VueScan Question

[IanBarber]

I am using VueScan with an Epson V800 to scan Black and White negatives.

After performing a Pre-Scan I select the area of the negative to be scanned indicated by the dashed lines.
In VueScan, I have the clipped highlight and shadow warnings enabled, blue indicates clipped shadows and Green indicates clipped highlights.

In the example below, notice how parts of the deep shadow areas of the rocks are showing blocked.




If I extend the area around the negative by extending the dashed lines, notice how the areas in the rocks are no longer showing clipped.



Can anyone explain why extending the crop box eliminates the blocked up shadows. Without a proper densitometer, I am unable to ascertain whether the negative has lost detail in those areas or whether VueScan is incorrectly showing areas with no detail

[Ted Baker]

If I understand your question.

Try and put the raw graph on, and scan (cntrl 1, or image menu -->graph raw). That will tell you if the exposure is clipped in anyway, I doubt it is. If there is no clipping in the exposure and there is clipping after the inversion then it is a result of the inversion logic and the settings you have used. Moving the crop box in combination with some of the setting will affect the inversion logic.

Hope that helps.

[IanBarber]

Ted, this is what the RAW graph looks like after Turing it on and doing another pre-scan

[Ted Baker]

Ted, this is what the RAW graph looks like after Turing it on and doing another pre-scan

Exactly no clipping in the exposure, you even increase the exposure a little bit if you wanted too, the graph will move to the right. In the DLSR world they call this ETTR.

The clipping you are seeing is caused by the algorithm, and the settings you are using. If you were to print this optically what grade paper would you likely use? This is the part of Vuescan that I definitely do not like, as it is very hard to understand what the controls are supposed to do.

[IanBarber]

Exactly no clipping in the exposure, you even increase the exposure a little bit if you wanted too, the graph will move to the right. In the DLSR world they call this ETTR.

To increase the exposure, would I Check the Lock Exposure option and then increase the RGB Exposure value

[Ted Baker]

Yes that is the one, when you uncheck it, you will see the factor that Vuescan already applied, it is likely to be range 1 to 2 try increasing it by about 50%. Your scan will take longer, you may not see any difference because the densest areas were already far enough away from the left. But do not clip it to the right, I don't go any further than about 90% it is not necessary to be exact. What I like about this feature is that it is repeatable which is useful for batch scanning the same stock, no preview required (for raw), its also useful IMHO if your interested in calibration for colour negative work.

The main point in your case is the clipping is from the software not the exposure.

[IanBarber]

Thats what I thought but something appears a drift here.

Perform Pre-scan with Lock Exposure Unchecked
Check Lock Exposure which reveals the RGB Exposure value which VueScan has set to 1.351
Increase value to 1.55 (15% increase) and it crushes the shadows even more and the graph now shows



Looking at the graph, I am presuming the shadow values are indicated on the right (opposite way to how photoshop displays them)?

[Ted Baker]

Thats what I thought but something appears a drift here.

Perform Pre-scan with Lock Exposure Unchecked
Check Lock Exposure which reveals the RGB Exposure value which VueScan has set to 1.351
Increase value to 1.55 (15% increase) and it crushes the shadows even more and the graph now shows



Looking at the graph, I am presuming the shadow values are indicated on the right (opposite way to how photoshop displays them)?

If you have the raw graph turned on the the least dense parts of the negative are on the right (the shadows). Try a smaller amount 1.4. Its with colour negatives and 35mm that you may need up to 2. Colour neg for example has a density base of around 0.9 in one of the layers.

You can also try a factor of 1, which is no exposure compensation and it will all move to the left, the main point point being that vuescan already does a pretty good job already.

With the raw graph think of the negative as a real scene, the brightest areas are the clear film, and the darkest has the most silver.

[mdarnton]

Can anyone explain why extending the crop box eliminates the blocked up shadows. Without a proper densitometer, I am unable to ascertain whether the negative has lost detail in those areas or whether VueScan is incorrectly showing areas with no detail

Extending the crop box includes the unexposed margin into Vuescan's automatic calculations. If your thinnest area on the neg is thicker than the black margin of the film, then it will be set to a level above zero, since the margin is automatically set to zero. Basically, you have fooled the program into thinking that the margin represents the lowest value in your picture. This only works if you expose enough to make sure there are no clear areas of the neg within your subject area (which is how I expose), and I use this trick to get a scan that's slightly flat at that end of the scale. If you don't expose that way, then including some of the glass around the film works, too, because the glass doesn't include the film+base fog densities.

Unless you want to use RAW in Vuescan, this is, in my opinion, a bad habit of Vuescan that isn't easily gotten around. The other end is more problematic, and I have though that I might include a black paper label stuck to my scanner window against the edge of the neg so that VS will be fooled into reading that as the densest area, which will give me some headroom at the other end, too

If anyone has any advice on preventing VS from snapping to what it thinks are the appropriate 0 and 255 points, leaving some room at each end, I'd like to hear it. It's fine that the clipping isn't in the exposure, as indicated on the RAW graph, but if you can't get that graph out of the software end without actually using RAW, well, that's a software defect, in my opinion.

Vuescan works well with my scanner, but the interface is complete garbage and non-intuitive. I've been using it for a couple of years and still haven't figured it out. There's probably a way around any problem, but try to find it. I remember learning PhotoShop, and this is much worse, for a program that does much less. I think part of it is the automatic features, even in the "Professional" mode. Just give me a RAW window with some points I could directly slide around, that would directly output a TIFF as I had mapped it and I would be fine, but that doesn't appear to be possible.

You shouldn't need to fool a program into delivering what you want!

[IanBarber]

What is interesting is that when VueScan shows areas in the preview as been blocked (blue n my case) when brought into Photoshop, those areas are showing RGB 0 which shows the VueScan is accurate.

Now... If I do the following, I can remove the blocked areas in VueScan..

After the Pre-Scan, Lock the Exposure and tweak the slider to get a more even spread, then in VueScan scroll down and check the Lock Film Base Color check box
Now under the Color Tab, three more options are revealed, Film Base Color Red, Green, Blue

By increasing these values slowly, you can watch the blue areas (Indicating blocked) slowly disappear and when scanned and opened in Photoshop, these areas are no longer RGB 0

I haven't as yet worked out why increasing the Film Base Colors appear to lift the shadow values.