken in very contrasty conditions. Topaz Denoise will remove noise in the shadows, but also removes all detail. So my solution has been as follows.

 

1) Take 2 shots of the slide, the first exposed to the right, the second exposed by an extra +3 stops (which totally blows the highlights)

2) Open the 1st shot (in LR or or PS ACR) and adjust the exposure, open up shadows etc. so it looks good (but will have noisy shadows) transfer image to PS

3) Open the 2nd shot (in LR or PS ACR) and apply the same settings as the first one (develop with previous),  but then reduce the exposure by 3 stops, (it will have lost all highlight info, but will have shadows that have been opened up and free of noise), transfer image to PS

4) In PS combine both images into one, as aligned layers, with the 1st shot on top.

5) Create a threshold luminance mask with feathered edge and erase the (noisy) shadows from the first exposure to reveal the same shadows (but noise free) in the second shot.

 

It's actually easier to do than it sounds. A simpler way might be to use merge shot 1 and shot 2 to HDR in LR or PS, but I've never found a way of getting an end result I like from Adobe's HDR merge. 

 

Mark

 

Edited April 23, 2020 by M.Chapman

[M.Chapman]

With respect to my tips above, I did some checking on sensor dynamic range and think that maybe this needs promoting in my list. I went to DPReview here and set up a comparison of the noise when opening the shadows by +3 stops. The results look like this.

 

 

It should be noted that DPReview says the results are only directly comparable between cameras with sensors of the same size. It should also be noted that the ISO settings aren't the same. The test uses the lowest standard ISO setting available on each camera.

 

Nevertheless, I must confess to being somewhat disappointed by the differences. My Lumix G7 has a smaller sensor (M4/3) and also a base ISO that is higher, which puts it at a disadvantage, which I expected. But I'm also using E-Shutter mode to avoid vibration which, for some reason, appears to have more sensor noise than standard (mechanical shutter) mode. It looks like, if I swapped to using the mechanical shutter I might get better results, unless that causes vibration in my tripod mounted setup. But using the ES-1 adapter would solve that.

 

Perhaps using a camera with a sensor with decent dynamic range should be at number 2 or 3 on my list?? Especially if digitising Velvia...

 

Mark

Edited April 23, 2020 by M.Chapman

[MDM]

21 minutes ago, M.Chapman said:

With respect to my tips above, I did some checking on sensor dynamic range and think that maybe this needs promoting in my list. I went to DPReview here and set up a comparison of the noise when opening the shadows by +3 stops. The results look like this.

 

 

It should be noted that DPReview says the results are only directly comparable between cameras with sensors of the same size. It should also be noted that the ISO settings aren't the same. The test uses the lowest standard ISO setting available on each camera.

 

Nevertheless, I must confess to being somewhat disappointed by the differences. My Lumix G7 has a smaller sensor (M4/3) and also a base ISO that is higher, which puts it at a disadvantage, which I expected. But I'm also using E-Shutter mode to avoid vibration which, for some reason, appears to have more sensor noise than standard (mechanical shutter) mode. It looks like if I swapped to using the mechanical shutter I might get better results, unless that causes vibration in my tripod mounted setup. But using the ES-1 adapter would solve that.

 

Perhaps using a camera with a sensor with decent dynamic range should be at number 2 or 3 on my list?? Especially if digitising Velvia...

 

Mark

 

That is interesting Mark. I don't understand what is going on with the mechanical versus electronic shutters but the differences are obvious. 

 

If you are going to be doing a lot of slide copying and you want to get the best setup then you can't really beat full frame Nikon. As I said above, it is not just shadow noise that is handled very well by the D810, it is also highlight detail. I have been astonished by what can be restored from very contrasty Velvia slides. But a quality (not necessarily expensive) lens is also needed. 

 

Given the fact that the ES-1 and ES-2 work on the recommended 60mm Nikkor macro lenses without anything else makes for a simple life for anyone with this gear already. The superb 55mm MicroNikkor can be found at very reasonable prices secondhand (needs a standard PK-13 extension ring) and the Tamron 90 also does a superb job with the advantage that the autofocus works very well even at very close range. The Tamron needs spacer rings in front of the lens but is a 1:1 lens so no extension rings required behind the lens. 

 

 

[Harry Harrison]

These DPReview comparisons are very useful, Wim shared another for resolution a couple of weeks ago, and there is a great deal of meticulous work that has to go on in the background to make them possible.  The Nikon results confirm the DxOMark lab tests, Sony seem very good also (do Nikon use Sony sensors?). I think from Ian's point of view though this might not fully explain the results he's seeing, the differences are much less obvious if you move the selection area over, say, the bank note or the lithograph/engraving, which might have more in common with a scan, and 3 stops correction is pretty extreme.

 

I would suggest to Ian that as a test he should deliberately over-expose his slide by a couple or even three stops to bring the shadows to a place where they are comfortably within the dynamic range and see if there is actually anything useful there, it may well be grain.

 

I was intrigued why Canon seem to do far less well than Nikon in terms of the dynamic range tests on DxOMark so I asked a Canon rep at the Photography Show, thinking there was some innocent explanation, testing bias perhaps. It didn't go down well, he was almost aggressive actually, well it was the last day.

Edited April 23, 2020 by Harry Harrison

[MDM]

34 minutes ago, Harry Harrison said:

These DPReview comparisons are very useful, Wim shared another for resolution a couple of weeks ago, and there is a great deal of meticulous work that has to go on in the background to make them possible.  The Nikon results confirm the DxOMark lab tests, Sony seem very good also (do Nikon use Sony sensors?). I think from Ian's point of view though this might not fully explain the results he's seeing, the differences are much less obvious if you move the selection area over, say, the bank note or the lithograph/engraving, which might have more in common with a scan, and 3 stops correction is pretty extreme.

 

I would suggest to Ian that as a test he should deliberately over-expose his slide by a couple or even three stops to bring the shadows to a place where they are comfortably within the dynamic range and see if there is actually anything useful there, it may well be grain.

 

I was intrigued why Canon seem to do far less well than Nikon in terms of the dynamic range tests on DxOMark so I asked a Canon rep at the Photography Show, thinking there was some innocent explanation, testing bias perhaps. It didn't go down well, he was almost aggressive actually, well it was the last day.

 

Sony were making the sensors for the Nikon D800s and D810s. I think Nikon started making their own when they introduced the new backlit sensors for the D850 and Z series but not sure if that is the case. The reason I didn't mention Sony above is that the ES-1 and ES-2 are made by Nikon and there is a big choice of Nikon lenses at reasonable prices including many secondhand as we discussed early on in this thread. There is no reason why the ES-1 wouldn't work with Sony cameras and the right lenses and adapters. 

 

 

[M.Chapman]

8 hours ago, MDM said:

If you are going to be doing a lot of slide copying and you want to get the best setup then you can't really beat full frame Nikon.

 

Indeed. Unfortunately I'm not sure that I could ever recover the cost through licence sales and hence can't justify the purchase (no matter how much I'd like to). So, for me, I think it must be a case of getting the best from the kit I have. 

 

On the plus side, I've sold 5 licences from slides I digitised this year already, although none on Alamy yet (but it's too soon for that).

 

Mark

Edited April 23, 2020 by M.Chapman

[M.Chapman]

49 minutes ago, Harry Harrison said:

and 3 stops correction is pretty extreme.

 

It is when applying to the whole histogram/image, but it's not when lifting just the shadows in PS.

 

Mark

Edited April 23, 2020 by M.Chapman

[Harry Harrison]

10 minutes ago, MDM said:

There is no reason why the ES-1 wouldn't work with Sony cameras and the right lenses and adapters. 

The 'what you see is what you get' advantage of electronic viewfinders, particularly high quality ones, on mirrorless cameras like Sony & Fuji makes using the 55mm Micro-Nikkor with an adapter a very viable option. Essentially 'Live View' but through the viewfinder.

[M.Chapman]

4 minutes ago, Harry Harrison said:

The 'what you see is what you get' advantage of electronic viewfinders, particularly high quality ones, on mirrorless cameras like Sony & Fuji makes using the 55mm Micro-Nikkor with an adapter a very viable option. Essentially 'Live View' but through the viewfinder.

 

I note that second-hand Nikon 810s are quite pricey. Is there a Sony camera that uses the same sensor that might offer the same dynamic range/low noise that might be cheaper?

 

Mark

[Harry Harrison]

34 minutes ago, M.Chapman said:

I note that second-hand Nikon 810s are quite pricey. Is there a Sony camera that uses the same sensor that might offer the same dynamic range/low noise that might be cheaper?

I think MDM's your man for that, I'd have to some research, I believe Sony have a whole range of MP that separates their models and after a quick look I can see that the A7R v1 is also 36MP. like the D810, not sure what the prices are like.

 

Edit: Actually quite pricy, wouldn't a D800 be OK, they seem cheaper?

Edited April 23, 2020 by Harry Harrison

[MDM]

I don't know anything about Sony cameras apart from the RX100 so not the man for that but I doubt you would get one secondhand cheaper than the equivalent Nikon. You could get a guaranteed secondhand Nikon D800 plus 55 Micro Nikkor for less than £1000 from Grays of Westminster which is the best place to buy more or less anything Nikon in the country. You could get these on ebay as well but you probably won't get a guarantee. The PK-13 and ES-1 would be around £175 new I guess. It would be a lot to invest just for slide copying but then a decent Nikon slide scanner was well over £1000 when they were available. That camera lens combo could be used for some serious close-up photography and the 55mm lens has multiple uses. 

[M.Chapman]

I did some testing to compare the effect of using the mechanical or the electronic shutter on my Lumix G7 on shadow noise and blur in digitised slides. My findings agree with those seen by DPReview, i.e. images taken with the electronic shutter show significantly more noise when the shadows are opened up than those taken with the mechanical shutter. Unfortunately in my current setup, the mechanical shutter causes slight vibration so there's some blurring. In order of quality of the end result (worst first) I see the following.

 

1) Electronic shutter + normal exposure (noisy but sharp)

2) Mechanical shutter + normal exposure (low noise, slightly less sharp)

3) Electronic shutter + blend of short and long exposures (minimal noise and sharp)

4) Mechanical shutter + blend of short and long exposures (lowest noise but less sharp due to vibration effect on longer exposure)

 

Here are some 200% crops

 

 

 

I need to improve the rigidity of my setup so I can use the mechanical shutter, or I'll have to stick with blending two E-Shutter shots when copying slides where I need to lift the shadows a lot. Improving the rigidity of my tripod mounted camera should be straightforward, but I'm conscious that my macro lens also has a floating element (for the IS which is OFF) which might be affected too.

 

Mark

Edited April 24, 2020 by M.Chapman

[Harry Harrison]

7 minutes ago, M.Chapman said:

images taken with the electronic shutter show significantly more noise when the shadows are opened up than those taken with the mechanical shutter.

Quite hard to understand why without knowing more about electronics, this article touches on it with respect to the Nikon Z7 but isn't able to demonstrate it conclusively.

[M.Chapman]

22 hours ago, Harry Harrison said:

Quite hard to understand why without knowing more about electronics, this article touches on it with respect to the Nikon Z7 but isn't able to demonstrate it conclusively.

 

Yes it's a real puzzle. I wonder if it might be to do with setting black levels for each pixel. If I use the mechanical shutter then the sensor is in the dark just before the exposure and all pixels can be "zeroed out". But this doesn't happen with the E-shutter.

 

After some further testing I've shown that the pixel noise reduction from using the mechanical shutter instead of the E-shutter is equivalent to the noise reduction I get if I increase an E-shutter exposure by 1 stop and then apply -1 stop correction in post. So it's not huge. When I really need to lift the shadows I find I need an extra 3 stops, which necessitates "HDR" merging of two exposures anyway which works fine with the E-Shutter. I've also been unable to totally eliminate the vibration caused by the mechanical shutter which causes very slight softening. So I'm sticking with my original method, i.e. use E-shutter (and multiple exposure merge only when needed).

 

Sure would be interesting to see what the Nikon 810 makes of one of my difficult slides using the ES-1....

 

Mark

Edited April 25, 2020 by M.Chapman

[Harry Harrison]

40 minutes ago, M.Chapman said:

Sure would be interesting to see what the Nikon 810 makes of one of my difficult slides using the ES-1....

Yes, that's a tantalising prospect, the same setup, the same lens, the same slide(s), different cameras. I don't know anyone with a Nikon 800/810 otherwise I'd be keen to try it myself. If you were to look at the resolution comparisons on DPReview that Wim linked to I can't see any difference between the D800 & D810, there's also the D800E without the anti-aliasing filter. The link that Wim gave to the resolution comparisons for their 'Studio scene' is instructive but it's difficult to get an idea about how much difference between sensors/MP one would see in a film scan.

 

https://www.dpreview.com/reviews/fujifilm-x-pro2/7

 

There's a slight problem with it in that it will depend to an extent upon the lens used for the tests which isn't consistent even within brands.

Edited April 25, 2020 by Harry Harrison

[M.Chapman]

56 minutes ago, Harry Harrison said:

Yes, that's a tantalising prospect, the same setup, the same lens, the same slide(s), different cameras. I don't know anyone with a Nikon 800/810 otherwise I'd be keen to try it myself.

 

Yes it would be interesting, maybe we could post the same slide(s) between us and then upload our RAWs for comparison? Maybe Micael (MDM) would be interested in participating too? I'd happily post a slide or two in return for a sight of some RAW files from other setups.

 

Mark