- Joined
- Feb 8, 1999
- Messages
- 18,311
- Real Name
- Robert Harris
This is a piece about heroes and passion.
About a classic film of extremely high importance, possibly the finest work of a great filmmaker.
And about a search for excellence through decades by archivists who would not take “no” for an answer.
The name of the film could not be shorter or more simple.
M.
I was fortunate to spend some time with the lead actor in 1961, after a chance meeting in the Fox parking lot, as he attempted to enter his car, someone having parked just a bit too close. Mr. Lorre, at that point in time, was not quite as wide as he was high. A charming man, who seemed genuinely surprised that a teen would want to discuss his Warner classics, and M, which he was shocked that I had seen.
What I didn’t understand at the time, was that while I thought I had seen M, what I had actually viewed (in 16mm) were but fragments of a once-great film. The fact that the film has been available over the past eight decades in only re-cut, revamped and damaged versions is not unique in the cinema.
But this is where the heroes enter, and there are several in this drama. And I apologize for, in many cases, simplifying their years of work.
We can begin with Enno Patalas, archivist at the Munich Film Museum, who decided to bring in all known extant prints of Fritz Lang’s masterpiece, M, as well as Metropolis, (much of the material coming from the Soviet Union). He dupiicated the footage and, bit by bit, was able to come close to reconstructing M to its original state, as described to him by Mr. Lang.
Some two decades later, archivist Martin Koerber was able to locate and photo-chemically reproduce, much of the original negative in its original 1.19:1 aspect ratio. This was of extreme importance, as all early duplicating materials had been handled incorrectly, with a phantom frame line appearing at the top of the image, and picture being lost. In addition, he was able to locate nitrate prints and bring them together, allowing missing material to be added. At that time, the original audio was rendered by Martin Sawyer. New fine grains and dupes were created, and the elements cleaned up both visually and aurally. This is what made its way to the Criterion DVD, and later to the Criterion Blu-ray.
Most recently (2010-11) Torsten Kaiser arrived on the scene, and taking the now preserved elements, located more material, missing in the 2001 restoration, completing the shattered masterpiece, even further. Among other things, he was able to fill in gaps in the preservation elements with the French preservation negative. It was at this time that image stability was handled. This had also been a huge problem, as damaged elements would not permit the film to be copied in the analogue world with stability. He was also able to match the correct original timing (gradation and density).
The film was once again digitally cleaned, frame-by-frame, of dirt, detritus, dust, tears and scratches, and where possible, frames long lost were replaced from other elements, and blended as transparently as possible, into a feature length film. The audio had always been a problem; with most of the difficulty stemming from the way the tracks had been created.
M was a silent film with sound passages. The silence as important as the sound. As an early German experiment into sound film production, and with no standardization at the time in Germany, the original M track negatives were literally rolls of clear 35mm film, with the sections of track containing audio, glued to the clear base. From an archival perspective, this created a myriad of problems, as glue, dirt, nitrate stock and wear and tear all came together to do their worst.
Here’s a quote from Mr. Kaiser regarding the audio. If it doesn’t place the problems faced by the archival team in perspective, nothing will.
[Audio] is an issue that was of great concern to us, but in a different way than you might think. What you heard [a sound akin to gently running water] is the noise floor that came to be due to a combination of a number of factors over the decades - wear, poor storage, dust, moisture, decomposition and the way the track and the element itself was produced. First something about the digital work we did compared to the Criterion track, before I get to how the soundtrack was actually produced at the time, which explains a bit why it is rather shrill.
There are 2 audio tracks on the [new] Blu-ray edition. The 2001 Preservation Soundtrack is presented here, in its original form, unfiltered. The 2011 Restoration Soundtrack is based on that preservation, and we made absolutely sure that the audio would neither be limited nor affected in dynamic range, definition and registration of the frequencies or by digital artifacts during our work. [That work] was essentially to reduce the noise floor as best as possible (not entirely as it would produce artifacts, as we expected and found out the hard way when even limiting the intensity of that noise floor) and to present the intended silent sequences Lang specifically wanted "dead silent" to create a very uneasy atmosphere as the audience (then more used to just the picture) would follow the drama of the hunt for the murderer.
And if you listen more closely, you will no doubt realize immediately that it is, in fact, the Criterion that has been heavily filtered especially in the upper band, but also in the mid and lower ranges. This is not what the original (that Criterion used in 2003 as basis, just as we did) sounds like. Note the ringing of the bell when the woman with the laundry basket comes up the stairs at the beginning of the picture. It is clearly reduced in dynamic range compared to the original, which we were able to refine even more by making a new pre-mix on the 2011 Restoration. There was a lot more detail hiding even in the original, and we carefully got it out. On the Criterion, detail both in recorded dialogue as well as ambient sounds so essential to this film is no longer present - due to filtering.
Now, to the noise floor. We analyzed the original preservation track, made in 2000 from a new sound master positive directly minted in very long, difficult work by Martin Sawyer in London off the original Variable Density Sound Negative, very carefully and did long, [and] very tedious tests on the noise floor that seemed to go nowhere, as always artifacts would become an audible issue and the dynamic range/detail would also be very much affected.
It took two months and several re-written programs [until] we finally found a workable solution that also surprised Martin Koerber in a very positive way. I played the track for him when we made the preparations for final playout, and he was genuinely surprised and relieved that the quality came out so good - as was I, after so many weeks of "dead ends" that turned out to be fruitful in the end.
As I said before, a lot of the issues you heard are very much tied to the sound element, its treatment (in the previous decades), storage and (de)composition beginning with its very production. This particular sound neg (a Variable Density Track Neg) was crafted right at the very start of German experiments with sound in 1930; the process was devised by a German company called TOBIS (who named their process/format KLANGFILM). Essentially, in this particular case, you have to view this element as a stack of reels of BLANK FILM, with the variable density track segments, literally, glued to the element.
When you examine the elements, you will find that the silent portions in the film are actually BLANK, no VDT, no trace of glue, no damages or hint of anything ever being (glued) there. As for the sound segments, part of what you are hearing is misregistration due to the "opposing" wobbling and shrinkage of the two different layers and the edges of the VDT layer. What makes this even worse is that the element has of wear and tear, together with dust and scratches.
Now add to that the most problematic part in handling this element: moisture. It serves practically as glue itself that binds together any dust and debris that came through in storage over the decades in sometimes evidently not so good conditions. Removing it photochemically was out of the question. Only very few labs/technicians still remain who can handle this kind of material, so Martin Sawyer was asked to do the job of making another - as pristine as possible - positive sound element in order to get the modulation of the VDT correct. It was very difficult, as you may imagine, getting the registration off the thinly produced, deteriorated VDT Neg just right in balance of gray scale despite shrinkage, decomp, wobble etc. was a huge challenge - but the worst part was that all the dust, scratches, debris, wobble, tears etc. had to be imported into the digital realm and only here could these be "tackled".
This was what made the preservation track so "noisy" and this was why we had to work with the thought in mind that we would not get to the point where we could get everything perfect as on day one, but as best as possible and as CORRECT as possible. In other words: our job was not to get the track de-noised to get rid of background noise as was done on the Criterion, but to get it CORRECT in registration so that it would be audible as closely in the way it was produced.
Now, as for the silent sequences: Due to the copying process involved even back in the ‘30s the optical track on any minted positive off the OSN would have had a VDT record printed onto it, even if it was, under best conditions, in one shade of gray. The optical head would have read this during projection, especially after numerous uses with now added dust and scratches, as SOUND, albeit "ambient". This was a technical problem that Lang did not produce for; he even referred to it as something he would gladly get rid off. He wanted this silent sequence truly silent, which is what we made sure after very extensive analysis in these articular sequences would be the case. As a result, many sequences may sound now very alien to the "next generation" viewer, but one has to keep in mind that production in sound was still alien to the makers, then, as well. Still it is amazing how well this was done, gluing the VDT to the blank elements and all.
I’ve just screened the 80th anniversary Universum Blu-ray of M, and have come away amazed at the quality. There is now a transparency to the audio, allowing us to hear something even closer to what audiences heard in 1931 in Germany, as the audio image comes alive.
Likewise, there is now an overall stability to the film that I have never before seen, and bits and pieces previously missing, have been reclaimed.
To say that I’m thrilled by this release would be an understatement.
While the Criterion Blu-ray remains easily accessible, beautiful, and still highly recommended, those who desire to get as close as possible to the original are advised to import this very special Universum release.
Very Highly Recommended.
RAH
About a classic film of extremely high importance, possibly the finest work of a great filmmaker.
And about a search for excellence through decades by archivists who would not take “no” for an answer.
The name of the film could not be shorter or more simple.
M.
I was fortunate to spend some time with the lead actor in 1961, after a chance meeting in the Fox parking lot, as he attempted to enter his car, someone having parked just a bit too close. Mr. Lorre, at that point in time, was not quite as wide as he was high. A charming man, who seemed genuinely surprised that a teen would want to discuss his Warner classics, and M, which he was shocked that I had seen.
What I didn’t understand at the time, was that while I thought I had seen M, what I had actually viewed (in 16mm) were but fragments of a once-great film. The fact that the film has been available over the past eight decades in only re-cut, revamped and damaged versions is not unique in the cinema.
But this is where the heroes enter, and there are several in this drama. And I apologize for, in many cases, simplifying their years of work.
We can begin with Enno Patalas, archivist at the Munich Film Museum, who decided to bring in all known extant prints of Fritz Lang’s masterpiece, M, as well as Metropolis, (much of the material coming from the Soviet Union). He dupiicated the footage and, bit by bit, was able to come close to reconstructing M to its original state, as described to him by Mr. Lang.
Some two decades later, archivist Martin Koerber was able to locate and photo-chemically reproduce, much of the original negative in its original 1.19:1 aspect ratio. This was of extreme importance, as all early duplicating materials had been handled incorrectly, with a phantom frame line appearing at the top of the image, and picture being lost. In addition, he was able to locate nitrate prints and bring them together, allowing missing material to be added. At that time, the original audio was rendered by Martin Sawyer. New fine grains and dupes were created, and the elements cleaned up both visually and aurally. This is what made its way to the Criterion DVD, and later to the Criterion Blu-ray.
Most recently (2010-11) Torsten Kaiser arrived on the scene, and taking the now preserved elements, located more material, missing in the 2001 restoration, completing the shattered masterpiece, even further. Among other things, he was able to fill in gaps in the preservation elements with the French preservation negative. It was at this time that image stability was handled. This had also been a huge problem, as damaged elements would not permit the film to be copied in the analogue world with stability. He was also able to match the correct original timing (gradation and density).
The film was once again digitally cleaned, frame-by-frame, of dirt, detritus, dust, tears and scratches, and where possible, frames long lost were replaced from other elements, and blended as transparently as possible, into a feature length film. The audio had always been a problem; with most of the difficulty stemming from the way the tracks had been created.
M was a silent film with sound passages. The silence as important as the sound. As an early German experiment into sound film production, and with no standardization at the time in Germany, the original M track negatives were literally rolls of clear 35mm film, with the sections of track containing audio, glued to the clear base. From an archival perspective, this created a myriad of problems, as glue, dirt, nitrate stock and wear and tear all came together to do their worst.
Here’s a quote from Mr. Kaiser regarding the audio. If it doesn’t place the problems faced by the archival team in perspective, nothing will.
[Audio] is an issue that was of great concern to us, but in a different way than you might think. What you heard [a sound akin to gently running water] is the noise floor that came to be due to a combination of a number of factors over the decades - wear, poor storage, dust, moisture, decomposition and the way the track and the element itself was produced. First something about the digital work we did compared to the Criterion track, before I get to how the soundtrack was actually produced at the time, which explains a bit why it is rather shrill.
There are 2 audio tracks on the [new] Blu-ray edition. The 2001 Preservation Soundtrack is presented here, in its original form, unfiltered. The 2011 Restoration Soundtrack is based on that preservation, and we made absolutely sure that the audio would neither be limited nor affected in dynamic range, definition and registration of the frequencies or by digital artifacts during our work. [That work] was essentially to reduce the noise floor as best as possible (not entirely as it would produce artifacts, as we expected and found out the hard way when even limiting the intensity of that noise floor) and to present the intended silent sequences Lang specifically wanted "dead silent" to create a very uneasy atmosphere as the audience (then more used to just the picture) would follow the drama of the hunt for the murderer.
And if you listen more closely, you will no doubt realize immediately that it is, in fact, the Criterion that has been heavily filtered especially in the upper band, but also in the mid and lower ranges. This is not what the original (that Criterion used in 2003 as basis, just as we did) sounds like. Note the ringing of the bell when the woman with the laundry basket comes up the stairs at the beginning of the picture. It is clearly reduced in dynamic range compared to the original, which we were able to refine even more by making a new pre-mix on the 2011 Restoration. There was a lot more detail hiding even in the original, and we carefully got it out. On the Criterion, detail both in recorded dialogue as well as ambient sounds so essential to this film is no longer present - due to filtering.
Now, to the noise floor. We analyzed the original preservation track, made in 2000 from a new sound master positive directly minted in very long, difficult work by Martin Sawyer in London off the original Variable Density Sound Negative, very carefully and did long, [and] very tedious tests on the noise floor that seemed to go nowhere, as always artifacts would become an audible issue and the dynamic range/detail would also be very much affected.
It took two months and several re-written programs [until] we finally found a workable solution that also surprised Martin Koerber in a very positive way. I played the track for him when we made the preparations for final playout, and he was genuinely surprised and relieved that the quality came out so good - as was I, after so many weeks of "dead ends" that turned out to be fruitful in the end.
As I said before, a lot of the issues you heard are very much tied to the sound element, its treatment (in the previous decades), storage and (de)composition beginning with its very production. This particular sound neg (a Variable Density Track Neg) was crafted right at the very start of German experiments with sound in 1930; the process was devised by a German company called TOBIS (who named their process/format KLANGFILM). Essentially, in this particular case, you have to view this element as a stack of reels of BLANK FILM, with the variable density track segments, literally, glued to the element.
When you examine the elements, you will find that the silent portions in the film are actually BLANK, no VDT, no trace of glue, no damages or hint of anything ever being (glued) there. As for the sound segments, part of what you are hearing is misregistration due to the "opposing" wobbling and shrinkage of the two different layers and the edges of the VDT layer. What makes this even worse is that the element has of wear and tear, together with dust and scratches.
Now add to that the most problematic part in handling this element: moisture. It serves practically as glue itself that binds together any dust and debris that came through in storage over the decades in sometimes evidently not so good conditions. Removing it photochemically was out of the question. Only very few labs/technicians still remain who can handle this kind of material, so Martin Sawyer was asked to do the job of making another - as pristine as possible - positive sound element in order to get the modulation of the VDT correct. It was very difficult, as you may imagine, getting the registration off the thinly produced, deteriorated VDT Neg just right in balance of gray scale despite shrinkage, decomp, wobble etc. was a huge challenge - but the worst part was that all the dust, scratches, debris, wobble, tears etc. had to be imported into the digital realm and only here could these be "tackled".
This was what made the preservation track so "noisy" and this was why we had to work with the thought in mind that we would not get to the point where we could get everything perfect as on day one, but as best as possible and as CORRECT as possible. In other words: our job was not to get the track de-noised to get rid of background noise as was done on the Criterion, but to get it CORRECT in registration so that it would be audible as closely in the way it was produced.
Now, as for the silent sequences: Due to the copying process involved even back in the ‘30s the optical track on any minted positive off the OSN would have had a VDT record printed onto it, even if it was, under best conditions, in one shade of gray. The optical head would have read this during projection, especially after numerous uses with now added dust and scratches, as SOUND, albeit "ambient". This was a technical problem that Lang did not produce for; he even referred to it as something he would gladly get rid off. He wanted this silent sequence truly silent, which is what we made sure after very extensive analysis in these articular sequences would be the case. As a result, many sequences may sound now very alien to the "next generation" viewer, but one has to keep in mind that production in sound was still alien to the makers, then, as well. Still it is amazing how well this was done, gluing the VDT to the blank elements and all.
I’ve just screened the 80th anniversary Universum Blu-ray of M, and have come away amazed at the quality. There is now a transparency to the audio, allowing us to hear something even closer to what audiences heard in 1931 in Germany, as the audio image comes alive.
Likewise, there is now an overall stability to the film that I have never before seen, and bits and pieces previously missing, have been reclaimed.
To say that I’m thrilled by this release would be an understatement.
While the Criterion Blu-ray remains easily accessible, beautiful, and still highly recommended, those who desire to get as close as possible to the original are advised to import this very special Universum release.
Very Highly Recommended.
RAH