Friday, November 20, 2009

Cross Processing Myths RIP

Cross Processing Myths RIP

The 2 myths I will be talking about is that cross processing E6 slide film in standard C41 colour negative gives unpredictable hence random results. Hence the popularity with lomography crowd.  Before you stop me I know plenty of people that take wonderful dreamy shots with lomo and other toy camera but none of them rely on random luck it's skill and artistic ability. Second lesser myth is that it is better to over expose expired slide film when you cross process.

First off in a past life I have a background in imaging and high speed microscope scanners including some patents that I'm proud of but most of what I am going to be talking about is based on observation and intuition. So bear with me while I go through some technical details:

Colour Slide Versus  Colour Negative film

Slide film in general has a much smaller latitude when compared to negative film this means that your exposure settings for slide film has to be more accurate then with negative film or your highlights will be blown out (sort of sound like digital doesn't it) on the other hand if you under expose the darks will be one shade of black.  With todays colour film a couple of stops under or over exposure doesn't make much difference it will look pretty much the same. Film records light in a nonlinear way so instead of clipping like in digital the highlights and shadows are compressed in a natural eye response type of way.

So why use slide film because if you get it right the resulting image has a much greater dynamic range (the difference in film density between the lightest to darkest part of the image), blues go from the palest lightest shade to inky almost black blues, the same for the greens and reds. A 8x10" colour slide on a light table is a glorious thing.  Maybe paradoxically it is more difficult to scan when compared to the more compressed lower contrast ratio colour negative. This is true even when E6 type slide film is processed in c41 (standard developing for colour negative films)

If you want better technical reading go


Dynamic Range By Bob Atkins

and here:

Film Contrast by Perry Sprawls, Ph.D.

Film is not like fine wine it's designed to be used when it is still young. As film ages it gets bombarded with cosmic rays and maybe more important the layers oxidize and chemicals diffuse and mix with each other sort of like the ripples in hundred year old glass windows. Refrigeration  or freezer will extend the life but only by so much.  As the film ages the unexposed film slowly fogs meaning black  is no longer black but shades of grey. The colours in the dark areas become muddy and the grains of silver become more obvious and dyes start to bleed. (This paragraph is conjecture and observation on my part.) Astronomer super charge film sensitivity to low light by annealing it with special gas mixtures).

The lower the ISO the slower the film ages (this shouldn't surprise anyone). What sort of surprised me was that expired film  seemed to be fairly constant in exposure (ISO), with only the latitude and dynamic range  decreased by the level of fogging.

Now what does this all mean when it comes to exposure. Conventional wisdom seems to be to over expose to compensate for the fogging. I have tried this by a half a stop for 15 year old slide  film and all I did was sacrifice the highlights for a bit better shadows, I also tried over exposing by 1 stop and pulling the development by one stop, this was even worse, colour film in c41 seems to push better then it pulls. My best results is to shoot at box speed and meter the exposure for the subject of importance. It also saves me the time  money at the lab as most lab charge extra for pushing, pulling and cross processing if they can even really do the pulling correctly.

So shoot at box speed and meter for the subject, it's old film what do you expect new film? I'm only talking about colour slide film it's different for B&W and negative film.

Hotel Victoria
Lab Scanned Early Work

The big myth that cross processing gives random inconsistent results. Except for the occasional very rare damaged roll. I find the result from each type of film is repeatable and consistent.

When I first tried cross processing I didn't really know why the results  appeared to be inconsistent. I believed all the myths. First I thought it was how it was developed, then how it was exposed. A couple of years ago I started to examine the colour curves. I couldn't make any sense out of them so I thought it was cross talk between the RGB channels red getting into the green channel, blue into red, ... sort of like the current lomo fad (ducks) using redshft film to shift the colours into the yellow orange red spectrum. So I tied to build a colour profile that would correct this cross talk, no such luck. it wasn't until a lab scanned the cross processed  as colour positives effectively supplying me with colour negatives (excuse the brain twisting double negatives).

Examining the colour negatives curves it was plain to see that the rgb curves had a much different shape then normal and the centre of symmetry was offset and different shaped for each of the rgb channels when compared to a normal colour negative curve.

What does this mean:

That colour is dependent on and intertwined with intensity.
If you remove this dependency you get consistent slide to slide, roll to roll results.

If you let a lab scan your film you will get highly inconsistent results that depend how their automated scanner treats a cross processed image when expecting a normal colour negative..

If you want a consistent results you have to adjust every image at either at the scanning stage and/or in post.  Usually it's enough to correct white balance at 1/3, 2/3 and %100 white.

 The goal is not to get perfect white balance, if that was the target then why bother cross processing.

Whats with the ABCD in the above polytch using the same negative Kodak 64T expired 1997.

A is lab scan not too bad except for the blown highlights and lack of shadow detail.

B is scanning as a colour negative, good details an colours but hard to get the highlights as tungsten balance film has a tendency to have a lot of red noise in the highlights, this can be used to good effect  to give a gold look.


Mid Summer Dream
Kodak Ektachrome tungsten balanced film is my favourite film to crossprocess

C Scanned to give the expected digital cross processed look.

D is scanned as a colour positive and then colour reversed in post. This gives you the most control at the expense of extra time.
Note: Scans B,C and D were done on my Epson V500 scanner.

I cross process because I have cheap expired slide film and I like the surreal colour shifts even if it is at the expense of some colour bleed and lose of details. I really like Kodak Kodak Ektachrome tungsten balanced film is my favourite film to crossprocess because I can bring out golden highlights.

Lab Versus Y500
This expired  Fuji Provia 100f is not too bad it still has a good DMAX - DMIN. But of course with all expired film it has lost some shadow detail. By now  you should be able to guess which scan is the v500.

Springtime At The Necropolis
Cross processed Fuji Velvia 50 scanned as positive and then reversed by colour negation  in post. Velvia 50 is very difficult to scan directly as a negative because the DMAX-DMIN (dynamic range) of the film is so high.

If you disagree with anything I have said then there is nothing to stop you from doing it your own way:)

If you are bored you can have a look at a slide show  of my cross processed work on flickr

Thursday, August 13, 2009

Faux Redscale

Faux Redshift

Redscale is done by putting the roll of colour film into the canister with topside and backside reversed. Originally a DYI project Lomo is now selling the reverse film at a premium and Rolli is now manufacturing films with the filter layers reversed also at a premium.

The gel filters are reversed ordered hence the CMY channels are mixed to an approximation

C' = C + Y
M'= M + Y
Y'= .2Y

This example was done by

1) Change RGB positive to colour negative (don't use the grey scale negative).

2) Remix the RGB channels (I don't know it might be better mixing CMY channels

R' = .2R (Y on negative, Blue on positive)
G' = (G + R)/2 (M on negative, Green on positive)
B' = (B + R)/2 (C on negative, Red on positive)

3)Reverse the colour negative back to a positive (at this step you have a chance to change the colour gamma to bring out the yellows and greens)

3)Adjust contrast and brightness curves

If you are using film you could start with a colour negative and skip step 1.

This is my first attempt at faux redscale so it needs some tuning but overall I find real redscale has a creative yawn factor of 8 out of 10 so faux isn't going to be any better. Give me IR film with the filter built in now that would be creative!

Original Photo

Rain Rain Go Away

Thursday, July 02, 2009

Black & White Development In One Shot D76

Sun Bathing In The Don
Mamiya C220; Film Tmax 400; shot at ISO 400; 12 minutes in the developer

I have been doing all my B&W developing with Kodak D76.

I keep my d76 in a mason jar in powder form and mix each batch fresh one shot use. Conventional wisdom is that the powder might settle into it's components but it doesn't (don't use the last table spoon). Unconventional wisdom from people that have been doing it for years. I also develop in 1:3 dilution so development time isn't so critical and I get good shadow detail without blowing highlights. Just beware of chemical dust when measuring out the powder.

First time you empty the package into the mason jar stir and shake and roll the powder in the jar then put it on the shelf, no need to shake each time you use .

Disclaimer: I am not an expert and I scan my film rather then print in the darkroom.


See massive development table for times don't forget to adjust for temperature.

I use 1.5 level teaspoons (7.5ml) d76 fill to 420ml with water (distilled if you want to be fussy) Stir well let sit for a few minutes and stir again until all powder is completely dissolved. This makes up an 1:3 dilution stock solution. Be careful not to breath the dust from the powder. Only use it for one roll of film.

  • 2 minute water presoak
  • xx minute development {shake well for the first 30 seconds (instead of shaking invert if possible) tap once on a hard surface to remove bubbles, thereafter every 3 minutes gently shake for a few seconds}
  • rinse with water stop 4 times
  • yy minute fix
  • rinse with water 4 times and soak in clear water for 3 minutes
  • dip in water a bit of rubbing alcohol and a few drops of hypo clear
  • Go here for final rinse instructions :
  • hang to dry for at least 2 hours (the leader should feel dry and not tacky) I find weighting the bottom of the film keeps from coming out curled

My 2 favourite general purpose films are Kodak TMAX 100/400 and Fuji Neopan. So far this method has given consistent results with all films I have used.

Kodak TMAX needs loads of fixing time. You can drop a small piece of film that hasn't been in the developer and watch the fixer clear the film, Multiply the time to clear by 2 and that should be safe.

Mamiya C220; Film Fuji Neopan 400; shot at ISO 800; 20 minutes in the developer

Please go here for more details and my notes on developing B and W

Tuesday, May 12, 2009

Using The V500 Scanner and HDR Techniques To Get The Most Out Of The Negative

Example of a technique for scanning high dynamic range films. Plus a digital darkroom dodge and burn

I took this shot using my mamiya c220 with a 180mm lens. Film was TMAX 400 shot at 1/2 stop faster. As usual I screwed up the loading of the film into the developing spools so there are a few handling flaws.

- Developed for 12 minutes in 1:3 diluted D76.
- Scan 1 scanned to make sure highlight detail was retained
- Scan 2 was for shadow details both scans were 16 bit tiffs
- Used a program called enfuse and enfuse GUI to automatically blend both scans together. Enfuse by passes the HDR step going directly to the tone mapping but with more subtlety, less noise and artifacts.

It's hard to tell in small sized but the enfused image has rather harsh contrast in the area of the gravel.

Final step I blended Scan 1 and the enfused image together using a hand painted selective mask. With a bit more care I could tailor multiple selection masks to get the result exactly how I like it.

History of 158 Sterling Rd

Another Example using HDR Tools

Image on the left is 14 bit scanned with some brightness curve adjustments in Lightroom. Normally I would do some dodge and burn.

The image on the right is the same scan plus an additional 2 scans at different scanner exposure levels. The 3 scans were combined and then tone mapped using HDR (high dynamic range) software.

I used Fuji Neopan ISO 400 film if I had used ISO 100 film the dynamic range could have been increased significantly. I think this does demonstrate how much information is contained within a film negative. Scanners have both a programmable offset and gain so there is actually more information extracted from the negative when you scan 3 times rather then manipulate a single scan.

As tone mapping increases local contrast at the expense of flattening out the overall image blending together portions of both images might result in the best of both worlds and focus the composition a bit more. Actually took the HDR into Lightroom to adjust curves and the results are better then above.

There is a trick (an extra 2 steps) to do the same thing in colour without destroying the colours I'm not sure it's worth the trouble but ask me if you are interested.

Thursday, May 07, 2009

Testing Epson V500 Scan of a 35mm Zeiss Contax IIIa Colour Negative

Unlike my Canon 4400f the Epson V500 scanner is in focus, also the slide holder are much better and easier to use. It also works on medium format. The new scanner will allow me to do more B&W in the future. This shot is nothing special but the I choose it as it needed a lot of dodging and burning
Notice the camera information in the additional information.
Yes the v500 is excellent price performance wise. Unfortunately the scanner software does not provide exif information. Unfortunately the program I use to add EXIF information is no longer available.

I did a test of my v500 on a resolution target and found the focus was good for about 800 lines per inch from 0 to .5mm plus the film thickness from the surface of the glass, going from 2400 to 4800 dpi only made a minor improvement differences, kind of explains why you have to go to medium format to get a worth while size print. Also explains why a special film holder is not going to buy you much  except with curly negatives considering the v500 optics and it's LED illumination. Other have measured the V500 resolution as less than 1600 dpi but there is still merit to over sampling.

This confirmed the results of wet scanning with the film right against the glass

My Current Work Flow
  1. Clean everything with antistatic cloth 
  2. Set scanner to pro mode, 3600 DPI for 35mm and 2400 DPI for MF in 48 bit tiff output.  
  3. Insert negatives If they are too curly use anti-glare glass instead of plastic holders. 
  4. Blow any dust off with a rocket blower (If you are rich use canned air). 
  5. Preview
  6. Select all negatives and turn off sharpening and hit auto curve button to see what the epson software things it should look like 
  7. For each negative adjust curve so there is no clipping using the droppers to set white and black and gamma and colour balance. The epson software is so primitive here, that my main concern that there is no highlight clipping or shadow blocking. Do this for each colour as the over all may look good but one channel might not. The point here is not to get a perfect scan but to get all the information out of the negative. 
  8. Scan
  9. Optional If you need more shadow or highlight details then do an additional scan to expand the contrast in the highlight or dark areas blend the images back together in post. I don't have to do this very often. In post I use Lightroom to correct the image and remove any dust.
Some help with settings for the epson scan software:
epson 1.png
epson 2.png

Important because you have turned sharpening off you need 2 levels of sharpening one for the scan and one just before final output. Fortunately lightroom has pretty good control here. The first level of sharpening is important and should be matched to the grain size and the scanning DPI it is not there to sharpen edges.

For difficult colour balance cases check out this note of mine.
For scanning cross processed films
For scanning high dynamic range (mostly B&W) go New herehere and maybe here.
For curly negatives the Lomography Digtaliza works well see the bottom of my blog entry

A Two-Pass Approach to Sharpening

This is a good reference on resolution and scanning: