Opening night August 18
Marlon Brando’s only directorial effort, restored by NBCUniversal in collaboration with The Film Foundation
Introduced by Peter Schade, Universal Pictures
Friday night August 19
BEAT THE DEVIL, directed by John Huston, restored by Sony Pictures in collaboration with The Film Foundation to the original, uncensored version
Introduced by Grover Crisp, Sony Pictures
Saturday night August 20
McCABE AND MRS. MILLER, restored by The Criterion Collection
Introduced by Lee Kline and Karen Stetler, Criterion Collection
The Speed of Cinema III: Digital Projection of Archival Films
Jonathan Erland, Composite Components
This presentation is a progress report on the continuing efforts of the Pickfair Institute for Cinematic Studies to find a DCP solution for the display films created in non-standard frame rates. Historians, archivists and film-lovers have long had to contend with serious limitations in programming of our “silent” film heritage, including the fact that digital cinema has been unable to display images at silent film rates. Since DCP has become the ubiquitous platform for theatrical projection, it is now imperative that technical solutions for silent film frame rates be available. The approach that has been taken by Erland and his colleagues (including Qube Cinema, the manufacturer of the server and IMB board capable of supporting multiple silent film frame rates) has been an exercise in the analysis of films to derive a more accurate assessment of frame rates, and a study of the design of projection equipment to understand the role of the shutter in producing the screen effect of silent film. But the presentation problem of silent film is not only that of the image and its intermittence, but also how to synchronize recorded musical tracks prepared for silent films. Since these audio accompaniments run at the standard digital audio sampling rate their integration into a projection system for silent film constitutes another category of technical challenge for the project. A variety of other technical problems, such as certain content-related display artifacts and the adaptation of multiple projection speeds within a single film, remain on the research agenda. Ultimately, this work has meant the reverse engineering not only of the silent film projector and its functions, but also of the modern, DLP-based projection system and its server. Jonathan Erland will present a demonstration including new clips of silent films created to run at various silent film rates (i.e., 16, 18, 19, 21 fps).
The Academy Film Archive and The Film Foundation are nearing completion of the digital restoration of the recently uncovered and rarely seen US domestic version of The Front Page (1931), which is having its European premiere at the Bologna Film Festival. The team will discuss discoveries and challenges that took place during the 4K digital restoration.
In preparation for the restoration, detailed shot comparisons took place between 35mm elements held at the Library of Congress (thought to be best surviving material and the source of many 16mm prints, Blu-ray and DVD releases) and the Academy Film Archive’s recent acquisition of a 35mm safety composite print in the UNLV Howard Hughes Collection. Differences including actor movements and their placement within a scene, dialogue and camera placement were identified in both elements in every shot. This indicated at least two versions of The Front Page were created at the time of production.
Further research into The Front Page production files in the Lewis Milestone Collection at the Herrick Library confirmed the creation of three camera negatives: two for foreign and one for domestic distribution. The domestic version revealed more sophisticated editing of picture and sound as well as different dialogue to cater to American audiences. This domestic version has not been seen by audiences in decades.
The presentation will discuss the wet gate scanning, 4K workflow and hurdles to picture restoration. While the 35mm safety composite print in the Hughes Collection (made from the now lost camera negative in the 1970s) had good image structure, there was intermittent chemical staining built into the picture source that had to be addressed.
The sound for The Front Page posed its own unique challenges. The movie was distributed in two versions: sound on film and a sound on disc. The soundtrack on the only surviving 35mm print sounded less than ideal, however since the film was also released in disc form, the original master metal disc stampers were in the UNLV Howard Hughes collection. Utilizing the stampers as the primary audio source was not a simple matter, but the end result yielded a restored track that sounds better than it has in decades. The restored sound track provides well-deserved clarity to a film that is best known for its rapid fire dialogue.
IETF’s CELLAR Group is currently standardizing EBML, the Matroska container, the FFV1 video codec and the FLAC audio codec. This presentation starts with a brief summary of the recent Berlin Symposium on those topics. The main focus will show how archives can use Matroska and FFV1 not only for the preservation of video streams, but also of single image RGB content, like DPX and TIFF, which may be an alternative for small archives.
Useful Acronyms for this presentation:
- CELLAR = Codec Encoding for LossLess Archiving and Realtime transmission
- DPX = Digital Picture eXchange
- EBML = Extensible Binary Meta Language
- FFV1 = FFmpeg Videocodec 1
- FLAC = Free Lossless Audio Codec
- IETF = Internet Engineering Task Force
- RGB = Red Green Blue, a current additive color space
- TIFF = Tagged Image File Format
35 years ago this month, MTV was launched to a handful of cable providers with 24/7 airings of music videos, news, and interviews. Within 18 short months, this rag-tag network grew quickly to not only revolutionize the music industry, but worldwide pop culture in general. But as it turns out, music videos weren’t all that new of an idea in 1981, because 31 years before on KTLA, the first music films shot for television premiered on a warm Friday October 27th evening in 1950.
The Snader Telescriptions were 3-minute musical films produced by Lou Snader, a Los Angeles real estate developer who wanted to get in on the ground floor of the then growing medium of Television. Shot in 35mm and sold to local stations as a library package in 16mm, Telescriptions turned out to be the television debut of many legendary performers. Over the next 2 years, Snader and his team produced more than 750 music films in the Pop, Jazz, Country and Novelty genres. Big stars appeared in multiple Telescriptions for Snader such as Nat King Cole, Peggy Lee, Count Basie, Duke Ellington, Lionel Hampton, Mel Tormé, “Tennessee” Ernie Ford, The Four Freshmen, Burl Ives, George Shearing, Lawrence Welk and many others. Up and coming performers were also given their first exposure on Television by Snader, like Steve Lawrence and Eydie Gormé—2 years before they met on TONIGHT Starting Steve Allen. Unfortunately for Lou Snader, Telescriptions never caught on with the American television public and production was discontinued completely by the spring of 1954.
This presentation will explore the production history and technical challenges of Telescriptions, as well as their appeal, decline and issues surrounding the preservation of this important chapter in television history. We will also answer the question, “Who was the first person that dreamed up Music Videos for Television?”, plus present a rare screening of 5 of the very best Telescriptions from original 16mm prints.
- Korla Pandit – Chiu Chiu
- Peggy Lee – You Was Right Baby
- Delta Rhythm Boys – Never Underestimate The Power Of A Woman
- Sarah Vaughan – Perdido
- Nat King Cole – That’s My Girl
We are fortunate that Les Documents Cinématographiques has entrusted Eclair with the restoration of Louis Delluc’s films. This presentation will detail the restoration of four of Louis Delluc’s movies: “L’inondation” (1923), “Le chemin d’Ernoa” (1921), “Fièvre” (1921) and “La femme de nulle part” (1922). There were multiple technical issues to deal with and we will present the various stages of restoration related to each film that has allowed them to be preserved and finally screened, including physical repairing of the nitrate source elements, the reconstruction from different sources, the grading and simulation of tinting and toning, manufacturing of new digital duplicate negatives, re-creation of different film speeds for DCP, and preservation of the digital data.
A primary goal of the work was to reconstruct the films in order to present them in a version as close as possible to the director’s original cut. Use of original camera negatives was critical for much of the work, as were previous photochemical restorations as editing and tinting references. Original tinted prints also served as references and we created a complex digital grading method in order to emulate as faithfully as possible the original tinting, toning and toning with mordant. We were able to use intertitles extracted from different sources and re-created the intertitles when there were none available, based on the original typography used in Louis Delluc’s movies. Finally, a respectful digital restoration workflow was followed in order to correct for film aging or damage: fungi, decomposition, missing frames, heavy scratches, multiple scratches, etc.
For screening purposes, a simulation of the original screening speed was created by interpolating or duplicating frames to match current 24 frames per second screening standards. In conclusion, we will cover the preservation in climate controlled vaults of a new digital image negative and digital data assets of the restorations.
Digital file-based audio post-production workflows have been commonplace for over a decade and produce a myriad of digital audio assets for archiving. Some of the assets are easy to identify and archive as industry-standard broadcast WAVE files. Others are delivered and received as Pro Tools (or other DAW) sessions with complex EDL components that present challenges for long-term archiving. The recent introduction of immersive audio formats such as Dolby ATMOS has further expanded the quantity and complexity of assets for archiving. This presentation will provide an overview of digital file-based audio post-production and immersive audio workflows, along with the derivative assets they produce. Archival challenges will be discussed, along with recommendations and possible solutions for the future.
Being a film restorer today is like a dream. The scanners, the algorithms and the software restoration packages get better every year and you can freely choose between them. Things that took weeks can now be almost fully automated and completed in couple of days. This is all true for film restoration, but what about video? It has always been the “worst” medium, something you can use for news footage, game shows or talking heads, but there was little place for video in the “big art”. But what about “Doctor Who”, “The Twilight Zone”, “The Celebration”, “28 Days Later”, unique news war footage from 80’s and 90’s, Oscars, BAFTA and other ceremonies?
This presentation is built around a case study of the restoration of the video legacy of the BAFTA Awards ceremonies. The source material was of very poor quality of PAL videotapes, mostly 4th and 5th generation VHS copies of different format tapes from the early 1070’s up until just a few years ago. The study will highlight some of the unique problems in video restoration and propose a strategy to deal with them.
Cinematographer Vilmos Zsigmond and Director Robert Altman had a specific vision for the look of MCCABE & MRS. MILLER. Flashing the film, creating optical snow and trying to outsmart the studio with something that was new and different, the legendary cinematographer was very excited to use digital tools to finally get it completely right for a new 4k remaster. Unfortunately, Zsigmond passed away just before the work began. Referencing multiple prints, talking to different filmmakers and cinematographer (and losing lots of sleep), Lee will show examples and interviews to demonstrate how he tried to get the look a filmmaker wanted when they’re no longer around to speak for themselves.
Comparing completed moving image archives to their corresponding master uncompressed source is generally done by displaying them side by side or butterflied and visually comparing them. This method relies on humans that can miss differences that may be subtle or nearly impossible to detect and is not feasible for large collections.
An image comparison process will be described that uses an automated system with advanced algorithms. First, phase correlation is used to measure image misalignment creating a pseudo-pixel registration. A multi-tap FIR filter is then used to resize one image to match the other. If the original transfer function and color space characteristics are known, then one image can be corrected using that information so that its transfer function and color space matches that of the other. Otherwise, the corresponding pixel values from the two images are loaded into a curve fitting algorithm, using either polynomial or logarithmic regression, which then predicts a transfer function that can be used to adjust the pixel values of one image to match the pixel values in the other image. An automatic color correction algorithm can determine remaining color differences and apply corrections. The images will match closely so a difference image is then made and a cumulative histogram is created of those differences. The histogram is analyzed. The larger the difference the fewer pixels with that much difference are allowed before triggering a failure in comparing the images. A list of problem frame numbers/timecode can be compiled and a difference image is created for inconsistent frames that can be viewed to pinpoint the issue.
What exactly are UHD and HDR? Are they related? What’s driving their proliferation? How do we view UHD and HDR content? Have common practices developed for the creation of UHD and HDR deliverables on new films and TV shows? How do we take UHD and HDR into account when remastering or restoring older titles? Are there universal standards for each, or do these vary based on viewing platform? Are there special considerations when archiving UHD and HDR content? Join us for a brief overview of two of our industry’s new frontiers.
I have often seen myself in motion pictures, and the sight has made me very sad. I have wondered if I do walk like an animated jumping jack, or move around with such extreme rapidity as I appear to? – President Woodrow Wilson
Projection speed was much debated in the 1910s and 1920s primarily in the context of motion pictures, rarely in regards to newsreels; often these were thrust into theater schedules at even greater projection speeds than the established SMPTE standard, for both shooting and projection, of 60ft per minute, equivalent to 16fps.
But whilst speed is predominantly an issue for this presentation it is only one of the issues in regards to the restoration of films from the silent era. There is the legacy of bad duplication causing cropping with an optical mask, the truncation of title and inter-title cards to a single or even a partial frame, the re-cutting of films leaving no extant copy of the original form to rely on and scarce information on a film’s duration other than the number of reels and an overall footage length, a length that may or not include head and tail leaders.
Whilst this presentation will touch on all those issues its primary focus will be on original capture speeds and how to address that in digital restoration today, based on my personal experiences over the past decade. In the course of my work in documentary production and more recently in overseeing the restoration of documentary films the shooting speed of content and its huge variations has been not only increasingly apparent but a matter of considerable significance to me.
It was in the late 1970s, when viewing the surviving materials of the lying in state and funeral of the late Diarmaid Ó Donnabháin Rosa, in City Hall, Dublin in 1915, it was clear that the lack of available light for the camera led to an exposure rate in the region of 6-8fps. The only solution then available to reuse this material sensibly for broadcast purposes was to use optical step-printing so as to duplicate each frame 3 or 4 times for screening, this was far from satisfactory but managed to illustrate a sense of motion. The published proceedings of SMPTE, professional journals and handbooks of the period offer a record of the discussion and debate taking place around both ‘taking’ and projection speeds. They also record the admissions of projectionists in regards to the pressure they were under from theater owners and managers to maximise the amount of material projected within a working day.
Whilst certain higher profile directors may have specified projection speeds for particular sequences, as well as sound effects to be added within musical scores, the vast majority of material was left to the discretion of projectionists or the diktat of theatre owners. If a true visual record was to be not only preserved but also presented of sporting events, either summer or winter, accuracy is paramount. If a director had creatively used undercranking for dramatic effect then textural records need to survive that endorse that intent. From analysis of the surviving materials it is clear that multiple shooting speeds were utilised. What cannot be stated is whether this was for financial expediency by reducing the amount of raw stock exposed or dramatic creativity.
What can be achieved today for digital presentation through 48fps and 24fps DCPs is ‘natural’ movement that hopefully approaches the reality of the actuality of the day. Perhaps one day even greater possibilities will exist to address the criticisms of the late President Wilson.
This presentation is a Case Study of digitizing “MAMA, ICH LEBE”, a 1977 film by Konrad Wolf, for the DEFA-Foundation at OMNIMAGO GmbH and utilizing for the first time the new WetGate on DFT’s Scanity 4K scanner. Beginning in July, 2016, various film elements for the film “MAMA, ICH LEBE”, from one of the most influential directors of the former eastern republic of Germany, Konrad Wolf, arrived for digitization at OMNIMAGO GmbH. Received were the Original Camera Negative (OCN), Interpositive, Internegative and several release prints, one for the eastern republic and one for the western world, along with several different soundtracks. The client for this project is the DEFA-Foundation, which is the rights owner of all films produced in the former DDR. The actual material belongs to, and is stored at, the Bundesarchiv – Filmarchiv, a third institution. No preparation or examination of the material was done prior to receiving the material.
The OCN from 1977 has average damage, such as fragile splices, double sided notches and the most obvious defects are scratches throughout the material. Numerous copying and the cutting of different versions left their handling marks and defects over the years. But the OCN is the most complete material and will serve as the main and best source for the digitization project. The first step was to scan the OCN with the new WetGate Scanity. We scanned the first of 11 reels dry and wet and compared the results.
Right now we are still right at the beginning of this project and in the process of determining the content differences between the various elements, but a wetgate scan of the OCN is essential. Requirements for the project are a 4K 10 bit scan and to proceed in 2k with the color correction and restoration. In charge for this project during the digitization process at the lab is Korinna Barthel. She will talk about the progress made to date, the difficulties of the material and the decisions made so far.
The second part of the presentation comes from a more technical aspect. Henning Hahn will explain the benefits of the WetGate workflow on this film project, point out the differences of dry and wet scanning; also he will report on the ideas and challenges during the design process. The WetGate is designed in a way that the handling of the film is extremely gentle. The film passes through a fluid tank completely submerged in a liquid close to the refraction index of the film itself; the liquid fills in the surface imperfections, such as scratches and other surface damage, to minimize them in the scanning process. The WetGate is as an option to the existing scanner technologies to ensure low running costs, little setup time and especially the absence of sealing, rubber libs and vacuum systems that would stress the film during the process of scanning.
Over the past two decades, a few developers and researchers have been trying to solve the issue of image spread correction in variable area negative soundtracks. They were initially aiming to save time and money on the optical transfer process by avoiding the printing of new sound positives. This presentation will emphasize the higher quality that is achievable with the optical soundtrack scanners that are available today, rather than focus on the economic advantages through scanning.
After a brief technical explanation of what image spread is, how it is generated by optical recorders and fixed by either positive prints or scanners, this proposal will analyze some specific, yet broadly recurring issues in variable area optical soundtracks, and explain how these can now be solved by scanner technology. The approach proposed by this paper will link each setting of optical scanners to precise examples of audio and optical damage or human mistakes coming from a particular period, film stock, or film sound technology.
Finally we will try to envisage some possible developments of scanner technology, especially the increased collaboration between archives and laboratories, achieved by splitting raw scan and post processing and deploying the two processes to their most suitable environment.
In August 2016, a group of over 30 representatives of state and local law enforcement, prosecutors, technology vendors, civil liberties groups and public interest advocates, digital archivists and media preservationists convened at UCLA’s Department of Information Studies for “On the Record, All the Time,” a three-day meeting focused on the management and preservation of audiovisual evidence. Funded with generous support from the Institute of Museum and Library Services, the forum addressed a simple question: “What will people working with large volumes of video data, now and in the future, need to know and do?” Body-worn camera programs being proposed, piloted, and rolled out by police departments nationwide have provided a complex (and timely) use case through which to explore this question and devise educational programs and research priorities to meet anticipated needs. Forum co-convener Snowden Becker will present a hot-off-the-presses summary of the meeting’s outcomes, which will include the identification of training requirements and core competencies for information professionals managing audiovisual evidence for the long term. National Forum participant and Chain of Custody & Authenticity working group co-chair Blaine Davison will conclude the segment with an overview of the Norman (OK) Police Department’s development of a scalable solution for secure, in-house storage and management of their exponentially increasing volumes of in-car, body-worn, and surveillance video.
The Criterion Collection’s audio restorations of three film soundtracks featuring icons of classic rock – Bob Dylan in Don’t Look Back, The Who in Quadrophenia, and The Beatles in A Hard Day’s Night – provide a solid framework for exploring Criterion’s technical approach to digital audio restoration as well as ethical concerns when up-mixing or down-mixing soundtracks. The audio restoration of Don’t Look Back hinged on the unlikely successful digitization and restoration of the outstanding original 1/4″ magnetic tracks, thus allowing the film to be heard in the original mono for the first time in decades. The issue of channel status for Quadrophenia presented the opposite problem. Though it was originally conceived to have quadrophonic sound, the movie was originally released in stereo. This restoration allowed director Franc Roddam and The Who to realize the original concept of the film with quadrophonic sound, remixed by the band especially for this release. Finally, A Hard Day’s Night allows the comparison on both a restored original mono track (which was preferred by director Richard Lester) with a new 5.1 upmix created by Giles Martin and Abbey Road.
Since its inception in 2013, Cinelicious Pics, a boutique Art House distribution company, has released a mix of carefully selected first run independent features and repertory titles, the latter often restored on-site through sister company and post-production house Cinelicious. By collecting the means of acquisition, restoration, and distribution under one roof, Cinelicious Pics is able to approach the Art House market through a variant economic model, one which immediately informs the acquisition mandate, scope of restoration, and scale of distribution.
This presentation aims to highlight the ways by localizing the distribution and restoration mechanisms, and building a yearly slate of new releases directly alongside revivals, has influenced strategy behind the licensing and marketing of repertory titles, restoration workflow, preservation standards and practices, and delivery and exhibition across various venues and platforms.