Exposing for negative film

[skysh4rk]

I have a Plustek scanner, and have used both multipass (which I hink reduces noise) and multiexposure, which does an extra pass at higher exposure to get a bit more detail out of the denser parts of the image (presumably shadows for transparency and highlights for negative film). I've often felt it was a useful feature.

I also suspect that if I understood how to use curves, I might be better at getting more of that detail off the film and into the scanned image. (And/or/maybe if Vuescan's curves implementation was a bit better?)

I don't really think that the additional scan is necessary for negative film; the scanner should be able to blast through the denser parts of a negative without any problem (a transparency on the other hand...). You could simply do one scan, process it for the highlights, take another copy of the same scan, process it for the shadows, and then combine them.

I don't think the curves adjustment will do much to address the problem, because you still have local areas that are much brighter than others in relative terms and the curves tool will only make global adjustments.

Surprising that UKFL only offer JPEGs; I really did expect they would provide 16/48-bit TIFFs, considering their target market.

I imagine that their target market is wedding/event film shooters that don't want to mess around on the computer editing their images. You're basically paying them for the service of providing the finished product (or near enough), as they individually adjust each scan according to your preferences, hence no need for TIFFs in theory.

 

[skysh4rk]

Thanks RJ. I used this shot for illustration, as the reflection shows what the sky was like. Pretty much all shots on the roll have little visible detail in the skies, from the scans. I'm a bit confused about whether the foreground is really in shade (as you suggest); the waterweed is definitely in sun, but I suspect the other side of the pond is in shade from a cloud. I don't remember thinking the sky was particularly bright at the time. I've generally been under the impression that actual ND Grads are rarely needed with negative film.

Here is a visual explanation of what I've been talking about with regard to the differences in brightness between the foreground and the sky.

The pictures below were taken just moments apart, but in the first one Clyde is lit by the sun (notice the directional shadows), but in the second one he is now shaded by a cloud that quickly passed overhead (see that the shadows have all disappeared). Both photos were shot at the same EV (metered for the shadows), so there isn't a dynamic range issue as you can see detail in the sky in the first one, but notice the difference in the colour of the skies in the second even though the sky was exactly the same in both (i.e., cloudless and blue).

Because Clyde is sunlit and of a similar brightness to the sky in the first photo, when the scan is adjusted to suit the foreground, the detail in the sky is retained. In the second photo, Clyde is now several stops darker as he is in the shade, so when the scan is adjusted to get him to a similar level of brightness as in the first photograph, the sky is much brighter and loses some of the detail (although it does not blow out as it might have on digital).

So we have the exact same sky, the exact same EV (so we know it's not an issue of dynamic range, exposure, metering, or film), but because of the different light falling onto the subject, it has affected the look of the sky.

Negative film is amazing at capturing lots of detail, but it still doesn't change the light falling onto the scene.

View attachment 20107 View attachment 20108

 

[ChrisR]

Thanks RJ, very nice example. I'm currently scanning my sample frame to TIFF, to see what difference that might make. It's already obvious that there is much more detail on the film than I was seeing before (as you suggested). I'm not having much luck with Hamrick's suggested method of scanning the leader and locking exposure and base colour, although I have done a scan that way; now doing a second with the Portra 160VC preset (Vuescan doesn't have a preset for modern 160). However in this case it's not the colour that's the issue.

 

[ChrisR]

OK, this is interesting. Scanned to TIFF on the Plustek 7500i with Vuescan, 4 passes and multi-exposure (as suggested, this may not have helped but it's what I usually do). A very slight curve in Vuescan to increase contrast in the brighter areas. I did one with locked base colour, but I think I must have messed up as it looked awful. This was scanned again with the Kodak Portra 160VC preset (no detectable difference to my eyes from the Portra 160NC preset). Uploaded to Aperture, Levels adjusted to get the histogram just within the range.



I'm not quite sure which element above made the most significant difference from the scan done by Photo Express (adjusted version(s) shown above). As I've said elsewhere I've been very happy with their scans in general. When I initially had problems with my own colour negative scans using the same Plustek with Silverfast and a trial version of Vuescan, I sent the film to Photo Express and got a really nice scan (for £5 the film) with excellent colours. It looks as if there's something about how this current scan was set up that hasn't worked for the sky areas. Maybe I should ask them to have a look at this thread and comment?

In this case I don't think scanning to TIFF made any difference, as I've just saved the scan to JPEG and the cloud detail is still all there in Aperture (I'm sure it has an impact with some kinds of adjustments).

So, going back to my original concern: contrary to my previous worry, I'm now happy that I can rate the fim half box speed, expose for the shadows, and still get proper detail in the brighter areas! But I may need to think about how the scan gets done.

 

[s162216]

Thanks RJ, very nice example. I'm currently scanning my sample frame to TIFF, to see what difference that might make. It's already obvious that there is much more detail on the film than I was seeing before (as you suggested). I'm not having much luck with Hamrick's suggested method of scanning the leader and locking exposure and base colour, although I have done a scan that way; now doing a second with the Portra 160VC preset (Vuescan doesn't have a preset for modern 160). However in this case it's not the colour that's the issue.

I wouldn't bother with the pre-sets as they are all very out-dated. Locking the exposure and base colour does essentially the same job anyway, but it's essential that you adjust the levels of the scan afterwards or in Vuescan as obviously the scanner has no idea of what should be the brightest or darkest part of a picture.

Using the exposure lock setting should guarantee you get the maximal range of the negative, but then adjust the levels until just before clipping is induced (if you want to do it in Vuescan, turn on "pixel colours" in the "colour" tab and set the colour balance to "neutral", then go to the "image" menu and select "graph B/W". This will bring up a histogram, pull the arrows at the bottom to adjust the output levels. I usually adjust them until just literally before clipping is introduced [look for the warning colours]. One problem with doing it in Vuescan though is that it doesn't have anything to adjust the midtones, which is what you need to adjust to get into the shadow/highlight detail so I usually just do it manually in Photoshop Elements, plus you can't adjust each channel individually in Vuescan).

Scanning to TIFF is unlikely to actually make much difference unless its a 16 bit one as theres practically no difference between a maximal quality JPEG and a TIFF otherwise if they are both 8 bit.

Sam.

 

[ChrisR]

Interesting Samuel; I did save the locked base colour TIFF and I'll have another look at it with your comments in mind. I used a clear part of the leader, but when scanned it seemed to have quite a colour variation, and it was hard to get the cropping to a consistent area (in fact I'm pretty sure there were variations, which is why I commented that I'd messed up). I did use a 16/48-bit TIFF; they are over 40 MB which is annoying (I wish more software would use JPEG2000, which can use 16 bits per channel with a better compression algorithm than TIFF, but that's another story).

 

[s162216]

It will vary slightly, just select the whole frame or the largest consistent area. I use JPEG2000 (with a Photoshop plugin) to save some of my edited versions (prior to down-converting) and it does help to save space whilst being lossless, but I suppose if you have the the option to save as a JPEG2000 elsewhere (in Aperture etc), you could just possibly save the original scanned TIFF as a lossless JPEG2000, and just delete the original TIFF? Turning on compression for TIFFs in Vuescan does also help to reduce their size a bit (and it is lossless).

Sam.

 

[skysh4rk]

I'm not quite sure which element above made the most significant difference from the scan done by Photo Express (adjusted version(s) shown above). As I've said elsewhere I've been very happy with their scans in general. When I initially had problems with my own colour negative scans using the same Plustek with Silverfast and a trial version of Vuescan, I sent the film to Photo Express and got a really nice scan (for £5 the film) with excellent colours. It looks as if there's something about how this current scan was set up that hasn't worked for the sky areas. Maybe I should ask them to have a look at this thread and comment?

Chris, you still need to address the imbalance in brightness between the foreground and the sky at exposure if you really want those colours in the sky in your photographs; nothing you ask Photo Express to do will change this unless you're prepared to sacrifice lots of foreground/shadow detail.

This is what I was trying to show with the two photos that were taken at about the same time, with the same sky, at the same EV, scanned on professional equipment, and scanned by superior operators (better than I am anyway), but just with different light falling on the foreground.

 

[FujiLove]

Going back to the question of how to meter and develop black and white film...

I now have a few rolls of Tri-X 400 that I've shot at half box speed, so it's effectively over-exposed by a stop. I've used my light meter to take incident readings in an area of shadow, which again will have added around a stop or so of extra exposure. Now I'm looking at the massive dev chart and wondering which time to use. I don't want to pull the film, so I guess the values for 200 ASA are incorrect? Should I be using the time for 400 ASA so all the over exposure is retained, or reducing it by 20% as mentioned above? I don't think I understand the reasoning behind the 20% reduction. Wouldn't that cancel out the over exposure and leave you with something similar to shooting at box speed?

 

[StephenM]

This is an oversimplification, but it's still fairly accurate.

1. Beyond a certain minimum developing time, shadow density doesn't increase much. (That's why you can develop overnight to try to get some detail, without everything going black.)

2. After that minimum point, it's highlight detail that builds up.

3. That's another way of saying that contrast is controlled by development time.

4. And that's why there's the old saying "expose for the shadows and develop for the highlights".

To reduce contrast when the subject brightness range is high, give more exposure and less development. Fot flat subjects, do the reverse (and note that low contrast subjects are by definition ones where the film's latitude is greatest).

After that comes the watershed. If development is pushed and the highlights go dense, you have a problem fitting the contrast onto a conventional darkroom print. There's a maximum advised density set to allow for this. But if you're going to scan and print digitally, all bets are off, because you have more control of the contrast when you have a digital file.

Personally, given the probable/possible overexposure, I'd develop normally. If the film is outdated, the contrast may have dropped along with the speed, and more development would be indicated.

If your negatives are too dense, there's always Howard Farmer to give you a helping hand.