Jump to content

Slide film: Using ColorPerfect/linear scans vs non-linear TIFF


philipus

Recommended Posts

Advertisement (gone after registration)

In Steve Blitz's recent interesting thread Color film choice I was asked to expand on why I usually avoid using Color Perfect and linear scans for my slide films. I was invited to put my explanation in that thread but on second thought it might be useful to have it in its own thread (also because it would perhaps open the topic to more viewers which might mean that I could get more feedback on the workflow; I certainly don't claim that the following is any rule or law to go by).

 

When I began using CP a few years ago it was in order to find a way of dealing with the orange mask and arrive at an OK-ish colour temp as easily as possible. For C41 CP is, in my limited experience, very, very good but there are two main reasons why I don't use it for slide film, one is image quality-related and the other is workflow-related.

 

I don't subscribe to the view that an image is final out of the scanner and will virtually always post-process my scans. With this in mind I try to keep my workflow as simple as possible; if I can cut steps I will.

 

CP's default setting appears to be to blow highlights by varying degrees. I seem to remember having read that this is by design because the plugin was designed mainly to be used with colour negative film which has a considerable latitude. How much the highlights will be affected depends on the scan so it is difficult to be generic but in virtually all slide film scans I make the Black slider (which controls highlights) will be above, and sometimes far above, zero.

 

CP does include highlight and shadow recover functions but they - and in particular the highlight recovery function - don't work very well with slide film, generally speaking (examples to follow) and sometimes also not very well with C41.

 

Because the highlights are 'stressed' by default I will have to adjust the image within the plugin itself. This adds a step to the workflow (with several sub-steps). I would happily accept this extra step if the results were good, but in my view they rarely are. In fact, the results will often be unpredictable. Beginning to edit the image in CP will set the whole post-processing exercise off on the wrong track. Naturally, for C41 going into CP is OK as a matter of workflow because I need the plugin to deal with the orange mask. However, even if CP had not caused poor results in the highlights, there is very little else that is useful (imho) in CP for slide film scans. Take for example its colour controls. For slide film, CP's main colour correction tools are the "Ring Around" function and the auto colour function. While Ring Around is quite intuitive as a matter of theory (it's based on adding and subtracting opposing colours a bit like Lab works) it is in use quite fiddly with nine smallish windows representing various colour temperatures. It should be noted that CP's image preview is generally considerably worse than what Photoshop shows (apparently this has to do with how plugins work within Photoshop, or something to that effect) and this has a negative effect on the Ring Around function, too. The auto colour function is the other main colour tool in CP but I find that it doesn't work very well with slide films because it requires using a long vertical slider to achieve a useful colour temp and it is just not very intuitive (for some reason it works better with C41 actually, I don't know why; and of course for C41 there are lots of film profiles which can - but don't always do - get a very accurate first colour temp). 

 

Here are a few examples of what I mean. I know this is not the best image but I hope it gives an indication. This is a linear scan (using the Vuescan settings recommended by the CP folks). This makes a slide come out like this at the other end. 

 

2_Linear_raw_s.jpg

 

Opening this file in CP and setting the mode to L(inear) in ColorPos results (having only OK'd out and saved as JPG) in this. To make it easier to see what I mean I'll use a crop of this frame in the following.

 

2_Linear_raw_Color_Pos_L_s.jpg

 

And the crop:

 

2_Linear_raw_Color_Pos_L_crop.jpg

 

As can be seen, due to CP's default setting the hat's highlights (and the shoulders) have been severely blown. Using this image for further post-processing will cause very negative effects in the highlights as, for instance, curves are added. 

 

To get a more neutral scan using CP I would need to counteract the blown highlights by pulling the Black slider to zero (I'll come to the highlight recovery tool in a minute). Oddly the White slider, which controls the shadows, is often by default at the equivalent of the Black slider's zero (even though the number value isn't actually zero). I say 'equivalent' because moving the White slider towards zero will not make the shadow areas less dense (in the same way that moving the Black slider towards zero brings out more detail in highlights); rather miving the White slider towards zero will only make the shadows wash out and become less contrasty.

 

In this version, I have pulled down CP's "Black" slider to zero. This brings back detail in the hat and makes the image slightly duller.

 

2_Linear_raw_Color_Pos_L_zero_highlights

 

In many cases it will not be possible to recover highlights very well in slide film. So, depending on the image, scanning linear and using CP may result in completely irrecoverable highlights. Very contrasty scenes will often result in particularly poor scans in this respect.

 

About CP's highlight recovery function. This function (and it's shadow-related sibling) is designed to recover blown highlights (or shadows) by selecting by how many "f stops" these areas of the image should be corrected. As great as this sounds I have found that the highlight recovery function often results in "unnatural-looking" highlights.

 

Below is the linear scan 'developed' in ColorPos L (without zeroing the Black slider but) with highlight correction set to 0,5 stops. The brim of the hat looks a mushy yellow. In my experience, the more blown highlights are, the worse will they look after using the highlight recovery function. Very often for very contrasty images (and even for colour negative images) the highlights take on a blueish/greyish/pale and generally quite terrible look.

 

2_Linear_raw_Color_Pos_L_0_5_stops_highl

 

Now for a non-linear scan -- i.e. where the RAW checkbox on Vuescan's Output tab is unchecked and TIFF is instead checked. This gives the following out of the scanner. Note that in linear/RAW mode Vuescan will ignore the Color Balance settings on the Color tab (at least; possibly also other settings, I don't know). However in non-linear mode it makes sense (to me and my workflow, at least) to use one of the Color Balance settings to simplify the coming steps in the workflow. I usually use Autolevels which is "good enough" for further editing. Here is the non-linear scan of the full frame for comparison with the crop of a non-linear scan using the Autolevels setting.

 

2_non_linear_cb_autolevels_s.jpg

 

And the crop

 

2_non_linear_cb_autolevels_crop.jpg

 

Ignoring for a moment the difference in colour between this scan and the second one from the top above, this non-linear scan has much less blown highlights than the linear/CP scan. And note also that the shadow area of the hat is also less dark. The big difference between the two images is that to arrive at a similar result in terms of how much information is retained in the highlights with a linear scan I need go into CP and make adjustments which adds a step (with unpredictable effects depending on the image) in the workflow. True, the colour may (again depending on the image, based on my tests) look 'richer' using CP (as the above shows), but given that I will always edit my images further using Photoshop and/or Adobe Camera Raw that is of little concern.

 

Sorry for the long post, but I hope my thought process is at least a bit clear. I'd be very interested in hearing other members' thoughts on this. I will also see if I can find another image to show a more dramatic difference.

 

br
Philip

  • Like 8
Link to post
Share on other sites

This is very interesting. I'm going to digest this information and get back to you when I've tried some scans using your technique (probably will have to wait for the weekend). The highlight and shadow controls in CP have been a concern of mine for a long while, and I've discussed them several times with the application developer. This could be working towards a solution - at least for those really difficult, contrasty images that don't work well within CP.

 

Thanks so much for the clear and thorough explanation!

  • Like 1
Link to post
Share on other sites

Thank you for your comments. This was something I had been meaning to do anyway, so being asked in the other thread provided the necessary momentum to put pen to paper. 

 

Btw, I forgot to say that when I asked the developer in the CP Flickr group about CP's preview looking different from what Photoshop itself will show he mentioned that PhotoLine apparently handles this better. Could be of interest to some, perhaps.

 

I agree that one will have to tailor the workflow a bit depending on the image in question. With 35mm I try to be effective and aim to scan a whole roll in one go with "good enough but as good as possible" results. This is what has caused me to rethink regarding CP. But I will re-scan and re-edit a particularly important image that I want to, for instance, print. And in some, rare cases CP will actually produce a better image than what I get with a non-linear TIFF.

 

It's a bit different with 120 film. Because Vuescan doesn't recognise frames properly (at least I've never managed to get it to do that) I need to preview and scan each frame individually. Vuescan does offer the possibility of scanning linear and non-linear at the same time; it's just a matter of ticking both the Raw and TIFF boxes. Naturally, this works with 35mm film as well but I don't use it because I'm usually happy (bearing in mind my 'aim' above) with the results I get with non-linear scans. For 120 film I will sometimes scan in both modes simultaneously and then use the one that looks the best for further editing. In almost all cases that's the non-linear TIFF but not always.

 

br
Philip

  • Like 1
Link to post
Share on other sites

This is really interesting. 

Just some random notes if it helps, before getting into any testing at the weekend:
 
- I use PhotoLine and Photoshop in parallel. I like PhotoLine because practically everything can be modified with keyboard shortcuts, so that the workflow for opening and doing the initial ColorPerfect processing takes just a couple seconds. Photoshop also had serious issues with  previews in CP when the Retina screens first came to the MacBook Pros. PhotoLine is an almost secret dark horse in the image-manipulation field - no-one ever seems to have heard of the application. One thing I especially like, is that the dust-spotting tool is MUCH better than Photoshop’s tool (although without some of Ps’s bells and whistles). This alone saves me a lot of time.
- the workflow you describe here is mostly for transparencies, but a lot of the problems I recognize from C41. I’m really interested to know to what extent this might fix C41 pain-points.
- I think there are marginal issues when combining Vuescan and ColorPerfect. Vuescan seems to produce a slightly more contrasty ‘linear’ scan than Silverfast, for instance, and this causes problems when ColorPerfect ‘develops’ the negative. I’ve tried to get to the bottom of this for the last few years, without any real success (I’m just a hobbyist mostly taking family snapshots, so I haven’t really done anything concrete other than write occasional mails to the developers).
- There are some weird workarounds that I’m exploring for problematic high-contrast negs. One of them is that I noticed that the overall initial contrast of an image pulled into Cp can be moderated by importing the negative with a large pure-white square in the middle of the image (which becomes pure black), and going through Cp’s weird ‘save master’ settings (can’t remember right now how this is done) to keep this as a preset, then reimporting the image without the additional white area and using the preset as the initial process. All sounds crazily complicated, but I’m really just trying to get to some sort of answer about why Cp sometimes behaves the way it does.
- Just as a sidetone, I haven’t been using any film profiles in Cp for a few years. I actually import the color image with a B&W preset - which strips out all extra contrast and color-manipulation. I found the film profiles didn’t help in any way - I’d rather post-process the image the way I want it in Photoshop and Lightroom.
 
Long post with probably very little additional information. Still, I find the whole workflow area fascinating. 
And ColorPerfect is a great tool in my view, for 90% of the images I take. That last 10% is what I’d love to solve: mostly because I spend most of my waking life in front of a computer screen, and I’d like to cut my private computer time to the absolute minimum possible.
 
Thanks again, Philip! Fascinating stuff.
  • Like 2
Link to post
Share on other sites

Advertisement (gone after registration)

I have used CP for C41 previusly, but did after some time find the same issue with highlights. They burn out easily, and if try to recover they look plain ugly, and to be honest where digital. Today I scan as tiff or jpeg, and adjust in LR or C1 afterwards.

  • Like 2
Link to post
Share on other sites

I tried a few variations of Philip's approach above, but nothing helped my C41 workflow - so I'm sticking with the linear Vuescan capture, followed by Colorperfect conversion.

 
FWIW the tweaks that are a normal part of my workflow when importing to ColorPerfect are:
 
- I've saved a lower threshold for Black Point, so that the image is converted from the start with less 'black' in the shadows. This makes for a slightly less dramatic, flatter looking image, but I consider it a better starting point. I most often pull down the black point clipping even further, which flattens the curve even more - many people would see the result and feel it looks insipid.
 
- I've got into the habit of opening curves in Cp and flattening the image even more. The curves are implemented to resemble 'zones' in Colorperfect, and need a few clicks to open up: clicking on Options then Curves, then clicking the little '+' symbol to see the zones. I then click to select which zones I wish to edit and/or lock, and then slightly tweak the curve - usually to pull a little light into the shadow area. I try to keep the changes as subtle as possible - I don't want anything like an HDR look - but I want to moderate the accentuated contrast that Cp lays over my original image.
 
- I usually pull the highlight clipping down - as others have mentioned, sometimes this looks terrible. I generally think that if the highlight clipping looks totally off, then the image has an unnatural color cast that I haven't otherwise noticed. Sometimes it really doesn't go to fix, or I actually like the color cast as it is (I'm not always aiming for a totally neutral rendition). 
 
- The 'Ring Around' color tweaking works pretty well - most often it makes obvious something I've missed if I want to make the image neutral. I'm not sure whether there may be a handicap if your color vision isn't good - this is often missed in discussions about color: color acuity varies greatly from person to person, and in different parts of the spectrum, and also deteriorates with age. (I often think this is a critical factor when old men start saying stuff like 'digital color is so much better than film color'. ymmv)
 
The image that comes out of Colorperfect is most often not some neutral, 'perfect' image - but in many cases it's WAY better than anything I can do myself with Photoshop. It's a solid starting-point for Lightroom.
 
95% of the images I take are family snapshots in any case. so I always sense a certain amount of wonderment that I even bother when discussing these matters. I even got a PM to that effect over on RFF from one of the moderators, so my advice is really of the 'take-it-or-leave-it' variety. I'd just like to get a workflow that was a lot quicker to implement and gave me fewer marginal problems - like the filled shadows, or 'dirty' highlights that I feel are the weak points in Cp.
  • Like 2
Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...