Getting Autochrome Look

[Mike Panczenko]

A director and I are talking about how to acheive an autochome look, preferably in camera. One thought was to shoot on black and white, and create a behind the lens filter with the three colors suspended in it, and place an identical one in the telecine gate. Would this work? What would be the best way to do it? Thanks.

[David Mullen ASC]

Autochromes were popular during the Pictoral Movement in photography, so it's hard to separate the technical process from the aesthetics, which often liked soft lenses, shallow focus, etc. to look more painterly. That's why some Autochromes are sharper than others, plus the fact that it was a rather homemade process.

 

Some have a very warm tone overall while others are more neutral. Usually the blacks were not too black and often the grain was golfball size. Super-8 is probably TOO low-rez to create this look since we're talking about 8x10 glass plates often (I believe) for the original process, but grainy 16mm or 35mm would be one method combined with some desaturation in post, maybe some milkiness of the blacks. I might try something like Fuji F-400T pushed two stops with a Fog Filter on the lens to create the softer Autochrome look (and the look of older optics which flared more); if you want a sharper look, I'd use flashing or UltraCon filters to soften the contrast instead of a Fog Filter.

[David Mullen ASC]

Lost my original post by editing it instead of replying to it. Here is it recreated:

 

 

 

 

 

The last photo is a blow-up to show the grain but I'm afraid my flatbed scan does not have enough resolution to show it properly.

 

The look is like faded Kodachrome to my eye, sometimes very grainy, sometimes less so. I don't think the filter over the camera and telecine would work. It would have to be some sort of microfine RGB pattern and it would have to be lined up on the positive in the telecine (meaning you need to shoot b&w reversal or make a print) to match the line-up in the original. It could work if you shot sequentially three b&w images, one filtered red, then green, and then blue -- and recombined them in post, but it would have to be of static images.

[Mike Panczenko]

My worry is that the digital recreation of autocrhome would be too exagerated, not subtle enough, etc. How would doing the in-camera/in-telecine filter with further digital tweaking look? The director wants a look that looks like autochrome, not digital trying to look like autochrome. What do you think of flashing, soft lens, fuji, pushing, etc., combined with the filter. Also, in the end, would this filter, which would basically be made of materials in the random autochrome array, acting as the potato starch- I don't know if it would actually be potato starch, but something to act as it- would it even give that pointilism effect that autochome gave, considering the sharpness of modern film stocks? Basically, as you know, because of the random array of the starch flakes, and how it utilized the additive process, a pointilism, "dotted" effect was had. But, would modern film, being so crisp, have the same effect? Would this filter, which, if you can't tell, I'm kind of in love with :lol: , actually contribute anything?

[David Mullen ASC]

I don't think you can use a RGB patterned filter to create the Autochrome effect. I'm not sure exactly how the process worked but I believe from what I've read that you basically were viewing the original glass plate almost like a transparency after development -- it wasn't a neg-to-pos system.

 

So whatever the random arrangement of potato starch dyed grains were, it was built-in to the original, not carried over and applied like a filter over a later print. So your idea simply won't work. You can't shoot with one filter and then just somehow use it in a telecine bay over the piece of b&w film and end up with a full color image (and Autochromes were a full color process.)

 

You look at enlargements of Autochromes and basically what you see are very large colored grains, which is similar to what you'd get in modern color film if you could enhance the graininess, either by using fast film push-processed enough or a small format like Super-8.

 

You could also try the old Ektachrome VNF films pushed before they completely are unavailable to shoot with.

 

Just study some Autochromes and analyze the look (color saturation, grain structure, black level, sharpness, etc.), list its attributes, and then recreate them as closely as you can using photo-chemical and digital methods. I don't understand -- why deny yourself any digital technology IF it helps you achieve the look you want? Unless this project is for direct print only and you can't afford any digital post anyway.

 

You get as close as you can photo-chemically and just tweak it in post digitally to finalize the effect. We're not talking about taking normal footage and recreating the look entirely in post.

 

By the way, if you look at Coppola's "Dracula" there's some hand-cranked footage of Dracula on the streets of London. Originally shot on 5245, it looked too good even though they used an old hand-cranked camera, so they duped it to make it super grainy and it took on an Autochrome look. I don't know the stocks used in duping but they may have done YCM separations from the color neg onto particularly grainy b&w film and then recombined them onto a color IN. That's another method to consider.

 

There were some milky grainy shots in Wong Kar Wei's "Happy Together" that looked a little like an Autochrome; it was underexposed footage pushed two or three stops I believe, so there was a lot of build-up in the base fog density.

 

Whatever technique you use, it has to be combined with the visual aesthetic of Autochromes and turn-of-the-century Pictoral photographers or else it will just look like grainy photography of modern scenes. You have to apply a painterly approach to the compositions and lighting and focus, etc. If you run around and shoot handheld, it will look more like some Super-8 documentary from the 1960's than an Autochrome from the 1910's.

[Mike Panczenko]

Thanks for taking the time to respond in such detail, David. Yeah, what you're saying is probably the way to go. I'm liking that Fuji, pushed, flashed, etc. idea you were talking about in the first one. Just for the record, though, here's how autochrome worked- you said you didn't know exactly and I think it's really interesting. Potato starches dyed RGB were placed over a black and white emulsion. The light would be filtered by the corresponding starches, and the grain under it would be exposed. For example:

 

Here's the random starch array:

 

RBRRBGGBG

BRGBRGBRR

GBRGBRRGR

GRBRGRBRB

 

In one were photographing a totally blue room, for instance, this is what it would look like. The grains behind the red and green starch flakes would be unexposed, and remain black, obviously, but the blue ones were be exposed. So, sticking with the above array, here's what it would look like:

 

-B--B--B-

B--B--B--

-B--B-----

--B---B-B

 

This is part of the reason it has a pointilism effect. The filter I was thinking of would essentially take the place of the starch grains, so, I guess, theoretically it would work, but practically, after what you told me, I highly doubt what good it would do. I still think I want to test it, but, realistically, the pushed, flashed Fuji seems like it might be the way to go. David, I really appreciate you taking the time to respond to my queries. Thanks! I just learned how autochrome works, so I'm kind of excited to share with other film buffs- hopefully my explanation made sense! But, thanks again, David.

[David Mullen ASC]

But from your description it was a little like making a large glass plate transparency with the silver grain after development "stuck" to a corresponding dye potato starch grain. I don't think an optical filter over the camera and then over the positive in post would do the same thing because you'd never get the colored spot on the filter to ever line up again with the developed grain of the positive.

[Mike Panczenko]

I don't know if they really stuck, or just remained aligned, but what the director and I spoke about is to create, and this may not be very practical, totally, down to the starch level, identical filters for the camera and the telecine suite, to preserve the alignmenment of the grains and the starch flakes. The thing is, I don't know how much it would cost to make, and, also, if the effect would even get us any closer to the original autochrome look. So, if they were perfectly lined up in camera and in telecine, do you think that the look would get us anywhere near what autochrome looked like? I know it would need to be augmented digitally in post, but would it be worth doing to get a more realistic and less digital look, or would it effectively do nothing?

[Guest federico drasal]

hi,

did you visit www.autochrome.com? there you can get some info.

[John Sprung]

Lynn Trimble had some genuine autochromes, which he showed in his color tech class back in the early 1970's. I remember one of a flower garden that was shot in 1904, a 4x5 glass plate -- or maybe it was 3x4. It was quite sharp and deep focus, which apparantly wasn't the usual style at the time. And the color even after 70 years was amazing.

 

The way it worked is that the dyed starch grains were suspended in a sort of varnish-like clear coating, and painted onto the glass plate. Then a B&W emulsion was coated over the starch grain filter layer. The finished plate was exposed through the glass plate, then developed as reversal. Given the technology of the day, coating the varnish-filter layer over the emulsion would have sealed it, and that would have prevented any chemicals from reaching the emulsion to develop it. So, they had to shoot thru the glass.

 

The really neat thing about this process is that every color is viewed using exactly the same filter that was used to shoot it. It looks very grainy because the starch particles were a lot larger even than the silver halide grains they had 100 years ago.

 

If you could coat your own B&W emulsion, the rest of it is fairly straightforward, and it should be possible to make your own autochrome plates.

 

 

 

-- J.S.

[David Mullen ASC]

>So, if they were perfectly lined up in camera and in telecine

 

I don't see how that is possible on a granular level. Besides, there's a big difference between colored specs on glass in contact with a b&w emulsion like with an Autochrome and in front of a lens like with a camera filter.

[Phil Rhodes]

Hi,

 

I find all of this stuff absolutely gorgeous. It's quite reminiscent of the smoky post-battle stuff in Cold Mountain - I was playing with some of it today. Very grainy, lowish-con filmstock, smoke lowering contrast further - you're already halfway there.

 

Phil

[Boone Hudgins]

Increasing the noise on each individual channel in Photoshop gives a similar look. The noise is too sharp. The grain on the pictures I've been looking at is softer. I gave each channel a random amount of noise.

 

Maybe an effect like the old Technicolor dye transfer printing, but adding a healthy dose of grain, would look similar.

[Mike Panczenko]

I was thinking to do it as a behind the lens filter, but, yeah, it doesn't seem practical. Not only would it be really expensive and almost impossible to create the identical filters, but then registration problems in the camera and telecine would also mess things up. I guess digital is the way to go- I just hope it looks natural and not too toyed with. Thanks for all the responses guys!