What's new

A Few Words About A few words about...™ Dawson City: Frozen Time -- in Blu-ray (1 Viewer)

Robert Harris

Archivist
Reviewer
Senior HTF Member
Joined
Feb 8, 1999
Messages
18,271
Real Name
Robert Harris
Having been aware of the Dawson City Nitrate find for forty years or so, I really thought that I knew what to expect of a documentary on the subject.

Old gold mining territory in Canada. Hundreds of reels of 35mm nitrate film, left at the end of the film distribution trail, much like Australia and those environs.

Distributors didn't want them back, several years after they first opened.

Dawson City: Frozen Time is a two hour documentary on the subject.

Simple.

Well. Not exactly.

While it was nice seeing from friendly faces among the archivists involved, the film really doesn't start until past the thirty minute point, and then it's sometimes only tangentially involved with the film elements themselves.

The first half hour is a bit of a history lesson.

The gold rush.

A specific area of Canada.

Late 19th century.

Early settlers, native Canadians.

And then it becomes a bit of a tone poem.

A bit of a fever dream.

At two hours, it's also very long. My personal feeling is that it could have easily told the same story in 90 minutes or less.

But then it wouldn't be a tone poem, or a fever dream.

It's all very odd.

And not unenjoyable.

Once you realize where you are and where you're being taken.

If you're picking up this disc for a dry background on the nitrate find itself, there are easier ways of getting the information, albeit you won't see hundreds of (very) short excerpts from the film, most all with the very specific look and texture that you'll realize is a part of those film, after being encapsulated in permafrost for forty years of so.

Hundreds of films had already been burned, or dumped into the sea.

But these lucky survivors, left in an empty swimming pool that had been covered over, are an extraordinary time capsule, of the films that the good folks of Dawson City saw at their theaters (that seems to burn down almost annually).

Think of Dawson City as Deadwood. A bit further north. Possibly a bit less gunplay. But the same muddy streets, and tinder-ready buildings.

For those interested in film history, and the problems of film preservation, this should be required viewing.

But don't get ready for a Ken Burns type affair.

Bill Morrison's work is very different.

With a very specific personality all its own.

And once you realize that, it's all good.

Image - n/a

Audio - n/a

4k Up-rez - n/a

Pass / Fail - Pass

Recommended

RAH
 

bugsy-pal

Stunt Coordinator
Joined
Jan 28, 2010
Messages
223
Real Name
Paul
Thanks for your impressions. I must say, the screenshots on DVDBeaver that I saw recently fascinated me. I must see this!
 

bigshot

Senior HTF Member
Joined
Jan 30, 2008
Messages
2,933
Real Name
Stephen
I screened this last week. It was fascinating and the research and film clips were incredible. The music was like bland wallpaper, but I suppose they wanted you to focus on the screen instead of getting lured into following the storylines of the clips. There must be a music academy somewhere where they give musicians quaaludes so they can provide soundtracks for documentaries.
 

Robert Harris

Archivist
Reviewer
Senior HTF Member
Joined
Feb 8, 1999
Messages
18,271
Real Name
Robert Harris
I screened this last week. It was fascinating and the research and film clips were incredible. The music was like bland wallpaper, but I suppose they wanted you to focus on the screen instead of getting lured into following the storylines of the clips. There must be a music academy somewhere where they give musicians quaaludes so they can provide soundtracks for documentaries.

I’m afraid the music was part of the fever dream.
 

Alberto_D

BANNED
Joined
Jan 24, 2006
Messages
215
I always wanted to know why they (archives/studios) don't create a film archive in the most cold weather habitable places on Earth. The cold preserve films and reduces a lot the storage costs. Maybe a archive on Alaska could keep original nitrates for more 100 years..
 

Robert Harris

Archivist
Reviewer
Senior HTF Member
Joined
Feb 8, 1999
Messages
18,271
Real Name
Robert Harris
I always wanted to know why they (archives/studios) don't create a film archive in the most cold weather habitable places on Earth. The cold preserve films and reduces a lot the storage costs. Maybe a archive on Alaska could keep original nitrates for more 100 years..

One would presume, access.
 

Alberto_D

BANNED
Joined
Jan 24, 2006
Messages
215
I understand you point Mr Harris. But what about north of Canada? Not near but quite arround civilization.
The mot cold places could be used as a emergency alternative, since in many countries a lot films are still lost due deterioration. Even acetate, like the complete version of The Alamo, that now is faded in colors since the studio didn't listened to your recomendation at the time, of put in cold storage.

By the way, If you allow me, I have a idea to restore the color of very faded films, like some extreme fadding in the cutted scenes from the It's a Mad Mad Mad Mad World.

One would presume, access.
 

Robert Harris

Archivist
Reviewer
Senior HTF Member
Joined
Feb 8, 1999
Messages
18,271
Real Name
Robert Harris
I understand you point Mr Harris. But what about north of Canada? Not near but quite arround civilization.
The mot cold places could be used as a emergency alternative, since in many countries a lot films are still lost due deterioration. Even acetate, like the complete version of The Alamo, that now is faded in colors since the studio didn't listened to your recomendation at the time, of put in cold storage.

By the way, If you allow me, I have a idea to restore the color of very faded films, like some extreme fadding in the cutted scenes from the It's a Mad Mad Mad Mad World.

One would still need security and access. Elements move with some regularity.

Please do share your ideas for reclaiming faded elements.
 

Alberto_D

BANNED
Joined
Jan 24, 2006
Messages
215
My pleasure.

As we know extreme faded films lost information, dynamic range, and can't be restored even with digital filter, since most of the image information it's gone from the faded layer (usually blue).
Some misinformed people could think a "simple alternative" like : "Ooh, why not just colorize it Colorize it ?" But colorization needs a right B&W image, and the loss of a color layer would have a different contrast in the overal image, specialy in the sky. Besides colorization it's still not perfect, despite got better today.

There are still the alternative of "transplant" the missing layer, like your friends did with Disney'sd Pollyana, combining color layers of the camera negative and the protection master (which missed the green layer), but it require a surviving color layer from other source.

I propose the recreation of the blue layer using the surviving the red and green layers as base and production still and other sources as reference. For this some algorithms created for colorization would be usefull to do what is called image segmentation (rotoscope each object, piuece on the scene). Today rotoscope evolved and it's used even to cut images with precision enough to 3D conversions. It's not colorization, but recreation of a missing color layer.

Now there is the difficult part, recreate the color gradients. Colorization it's based in the fact that most objects follow a color that m ost times it's very close to a color gradient (variation of saturation and hue values along a B&W gradient). It works fine in some cases and not very good in other (like skin color) but in this case we would be recreating just one element and still having the original two color layers, since rarely a film fades up to lost green layer to a incorigible degree. And fortunatelly, red and green gives most of the difficult variance to recreate.

I did a concept test in Photoshop once, getting a color image, destroying a small portion of it in the blue layers (making a hole), futher rotoscoping the destroyed area and adjusting gamma from red & green layer to fill it. For example, a purple shirt, if I fill the hole area with a copy of the red layer, and adjust gamma with very fine tuning until it get the right color.
The rotoscope, in theory, with proper tools (I don't have) could be very precise, since it's not like rotoscope a B&W mage, which tools work with B&W tonalities, as the tools could work with tones and saturation values & hue values, since two color layers already render a image with tone, saturation and hue differences, like two color technicolor.

The surving layers can be used as base also to estipuilate deviation of colors in a object, helping to get where the recreated color layer should deviate beyounf the color sprectrum aproach. It's difficult to explain in words. Let me try a example, a scene with someone in a set with light from two sources, and this would create a a variance in color and saturation different in one side than in other side of the character, like a white shirt near a candle. The portions near the candle would be more yellowed. In the red and green layer this portions would be brighter, and in the missing blue layers this portions would be darker, compared to the portions of the shirt where the candle's light is close. Analyzing mathematicaly such differences it's possible, in theory, to a better recreation ot the blue layer for such cases than just gamma adjust from segmented (rotoscoped) pieces. This would make the difference, specially for things like the darker and less satured area of beard in shaved faces.

In the case of The Alamo, it's even easier, since the Laserdisc transfer was made before the print fadded, and it would work as a good guide to the reconstruction of the blue layer, despite the limitated reolution. For It's a MAd Mad MAd MAd World, there are similar scenes, similar shot from sets and landscapes, that gives the righ color for reference.
I imagine that a faded film to the point of no recuvery with filter, can still have some informations, specially in fully exposed shots, that can find some color references. A very faded layer start to loss dynamic range, but some informations can be rescued as reference, despite the overal look be very uglly.

It would require investments to design a software, but it could also became a excellent tool for digital color gradding in modern films.
It would also work as a a tool to digital film restoration of technicolor films. I once send a similar concept to MTI (film restoration software company) but just as a restoration for damaged frames in technicolor strips, and they said it was interesting.

May I ask you a example of blue layer very faded, to the point of the it lost most range and canot looks fine even ange after digital filter. If I am right that some few dynamic range survived (not enough to a decent color restoration look by actual methods), perhaps it's possible, in theory, that just the missing dynamic to be recreated or that the few surving information can serves a a guide somehow.
.
 
Last edited:

Richard Gallagher

Reviewer
Senior HTF Member
Joined
Dec 9, 2001
Messages
4,275
Location
Fishkill, NY
Real Name
Rich Gallagher
Whenever I read about Dawson City I can't help but think about the promotion that Quakers Oats had in the fifties whereby people could send in a box top from Quaker cereals and receive a deed for one square inch of land near Dawson. The deeds were issued by the Klondike Big Inch Land Company, Inc. The land was seized for back taxes in 1965 and now is said to be part of a golf course.

Klondike_Big_Inch_Land_Promotion_Certificate.jpg
 

Robert Harris

Archivist
Reviewer
Senior HTF Member
Joined
Feb 8, 1999
Messages
18,271
Real Name
Robert Harris
Whenever I read about Dawson City I can't help but think about the promotion that Quakers Oats had in the fifties whereby people could send in a box top from Quaker cereals and receive a deed for one square inch of land near Dawson. The deeds were issued by the Klondike Big Inch Land Company, Inc. The land was seized for back taxes in 1965 and now is said to be part of a golf course.

Klondike_Big_Inch_Land_Promotion_Certificate.jpg

I had one of those. Acquire several, and one can have a nice vacation property.
 

Robert Harris

Archivist
Reviewer
Senior HTF Member
Joined
Feb 8, 1999
Messages
18,271
Real Name
Robert Harris
My pleasure.

As we know extreme faded films lost information, dynamic range, and can't be restored even with digital filter, since most of the image information it's gone from the faded layer (usually blue).
Some misinformed people could think a "simple alternative" like : "Ooh, why not just colorize it Colorize it ?" But colorization needs a right B&W image, and the loss of a color layer would have a different contrast in the overal image, specialy in the sky. Besides colorization it's still not perfect, despite got better today.

There are still the alternative of "transplant" the missing layer, like your friends did with Disney'sd Pollyana, combining color layers of the camera negative and the protection master (which missed the green layer), but it require a surviving color layer from other source.

I propose the recreation of the blue layer using the surviving the red and green layers as base and production still and other sources as reference. For this some algorithms created for colorization would be usefull to do what is called image segmentation (rotoscope each object, piuece on the scene). Today rotoscope evolved and it's used even to cut images with precision enough to 3D conversions. It's not colorization, but recreation of a missing color layer.

Now there is the difficult part, recreate the color gradients. Colorization it's based in the fact that most objects follow a color that m ost times it's very close to a color gradient (variation of saturation and hue values along a B&W gradient). It works fine in some cases and not very good in other (like skin color) but in this case we would be recreating just one element and still having the original two color layers, since rarely a film fades up to lost green layer to a incorigible degree. And fortunatelly, red and green gives most of the difficult variance to recreate.

I did a concept test in Photoshop once, getting a color image, destroying a small portion of it in the blue layers (making a hole), futher rotoscoping the destroyed area and adjusting gamma from red & green layer to fill it. For example, a purple shirt, if I fill the hole area with a copy of the red layer, and adjust gamma with very fine tuning until it get the right color.
The rotoscope, in theory, with proper tools (I don't have) could be very precise, since it's not like rotoscope a B&W mage, which tools work with B&W tonalities, as the tools could work with tones and saturation values & hue values, since two color layers already render a image with tone, saturation and hue differences, like two color technicolor.

The surving layers can be used as base also to estipuilate deviation of colors in a object, helping to get where the recreated color layer should deviate beyounf the color sprectrum aproach. It's difficult to explain in words. Let me try a example, a scene with someone in a set with light from two sources, and this would create a a variance in color and saturation different in one side than in other side of the character, like a white shirt near a candle. The portions near the candle would be more yellowed. In the red and green layer this portions would be brighter, and in the missing blue layers this portions would be darker, compared to the portions of the shirt where the candle's light is close. Analyzing mathematicaly such differences it's possible, in theory, to a better recreation ot the blue layer for such cases than just gamma adjust from segmented (rotoscoped) pieces. This would make the difference, specially for things like the darker and less satured area of beard in shaved faces.

In the case of The Alamo, it's even easier, since the Laserdisc transfer was made before the print fadded, and it would work as a good guide to the reconstruction of the blue layer, despite the limitated reolution. For It's a MAd Mad MAd MAd World, there are similar scenes, similar shot from sets and landscapes, that gives the righ color for reference.
I imagine that a faded film to the point of no recuvery with filter, can still have some informations, specially in fully exposed shots, that can find some color references. A very faded layer start to loss dynamic range, but some informations can be rescued as reference, despite the overal look be very uglly.

It would require investments to design a software, but it could also became a excellent tool for digital color gradding in modern films.
It would also work as a a tool to digital film restoration of technicolor films. I once send a similar concept to MTI (film restoration software company) but just as a restoration for damaged frames in technicolor strips, and they said it was interesting.

May I ask you a example of blue layer very faded, to the point of the it lost most range and canot looks fine even ange after digital filter. If I am right that some few dynamic range survived (not enough to a decent color restoration look by actual methods), perhaps it's possible, in theory, that just the missing dynamic to be recreated or that the few surving information can serves a a guide somehow.
.

I presume you’re referencing faded print, as opposed to negative.

Usually, all that survives is the magenta dye layer, which is only helpful in creating an acceptable monochromatic example. That was the methodology for the production of black & white prints for reissue and broadcast use in the days before color broadcast.

A b/w dupe neg was created from the magenta protection record.

For a moderately faded negative, Why not simplify, and create a histogram of a very similar shot, and force the affected shot to match the available histogram.

A Histo-Match.

The problem with getting into too much technology is cost. The most financially feasible position, is still to recombine a 480i video record with a b/w representation of the affected footage.

Presuming you have one.

The overriding problem is that few people truly care to go to any expense, as most viewers have minimal concern for quality, and things look fine on an iPhone.

You mentioned the use of stills to create color. Different emulsions, only allow a basic color representation. Consider Kodachrome vs a 1950s or ‘60s Eastman Color print. Useful toward colorization, and selection of colors.

With today’s digital tools, one can try to dig very deep, and attempt to pull up a color dye layer by its bootstraps. When performed with a color neg, grain rises to the surface, along with the Y information. One could then degrain, and regrain.

One could continue, and eventually arrive at a point of discussion, similar to fairies dancing on the head of a pin.

Only film is easier to test.

Of course, there were also experiments with color masks, that were ongoing in the analogue world. Some reasonably successful.
 

Alberto_D

BANNED
Joined
Jan 24, 2006
Messages
215
There is already technology of low cost image segmentation of all piece/objects in a scene, and it's used for film colorization and 3D conversions by Legend Films. The software track the subt changes along countorns during motion, after a reference frame with all the masks is made. Little erros are corrected, since the algorithm it's not perfect, by a operator. That's how they could colorize many B films and get profity with just a medium DVD sales.
A histogram match of similar scenes it's welcome, ideal in the case of It's a Mad Mad Mad World.

May I ask you email.
I have some color corrected images from Blu Ray screen captures I would like to send to you, but it's in the conventional way, forcing histogran, with extras color forcing in a multi tools aprouch, and not using any segmentation/recreation. I believe I did a better color correction, despite the Blu ray was a digitally restored version of a film. I don't want to post here cause people can think I'm being arrogant.


I imagine there is cases of negatives which the blue layer is lost, but the green can be forced a bit, since the blue fades first.
The video transfers can work as guide, since it don't require pin sharp detail, just a root a guide.

I presume you refer about Peter Kuran's method for color correction of faded films, made in film lab process, creating a oposite mask from a faded layer and raizing the contrast.

I presume you’re referencing faded print, as opposed to negative.

Usually, all that survives is the magenta dye layer, which is only helpful in creating an acceptable monochromatic example. That was the methodology for the production of black & white prints for reissue and broadcast use in the days before color broadcast.

A b/w dupe neg was created from the magenta protection record.

For a moderately faded negative, Why not simplify, and create a histogram of a very similar shot, and force the affected shot to match the available histogram.

A Histo-Match.

The problem with getting into too much technology is cost. The most financially feasible position, is still to recombine a 480i video record with a b/w representation of the affected footage.

Presuming you have one.

The overriding problem is that few people truly care to go to any expense, as most viewers have minimal concern for quality, and things look fine on an iPhone.

You mentioned the use of stills to create color. Different emulsions, only allow a basic color representation. Consider Kodachrome vs a 1950s or ‘60s Eastman Color print. Useful toward colorization, and selection of colors.

With today’s digital tools, one can try to dig very deep, and attempt to pull up a color dye layer by its bootstraps. When performed with a color neg, grain rises to the surface, along with the Y information. One could then degrain, and regrain.

One could continue, and eventually arrive at a point of discussion, similar to fairies dancing on the head of a pin.

Only film is easier to test.

Of course, there were also experiments with color masks, that were ongoing in the analogue world. Some reasonably successful.
 
Last edited:

Angelo Colombus

Senior HTF Member
Joined
Mar 19, 2009
Messages
3,406
Location
Chicago Area
Real Name
Angelo Colombus
Just saw the dvd from the library and liked it a lot. To me the two hour running time went by fast and I did like the history on the town itself. A very haunting score to go along with the still pictures and a easy Blu-ray buy for me.
 

Users who are viewing this thread

Latest Articles

Forum statistics

Threads
356,710
Messages
5,121,108
Members
144,146
Latest member
SaladinNagasawa
Recent bookmarks
0
Top