Epson v750 scans not the greatest, looking for suggestions

[Mark D Segal]

If the orange mask is on the rebate section as well as on the area of the film that is exposed, would trying to change the rebate section to what it would look like without the mask be a way to develop an algorithm to get rid of the mask across the entire exposure?  If not, why not?

Oh, so by "rebate" you mean the unexposed header or toe of the film strip or the separation strips between the frames. Never heard the term "rebate" applied to unexposed film before. Yes, it's helpful but it doesn't do the whole job, and for each different film (manufacturer/film type/ISO) you would need a different recipe. See my previous articles on this.

[Alan Klein]

Oh, so by "rebate" you mean the unexposed header or toe of the film strip or the separation strips between the frames. Never heard the term "rebate" applied to unexposed film before. Yes, it's helpful but it doesn't do the whole job, and for each different film (manufacturer/film type/ISO) you would need a different recipe. See my previous articles on this.

Isn't the unexposed section on all films clear except for the orange mask?  If so, then a program applies an algorithm that eliminates the orange mask to see no color in the unexposed section.  Maybe that's the way Epsonscan does it.  It doesn't care what film you used.  Maybe it uses levels and uses then unexposed area as the black point and adjusts the color to get the black.  Just some ideas.  Does anyone have a DSLR shot with the unexposed rebate section that they can post here?  A strait flat shot with no adjustments.  Then we can try levels and some other tests.

[digitaldog]

Epsonscan does not know which color film it's scanning only that it is a negative.  So it's removal of the orange mask has to be a common and simple algorithm that applies to all negative color film. 

That's basically true of all scanning software although they may provide some film stock names as a starting point, in no way does that guarantee success of joy in the rendering. Not all software handles the conversions or even the auto (on what part of the film) equally. The Imacon mentioned did a good job with FlexColor in the old days. And it's sharp corner to corner and mentioned. The lens alone probably cost many times what anyone paid for the Epson in full. With the Imacon software, the trick with color negs was to crop small(er) areas of the exposed image and click Auto which when it found a 'good sampling' did a very good job inverting and removing the orange mask which is affected by exposure/development hence, one could save this auto setting and use it with good success with the rest of the film stock if treated equally. There's more to getting a good scan but the conversion and removal of the mask is the first and most important global scan setting to aim for.

[Alan Goldhammer]

The D850 negative digitizer has some serious issues so not a great loss that you have the D810!

https://petapixel.com/2017/10/24/review-nikon-d850s-negative-digitizer-isnt-ready-prime-time/

The Nikon ES-2 film adapter looks like a cost-effective solution for your D810 with 60mm macro. You can start processing slides today using just LR (any version) and negatives using LR and PS or LR and SilverFast HDR. Please refer to the article Mark and I co-wrote detailing both workflows.

https://luminous-landscape.com/scannerless-digital-capture-and-processing-of-negative-film-photographs/

Thanks!  I had downloaded the article a while back but had not read it yet.  I'll also look at SilverFast HDR as per Mark's earlier post.

[Mark D Segal]

Isn't the unexposed section on all films clear except for the orange mask?  If so, then a program applies an algorithm that eliminates the orange mask to see no color in the unexposed section.  Maybe that's the way Epsonscan does it.  It doesn't care what film you used.  Maybe it uses levels and uses then unexposed area as the black point and adjusts the color to get the black.  Just some ideas.  Does anyone have a DSLR shot with the unexposed rebate section that they can post here?  A strait flat shot with no adjustments.  Then we can try levels and some other tests.

Ya sure. You didn't read what I've been trying to tell you? Are you trying to re-invent the wheel? We've been over that stuff through and through, years ago. And do you think that companies like Hamrick Software, LaserSoft Imaging etc. haven't devoted far more resources to this area than all the rest of us?

[Mark D Segal]

That's basically true of all scanning software although they may provide some film stock names as a starting point, in no way does that guarantee success of joy in the rendering. Not all software handles the conversions or even the auto (on what part of the film) equally. The Imacon mentioned did a good job with FlexColor in the old days. And it's sharp corner to corner and mentioned. The lens alone probably cost many times what anyone paid for the Epson in full. With the Imacon software, the trick with color negs was to crop small(er) areas of the exposed image and click Auto which when it found a 'good sampling' did a very good job inverting and removing the orange mask which is affected by exposure/development hence, one could save this auto setting and use it with good success with the rest of the film stock if treated equally. There's more to getting a good scan but the conversion and removal of the mask is the first and most important global scan setting to aim for.

Yes Andrew, the Imacon uses Rodenstock lenses that are apparently worth a couple of thousand dollars - i.e. twice the price of an Epson v800/v850 for the lens alone. Imcon still does a good job with Flexcolor from what I've seen out of a friend's Imacon. With SilverFast and Vuescan you don't need to do any of that sampling you mention. Their negative film "profiles" get you about 90% of the way, and then in SilverFast you can get the final 10% by minor editing of the "Negafix" profile of interest using tools provided in the Ai8 version of the application.

[Alan Klein]

I'm re-posting my request. 

"Isn't the unexposed section on all films clear except for the orange mask?  If so, then a program applies an algorithm that eliminates the orange mask to see no color in the unexposed section.  Maybe that's the way Epsonscan does it.  It doesn't care what film you used.  Maybe it uses levels and uses the unexposed area as the black point and adjusts the color to get the black.  Just some ideas.  Does anyone have a DSLR shot with the unexposed rebate section that they can post here?  A strait flat shot with no adjustments.  Then we can try levels and some other tests."

I'll work with you.  Just post the DSLR "scan" with no adjustments, only a positive flat picture.  Include the unexposed rebate portion

[Alan Klein]

I modified my previous post. I deleted my criticism,  It was wrong to say it and I apologize to the person effected.  If anyone has a DSLR "scan" with no adjustments, please post it here.  I'd like to play around with the orange mask. 

[Mark D Segal]

I'm re-posting my request. 

"Isn't the unexposed section on all films clear except for the orange mask?  If so, then a program applies an algorithm that eliminates the orange mask to see no color in the unexposed section.  Maybe that's the way Epsonscan does it.  It doesn't care what film you used.  Maybe it uses levels and uses the unexposed area as the black point and adjusts the color to get the black.  Just some ideas.  Does anyone have a DSLR shot with the unexposed rebate section that they can post here?  A strait flat shot with no adjustments.  Then we can try levels and some other tests."

I'll work with you.  Just post the DSLR "scan" with no adjustments, only a positive flat picture.  Include the unexposed rebate portion

Alan,

What you are trying to do is not going to work properly with a posted digital image of a piece of unexposed film. It won't have the spectral properties of the media and by the time it is processed for posting it won't be an accurate representation of any particular film stock. You need the piece of film itself, put it in your scanner and make the adjustments you want to make in order to invert it and neutralize it preferably in the scan software. So what you really should be asking for is for someone to send you by snail mail a piece of developed unexposed colour negative film. That's where to start from.

[Alan Klein]

Alan,

What you are trying to do is not going to work properly with a posted digital image of a piece of unexposed film. It won't have the spectral properties of the media and by the time it is processed for posting it won't be an accurate representation of any particular film stock. You need the piece of film itself, put it in your scanner and make the adjustments you want to make in order to invert it and neutralize it preferably in the scan software. So what you really should be asking for is for someone to send you by snail mail a piece of developed unexposed colour negative film. That's where to start from.

OK   Thanks. I'll have to scan it myself with my V600.  I'll scan the negative with the rebate "flat" as a positive and work it from there.

[trshaner]

Alan, if you're looking for a LR only color negative workflow please read my reply #21 and #24 in this thread. Mask removal is easily accomplished using the LR WB eyedropper on ANY neutral area of the negative. As mentioned you'll need to edit the camera profile (Adobe Standard is good) using the Adobe DNG Profile Editor to increase the range of the LR WB Temp control, which will probably bottom-out at the minimum 2,000K. Here's the link to the Adobe DNG Profile Editor (scroll down) and instructions:

https://helpx.adobe.com/photoshop/digital-negative.html

Both Mark and I tried using a LR only workflow with only marginal success due to issues caused by the inverted Tone Curve. It's not just a matter of removing the mask, which is easy enough to do. The inverted RGB color channels also need to be "balanced" to provide proper color correction. This is accomplished in SilverFast HDR using its Negafix control, which is similar to PS's Auto Curves control. In fact Auto Curves is what I use in my PS workflow described in the article Mark and I wrote. There is no similar "automatic" control available inside LR.

LR 7.3 introduced the capability to create LUT based XMP camera profiles. It is possible to create a camera profile that includes the invert function and a color correction LUT that does provide a LR only workflow:

http://www.cuchara.photography/blog/2018/5/one-click-inversion-of-color-film-negatives-in-lightroom

I tried the above procedure and created a LUT based profile for Kodacolor II 5035 color negative film. The image I chose to use for creating the camera profile looks great in the LR only workflow. Unfortunately the LR adjustments don't work properly and the controls have non-linear behavior making it very difficult to adjust other frames from even the same roll of film. Until Adobe offers a "built-in" solution for inverting the image file raw data the two workflows Mark and I outline in the article are your best choice for processing camera captured color negative film images.

[Alan Klein]

Alan, if you're looking for a LR only color negative workflow please read my reply #21 and #24 in this thread. Mask removal is easily accomplished using the LR WB eyedropper on ANY neutral area of the negative. As mentioned you'll need to edit the camera profile (Adobe Standard is good) using the Adobe DNG Profile Editor to increase the range of the LR WB Temp control, which will probably bottom-out at the minimum 2,000K. Here's the link to the Adobe DNG Profile Editor (scroll down) and instructions:

https://helpx.adobe.com/photoshop/digital-negative.html

Both Mark and I tried using a LR only workflow with only marginal success due to issues caused by the inverted Tone Curve. It's not just a matter of removing the mask, which is easy enough to do. The inverted RGB color channels also need to be "balanced" to provide proper color correction. This is accomplished in SilverFast HDR using its Negafix control, which is similar to PS's Auto Curves control. In fact Auto Curves is what I use in my PS workflow described in the article Mark and I wrote. There is no similar "automatic" control available inside LR.

LR 7.3 introduced the capability to create LUT based XMP camera profiles. It is possible to create a camera profile that includes the invert function and a color correction LUT that does provide a LR only workflow:

http://www.cuchara.photography/blog/2018/5/one-click-inversion-of-color-film-negatives-in-lightroom

I tried the above procedure and created a LUT based profile for Kodacolor II 5035 color negative film. The image I chose to use for creating the camera profile looks great in the LR only workflow. Unfortunately the LR adjustments don't work properly and the controls have non-linear behavior making it very difficult to adjust other frames from even the same roll of film. Until Adobe offers a "built-in" solution for inverting the image file raw data the two workflows Mark and I outline in the article are your best choice for processing camera captured color negative film images.

I found that color negative film is harder to adjust.  I had tried some Kodak Ektar 100.  So I'm sticking now with Velvia 50.  I know exactly what I shot.  No question about exposure.  I also bracket consistently.  I also prefer its palette.  But I'm willing to give negative color a chance again.  It has a lot of advantages over slide film, especially Ektar.

Is there a way you could email a complete "scan" of an original negative image?   I'd like to try some edits on my own.  Thanks.

[trshaner]

Is there a way you could email a complete "scan" of an original negative image?   I'd like to try some edits on my own.  Thanks.

Below is a link to the article PDFs, other processing information, and a Fuji Super G100 color negative raw image file using camera capture. There is no unexposed frame area, but there are many "neutral toned" areas in the image for white balancing with the LR Eyedropper. This image was used in the article for the LR/PS processing discussion.

https://www.dropbox.com/sh/u5pnzbesvqen1hv/AABZyaWAn33jCwEinkT45iFAa?dl=0

[Alan Klein]

Below is a link to the article PDFs, other processing information, and a Fuji Super G100 color negative raw image file using camera capture. There is no unexposed frame area, but there are many "neutral toned" areas in the image for white balancing with the LR Eyedropper. This image was used in the article for the LR/PS processing discussion.

https://www.dropbox.com/sh/u5pnzbesvqen1hv/AABZyaWAn33jCwEinkT45iFAa?dl=0

Thanks for the image and links.  Is the image without adjustment?  How was it "captured"?

I'd also like to get the rebate area as well.  Do you or anyone else have one?

[trshaner]

It's a Canon 5D MKII CR2 file shot with 5,500K light source and being a raw file it has no processing applied. Sorry, I don't have any captures with the rebate non-frame area showing. That's unexposed film, which may be useful for mask removal, but then the image still needs to be white balanced dependent on actual scene lighting and film characteristics. The LR white balance eyedropper corrects both color temperature AND mask removal in one-step. As Mark mentioned this is academic stuff that has been discussed ad infinitum. What exactly are you trying to accomplish (details!) that hasn't already been discussed here or in the article and materials at the link supplied? Seriously, it's a futile effort to try and reinvent the wheel. Both Mark and I have been there and done that over hundreds of hours working with LR, PS and SilverFast HDR.

The area that still needs more investigation is the use of film specific color-correction LUT based XMP camera profiles. From my tests using an inverted Tone Curve in LR they don't work very predictably. To make this a viable color negative processing solution Adobe needs to provide a true raw data invert function.

[VincentDJohnson]

Finally got around to working this out.

Did all scans with sharpening off and ICE off.

Tried flat on the glass. Flat on top of the removable glass plate that came with the Epson. And then again with the 4x5 film holders but the little removable boosting claws taken out. The last option seemed to work best. Still not 100% where i'd like it to be. Have a little color issues too. Need to see if I can get the now out-dated version of Silver Fast that came with the scanner upgraded.

But once again, thanks for all the input. Just needed a good weekend to sit down and try it all. Also didn't hurt I have like 30 new 4x5 shots to scan in.

[Mark D Segal]

The scans need to be sharpened, but I wouldn't recommend doing that from scan software even though most of those products provide sharpening tools. The reason is that once you sharpen at the scan stage, that sharpening is baked-in and you can't change it unless you rescan, which is a nuisance. As well, the sharpening tools in scan software don't use the three stage sharpening work flows that Bruce Fraser introduced us to years ago, make so much sense and work very well, especially for film scans - Photokit Sharpener for example has an in-built sharpener at the Capture sharpen stage for film scans, which is also adjustable, adaptable and non-destructive s long you retain the sharpening layers. Sharpening film scans in LR is also a good option.

[Alan Klein]

Mark  What procedure do you follow to sharpen in LR?

[Aram Hăvărneanu]

Hey Vincent, you might want to take a look at Nick Carver's video about the best way to mount film for scanning on Epson scanners: https://www.youtube.com/watch?v=mfGKd_AEZtY.

[Mark D Segal]

Mark  What procedure do you follow to sharpen in LR?

LR provides for sharpening at two stages: Capture and Output. For capture, I start with the proposed presets for scenic or faces depending on the image, examine the results at 100% and see whether any fine tuning is needed. For output, one turns it on or off; it automatically calculates the needed sharpening based on dimensions. I keep it on.