Scanning Consistency (Warning: Pictures)

[plasticman]

Advertisement (gone after registration)

...It was a bolt out of the blue. I'm looking for some pointers, providing a thread where I was hoping everyone could pool ideas, like the other 2 scanning-related threads, and then someone who's most useful comment has been "Buy an M9" tells me that I'm ignoring the very information I seek.

Pete

 

Hi Pete - I for one really appreciate the lengths you're going to trying to make the Film forum a place for constructive discussion and an all-round useful reference for scanning and post-processing film. I'm possibly wrong, but in my view a lot of the scans you've shown in these several threads have been really good.

 

Anyway, I expect you've long ago seen this site, with lots of information on the whole process (this page for instance), but also with the craziest navigation ever - the English and German versions of the site are 'parallel' to each other, but not all pages are translated into English. Still worth a look though.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[Stealth3kpl]

Thanks Andy.

How do set Vuescan up to give a TIFF with no clipping of each RGB channel? What is auto levels in Vuescan? Can you say more about adjusting each RGB channel? Are you simply setting the black/white point for each channel or are you adjusting gamma or anything else?

 

Scanning success hinges on getting all the information off the negative so no clipping. I still don't know how this is achieved unless saving a raw file. The closest I've got to an ok looking positive tiff is this but you'll see clipping at each end.

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

Just thinking about it as I sit here, perhaps my file is clipped at both ends because I raised the contrast in Vuescan. Perhaps I should've left the image flat.

 

Next, scanning success hinges on how each RGB channel is manipulated.

 

Thanks for posting.

Pete

[Stealth3kpl]

 

Anyway, I expect you've long ago seen this site, with lots of information on the whole process (this page for instance), but also with the craziest navigation ever - the English and German versions of the site are 'parallel' to each other, but not all pages are translated into English. Still worth a look though.

 

Thanks for those links. I had seen them some time ago but once I'd worked out how to get Linear Raw scans I didn't return because I was trying to get an angle on ColorPerfect. Thanks for the reminder of their existence. I'll have a read of them now I'm coming at Vuescan from another angle. (edit: I've just realised these are the ColorPerfect pages but you've reminded me to take another look at the Vuescan help site).

Thanks for your kind words. I've been so frustrated looking around the web for scanning information. I was hoping we'd have some information-filled threads in the Leica film forum for us and future users so that they wouldn't have to go through what I've been going through for the last few months. I've got to the point that I'm thinking there must be a colour balancing step that is obvious to everyone but me and that's why it's never mentioned.

At some point it would be good to collate the good points into a sticky.

 

Incidently, some users might not realise that right clicking on the image in vuescan gives a mid-grey point.

Pete

[andybarton]

I don't recognise the software you are using there, so I can't help with that at all, I'm afraid.

 

I always scan "flat", with colour or b&w, and then adjust contrast in Photoshop.

 

You can adjust the sliders in Vuescan so that you don't clip any of the channels. Here, curve low has been adjusted upwards to avoid clipping blue. This scan looks pretty good in the Preview screen, actually, so won't need much doing in PS. This is contrastier than I would normally scan - each film is different and you need to learn how to use Vuescan to get the best from your scanner and film.

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Stealth3kpl]

STAGE ONE

 

I always scan "flat", with colour or b&w, and then adjust contrast in Photoshop.

 

That makes sense to avoid clipping

 

You can adjust the sliders in Vuescan so that you don't clip any of the channels.

Which sliders are you referring to here? Just the curve low/curve high sliders or the colour sliders or all the sliders?

 

You can adjust the sliders in Vuescan so that you don't clip any of the channels. Here, curve low has been adjusted upwards to avoid clipping blue. This scan looks pretty good in the Preview screen, actually, so won't need much doing in PS. This is contrastier than I would normally scan - each film is different and you need to learn how to use Vuescan to get the best from your scanner and film.

 

 

So

1) You dial the black point and white point down to zero.

2)You don't right click any mid grey.

3)One could adjust the individual colour sliders to avoid clipping a channel, but you didn't need to in this example because when you adjusted CURVE LOW and CURVE HIGH none of the channels happened to be clipped?

 

STAGE TWO

 

I don't recognise the software you are using there, so I can't help with that at all, I'm afraid.

The software doesn't matter. Imagine it's Photoshop and describe what you do to each colour channel once you've tranfered the TIFF into Photoshop.

1) Can you say more about adjusting each RGB channel?

2) Are you simply setting the black/white point for each channel or are you adjusting gamma or anything else?

3)Do you click a mid grey?

 

Pete

[andybarton]

STAGE ONE

Which sliders are you referring to here? Just the curve low/curve high sliders or the colour sliders or all the sliders?

 

The one in the shot above. I have shown you the settings in the screen shot.

 

 

 

So

1) You dial the black point and white point down to zero.

2)You don't right click any mid grey.

3)One could adjust the individual colour sliders to avoid clipping a channel, but you didn't need to in this example because when you adjusted CURVE LOW and CURVE HIGH none of the channels happened to be clipped?

 

1 - Yes

2 - No

3 - Correct

 

STAGE TWO

 

 

The software doesn't matter. Imagine it's Photoshop and describe what you do to each colour channel once you've tranfered the TIFF into Photoshop.

1) Can you say more about adjusting each RGB channel?

2) Are you simply setting the black/white point for each channel or are you adjusting gamma or anything else?

3)Do you click a mid grey?

 

Pete

 

1. I have shown you what I do in the screen shot

2. I don't adjust gamma

3. I don't click mid-grey

 

I think that you are seriously over-complicating things by clicking mid-greys (which you cannot be sure are mid grey unless you do fancy stuff with Threshold layers etc in PS) in both Vuescan and PS

 

Do you have any PS books to help you?

[biglouis]

Advertisement (gone after registration)

Andy

 

That is very helpful, to see how others use Vuescan and the post process. I am taking a similar approach although I read my TIFF into LR and then do a lot of my curves and colour adjustments there. I particularly find the black point slider useful for putting some body into otherwise anemic scans. Anyway, thanks again for showing that.

 

Louis

[Stealth3kpl]

Yes, thanks for taking the time to show us your routine Andy. I'd be interested to see your, and other's, colour manipulation techniques in Photoshop-type software. I don't have PS myself but would purchase it in a breath if I felt it had more useful tools than Capture NX2. I'm interested to see the book Louis referred to, and to see your Lightroom routine Louis.

My routine is at the start of this thread.

I chanced upon this earlier this evening but haven't had chance to read it yet. It looks like it might be useful.

 

I'm possibly wrong, but in my view a lot of the scans you've shown in these several threads have been really good.

 

I'm often quite happy with my scans but on each film, for what ever reason, there are usually some frames with which I achieve fairly poor results. There are more ways than one to skin a cat and it's interesting to practice all these methods, one of which may work for a given frame where another may fail.

Ultimately, I'd like to find a sure fire way to approach the magic that lays within those Imacon files without the expenditure!

The key lays in the manipulation of each of those R,G and B channels. I'm reluctant to fork out for scott kelby's book as I suspect it might not address rearranging the R,Gand B channels' relationship but pertains more to digital files from a camera which have a better controlled relationship. Tell me I'm wrong here and I won't regret buying the book.

Pete

[Stealth3kpl]

Well, try as I might I can't get results that compare with with simplicity of the system I'm using at the moment - but at least I know and I can get on with life! From the beginning I learnt that everyone has a slighly differing system settling on what works best for them. I know that scanning for RAW files in negative form has drawn some flak but it gives me a full unclipped histogram, and more control over the dynamic range and contrast than the other methods I've tried, and the conversion to positive takes a matter of seconds with saved conversion settings. The conversion shown at the start of this thread is actually only 4 steps

1) Inversion (a saved setting involving an inversion and this gamma2.2 boost).

2) Setting white and black points as Andy does at his 3rd picture here.

3) Contrast curve

4) Mid grey click if possible which seems to align the colour channels better through the range between black and white. I find it makes a vast improvement to reds often, making them less magenta.

 

For simplicity I'll batch scan using EpsonScan with this method with "No Color Correction" ticked here. I'll have to do some comparisons of linear raw scans from vuescan and from EpsonScan to see which will give me less noise and better overall colour.

 

Incidently, I'm finding the new Ektar 100 very easy to scan and convert:

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

 

 

 

 

This is Fuji 400H

 

 

 

 

 

Pete

[andybarton]

I can't stand Ektar (I have said this many times before), and when that is the only colour film that Kodak make, I will stop using colour film altogether. I keep threatening to stop using colour neg film and I need to buy some chemistry so that I can process it myself to avoid the Jessops monkeys.

 

One thing you must never do when scanning, if you want your colour balance etc to be correct, is to include any of the blank film around the frame. That will skew what your software reads and make it much harder to get a decent scan.

 

BTW, isn't Capture NX2 designed to work with Nikon digital cameras? I have a D700 and was disgusted that Nikon expected me to buy the software that they recommend to use with it.

[Stealth3kpl]

Yes NX2 is a powerful image processing tool to get the best out of your NEFs but the tools can be used on JPEG and TIFF too.

Pete

[andybarton]

How do you establish where your "mid-grey" point is to click on?

[Steve Ash]

While I find the results very impressive there is one thing that is in the way in my point of view.

 

The reason to shoot film is to do less pp as possible and to get the special look of film. The optimum scanning result in this respect would be an exact copy of what the film is delivering. I want to keep the character of the film and want to be consistent. Any kind of color correction and white balancing etc contradicts this somehow.

 

That is what I love with slides. I calibrate my scanner with a target and after that I simply get what is there.

 

Regards

Steve

[Stealth3kpl]

I use trial and error. Photoshop users might like to have a look at this.

I don't have photoshop so I click around on things that look grey. The result is usually better than if I had not. For example, the first image here is no grey clicked. The second is with grey clicked on a shadowy part of the grey jumper.

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

 

Pete

[Stealth3kpl]

While I find the results very impressive there is one thing that is in the way in my point of view.

 

The reason to shoot film is to do less pp as possible and to get the special look of film.

Regards

Steve

 

I agree Steve. This is why I've dropped my use of ColorPerfect and have removed processing of the image by Vuescan. I try to alter the contrast as little as possible. In the above example I didn't apply any contrast change at all after finding the black and white points of each colour channel.

I'd be interested to hear any method of retaining the inherent look of film.

Pete

[Doc Henry]

I watched with great interest this thread.

Without wishing to offend anyone here, I think the most simple, is not a straight line?

I scan with software (as fast and simple as possible), which gives me good results without correction.Then I look if the image is like what I see or what I saw

If it does not ,a little correction in LR good according to me (PS is too expensive)

Obviously, we need to start taking the picture with his camera the best possible because we can correct minor defects but not a failed photo

Great work Pete

Best

Henry

[plasticman]

I'm not sure whether there's a 'final' conclusion to the thread yet - but I still find that my best results come from the ColorPerfect workflow: that is to say, a totally unprocessed 16bit negative linear scan with (hopefully) no clipping at either end of the scale, on which I then use ColorPerfect to do the conversion followed by a minimal amount of post-processing (NOTE: I never use the built-in film profiles).

 

I may tweak the image slightly afterwards using curves in PhotoShop. But essentially the process is almost totally 'automated'.

 

For about 80% of the images which I choose to scan at high resolution this gives me results that I'm really pleased with. Some of the remainder can be tweaked to give acceptable results - but as Doc Henry says, if the original image is not good, then the final results will never be great.

 

Pete - those images look good to me, but if you still have a vague sense that the process isn't optimal, you're gonna have to take a week off work, and go through all the possible alternatives on just one image, until you think you've finally arrived at scanning nirvana. Report back here when you're done.

[Stealth3kpl]

On this particular frame I'm happier with the levels version which has the colour of the coat better (bright red). The ColorPerfect version is a bit too magenta. I'm sure this could be rectified in ColorPerfect but I don't know how. I wonder if the skin tones look better with the ColorPerfect version?

I do find that some frames process better in ColourPerfect, and others with the levels technique.

Pete

 

LEVELS

 

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

 

COLORPERFECT

 

 

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[Stealth3kpl]

OK last one. Again, with the levels technique the colour of the Borrowdale Green slate is more accurate than the ColorPerfect version which appears a little brown. I won't post any more as I'm sure people have had enough. I hope others have found it interesting.

Pete

 

 

LEVELS

 

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

 

 

 

 

COLORPERFECT

 

 

[andybarton]

What is the "levels" technique?