Digital Cinema Initiatives, LLC (DCI) is a consortium of major motion picture studios, formed to establish specifications for a common systems architecture for digital cinema systems. The organization was formed in March 2002 by Metro-Goldwyn-Mayer, Paramount Pictures, Sony Pictures, 20th Century Studios, Universal Studios, Walt Disney Studios and Warner Bros. Entertainment The primary purpose of DCI is to establish and document specifications for an open architecture for digital cinema that ensures a uniform and high level of technical performance, reliability and quality. By establishing a common set of content requirements, distributors, studios, exhibitors, d-cinema manufacturers and vendors can be assured of interoperability and compatibility. Because of the relationship of DCI to many of Hollywood's key studios, conformance to DCI's specifications is considered a requirement by software developers or equipment manufacturers targeting the digital cinema market. == Specification == On July 20, 2005, DCI released Version 1.0 of its "Digital Cinema System Specification", commonly referred to as the "DCI Specification". The document describes overall system requirements and specifications for digital cinema. Between March 28, 2006, and March 21, 2007, DCI issued 148 errata to Version 1.0. DCI released Version 1.1 of the DCI Specification on April 12, 2007, incorporating the previous 148 errata into the DCI Specification. On April 15, 2007, at the annual NAB Digital Cinema Summit, DCI announced the new version, as well as some future plans. They released the "Stereoscopic Digital Cinema Addendum" to begin to establish 3-D technical specifications in response to the popularity of 3-D stereoscopic films. It was also announced "which studios would take over the leadership roles in DCI after the current leadership term expires at the end of September." Subsequently, between August 27, 2007, and February 1, 2008, DCI issued 100 errata to Version 1.1. So, DCI released Version 1.2 of the DCI Specification on March 7, 2008, again incorporating the previous 100 errata into the specification document. An additional 96 errata were issued by August 30, 2012, so a revised Version 1.2 incorporating those additional errata was approved on October 10, 2012. DCI approved DCI Specification Version 1.3 on June 27, 2018, integrating the 45 errata issued to the previous version into a new document. On July 20, 2020, fifteen years to the day after Version 1.0, DCI issued a new DCI Specification Version 1.4 that assimilated 29 errata issued since Version 1.3. On October 13, 2021, DCI approved a new DCI Specification Version 1.4.1 that integrated the 23 errata that had been issued to DCI Specification Version 1.4. For the convenience of users, DCI also created an online HTML version of DCI Specification, Version 1.4.1. Due to the HTML conversion process, the footnotes in the DCSS now appear as endnotes. The PDF version contains pagination and page numbers whereas the HTML version does not. DCI Specification Version 1.4.2, dated June 15, 2022, includes revisions and refinements respecting Object-Based Audio Essence (OBAE), also known as Immersive Audio Bitstream (IAB). Version 1.4.2 also implements post-show log record collection utilizing SMPTE 430-17 SMS-OMB Communications Protocol Specification. Additionally, Version 1.4.2 incorporated two prior addenda: the Digital Cinema Object-Based Audio Addendum, dated October 1, 2018 and the Stereoscopic Digital Cinema Addendum, Version 1.0, dated July 11, 2007. Users using Version 1.4.2 no longer need to refer to the separate addenda. Previous DCSS versions are archived on the DCI web site. Based on many SMPTE and ISO standards, such as JPEG 2000-compressed image and "broadcast wave" PCM/WAV sound, the DCI Specification explains the route to create an entire Digital Cinema Package (DCP) from a raw collection of files known as the Digital Cinema Distribution Master (DCDM), as well as the specifics of its content protection, encryption, and forensic marking. The DCI Specification also establishes standards for the decoder requirements and the presentation environment itself, such as ambient light levels, pixel aspect and shape, image luminance, white point chromaticity, and those tolerances to be kept. Even though it specifies what kind of information is required, the DCI Specification does not include specific information about how data within a distribution package is to be formatted. Formatting of this information is defined by the Society of Motion Picture and Television Engineers (SMPTE) digital cinema standards and related documents. == Image and audio capability overview == === 2D image === 2048×1080 (2K) at 24 frame/s or 48 frame/s, or 4096×2160 (4K) at 24 frame/s In 2K, for Scope (2.39:1) presentation 2048×858 pixels of the imager is used In 2K, for Flat (1.85:1) presentation 1998×1080 pixels of the imager is used In 4K, for Scope (2.39:1) presentation 4096×1716 pixels of the imager is used In 4K, for Flat (1.85:1) presentation 3996×2160 pixels of the imager is used 12 bits per color component (36 bits per pixel) via dual HD-SDI (encrypted) 10 bits only permitted for 2K at 48 frame/s CIE XYZ color space, gamma-corrected TIFF 6.0 container format (one file per frame) JPEG 2000 compression From 0 to 5 or from 1 to 6 wavelet decomposition levels for 2K or 4K resolutions, respectively Compression rate of 4.71 bits/pixel (2K @ 24 frame/s), 2.35 bits/pixel (2K @ 48 frame/s), 1.17 bits/pixel (4K @ 24 frame/s) 250 Mbit/s maximum image bit rate === Stereoscopic 3D image === 2048×1080 (2K) at 48 frame/s - 24 frame/s per eye (4096×2160 4K not supported) In 2K, for Scope (2.39:1) presentation 2048×858 pixels of the imager is used In 2K, for Flat (1.85:1) presentation 1998×1080 pixels of the imager is used Optionally, in the HD-SDI link only: 12 bit color, YCxCz 4:2:2 (i.e. chroma subsampling in XYZ space), each eye in separate stream === Audio === 24 bits per sample, 48 kHz or 96 kHz Up to 16 channels WAV container, uncompressed PCM DCI has additionally published a document outlining recommended practice for High Frame Rate digital cinema. This document discloses the following proposed frame rates: 60, 96, and 120 frames per second for 2D at 2K resolution; 48 and 60 for stereoscopic 3D at 2K resolution; 48 and 60 for 2D at 4K resolution. The maximum compressed bit rate for support of all proposed frame rates should be 500 Mbit/s. == Related information == The idea for DCI was originally mooted in late 1999 by Tom McGrath, then COO of Paramount Pictures, who applied to the U.S. Department of Justice for anti-trust waivers to allow the joint cooperation of all seven major motion picture studios. Universal Pictures made one of the first feature-length DCPs created to DCI specifications, using their film Serenity. Although it was not distributed theatrically, it had one public screening on November 7, 2005, at the USC Entertainment Technology Center's Digital Cinema Laboratory in the Pacific Theatre, Hollywood. Inside Man (2006) was Universal's first DCP commercial release, and, in addition to 35mm film distribution, was delivered via hard drive to 20 theatres in the United States along with two trailers. The Academy Film Archive houses the Digital Cinema Initiatives, LLC Collection, which includes film and digital elements from DCI's Standard Evaluation Material (StEM), a 12-minute production shot on 35mm and 65mm film, created for vendors and standards organizations to test and evaluate image compression and digital projection technologies.
Simple interactive object extraction
Simple interactive object extraction (SIOX) is an algorithm for extracting foreground objects from color images and videos with very little user interaction. It has been implemented as "foreground selection" tool in the GIMP (since version 2.3.3), as part of the tracer tool in Inkscape (since 0.44pre3), and as function in ImageJ and Fiji (plug-in). Experimental implementations were also reported for Blender and Krita. Although the algorithm was originally designed for videos, virtually all implementations use SIOX primarily for still image segmentation. In fact, it is often said to be the current de facto standard for this task in the open-source world. Initially, a free hand selection tool is used to specify the region of interest. It must contain all foreground objects to extract and as few background as possible. The pixels outside the region of interest form the sure background while the inner region define a superset of the foreground, i.e. the unknown region. A so-called foreground brush is then used to mark representative foreground regions. The algorithm outputs a selection mask. The selection can be refined by either adding further foreground markings or by adding background markings using the background brush. Technically, the algorithm performs the following steps: Create a set of representative colors for sure foreground and sure background, the so-called color signatures. Assign all image points to foreground or background by a weighted nearest neighbor search in the color signatures. Apply some standard image processing operations like erode, dilate, and blur to remove artifacts. Find the connected foreground components that are either large enough or marked by the user. For video segmentation the sure background and sure foreground regions are learned from motion statistics. SIOX also features tools that allow sub-pixel accurate refinement of edges and high texture areas, the so-called "detail refinement brushes". As with all segmentation algorithms, there are always pictures where the algorithm does not yield perfect results. The most critical drawback of SIOX is the color dependence. Although many photos are well-separable by color, the algorithm cannot deal with camouflage. If the foreground and background share many identical shades of similar colors, the algorithm might give a result with parts missing or incorrectly classified foreground. SIOX performs about equally well on different benchmarks compared to graph-based segmentation methods, such as Grabcut. SIOX is, however, more noise robust and can therefore also be used for the segmentation of videos. Graph-based segmentation methods search for a minimum cut and therefore tend to not perform optimally with complex structures. The algorithm has initially been developed at the department of computer science at Freie Universitaet Berlin. The main developer, Gerald Friedland, is now faculty at the EECS department of the University of California at Berkeley and also a Principal Data Scientist at Lawrence Livermore National Lab. He continues to support the development through mentoring, e.g. in the Google Summer of Code.
Language model
A language model is a computational model that predicts sequences in natural language. Language models are useful for a variety of tasks, including speech recognition, machine translation, natural language generation (generating more human-like text), optical character recognition, route optimization, handwriting recognition, grammar induction, information retrieval and disaster response. Large language models (LLMs), currently their most advanced form as of 2026, are predominantly based on transformers trained on larger datasets (frequently using texts scraped from the public internet). They have superseded recurrent neural network-based models, which had previously superseded the purely statistical models, such as the word n-gram language model. == History == Noam Chomsky did pioneering work on language models in the 1950s by developing a theory of formal grammars. In 1980, statistical approaches were explored and found to be more useful for many purposes than rule-based formal grammars. Discrete representations like word n-gram language models, with probabilities for discrete combinations of words, made significant advances. In the 2000s, continuous representations for words, such as word embeddings, began to replace discrete representations. Typically, the representation is a real-valued vector that encodes a word’s meaning such that words closer in vector space are similar in meaning and common relationships between words, such as plurality or gender, are preserved. == Pure statistical models == In 1980, the first significant statistical language model was proposed, and during the decade IBM performed 'Shannon-style' experiments, in which potential sources for language modeling improvement were identified by observing and analyzing the performance of human subjects in predicting or correcting text. === Models based on word n-grams === === Exponential === Maximum entropy language models encode the relationship between a word and the n-gram history using feature functions. The equation is P ( w m ∣ w 1 , … , w m − 1 ) = 1 Z ( w 1 , … , w m − 1 ) exp ( a T f ( w 1 , … , w m ) ) {\displaystyle P(w_{m}\mid w_{1},\ldots ,w_{m-1})={\frac {1}{Z(w_{1},\ldots ,w_{m-1})}}\exp(a^{T}f(w_{1},\ldots ,w_{m}))} where Z ( w 1 , … , w m − 1 ) {\displaystyle Z(w_{1},\ldots ,w_{m-1})} is the partition function, a {\displaystyle a} is the parameter vector, and f ( w 1 , … , w m ) {\displaystyle f(w_{1},\ldots ,w_{m})} is the feature function. In the simplest case, the feature function is just an indicator of the presence of a certain n-gram. It is helpful to use a prior on a {\displaystyle a} or some form of regularization. The log-bilinear model is another example of an exponential language model. === Skip-gram model === == Neural models == === Recurrent neural network === Continuous representations or embeddings of words are produced in recurrent neural network-based language models (known also as continuous space language models). Such continuous space embeddings help to alleviate the curse of dimensionality, which is the consequence of the number of possible sequences of words increasing exponentially with the size of the vocabulary, further causing a data sparsity problem. Neural networks avoid this problem by representing words as non-linear combinations of weights in a neural net. === Large language models === Although sometimes matching human performance, it is not clear whether they are plausible cognitive models. At least for recurrent neural networks, it has been shown that they sometimes learn patterns that humans do not, but fail to learn patterns that humans typically do. == Evaluation and benchmarks == Evaluation of the quality of language models is mostly done by comparison to human created sample benchmarks created from typical language-oriented tasks. Other, less established, quality tests examine the intrinsic character of a language model or compare two such models. Since language models are typically intended to be dynamic and to learn from data they see, some proposed models investigate the rate of learning, e.g., through inspection of learning curves. Various data sets have been developed for use in evaluating language processing systems. These include: Massive Multitask Language Understanding (MMLU) Corpus of Linguistic Acceptability GLUE benchmark Microsoft Research Paraphrase Corpus Multi-Genre Natural Language Inference Question Natural Language Inference Quora Question Pairs Recognizing Textual Entailment Semantic Textual Similarity Benchmark SQuAD question answering Test Stanford Sentiment Treebank Winograd NLI BoolQ, PIQA, SIQA, HellaSwag, WinoGrande, ARC, OpenBookQA, NaturalQuestions, TriviaQA, RACE, BIG-bench hard, GSM8k, RealToxicityPrompts, WinoGender, CrowS-Pairs
Best AI Voice Assistants in 2026
Trying to pick the best AI voice assistant? An AI voice assistant is software that uses machine learning to help you get more done — it scales effortlessly from a single task to thousands. The best picks balance beginner-friendly simplicity with the depth power users need, and they ship updates often. Whether you are a beginner or a pro, the right AI voice assistant slots into your workflow and pays for itself fast. This guide breaks down the top picks, their pros and cons, and who each one is best for.
The Best Free AI Clip Maker for Beginners
Looking for the best AI clip maker? An AI clip maker is software that uses machine learning to help you get more done — it can save you hours every week by automating repetitive work. Most options offer a generous free tier, with paid plans unlocking higher limits, faster processing, and team features. Whether you are a beginner or a pro, the right AI clip maker slots into your workflow and pays for itself fast. Read on for hands-on impressions, pricing tiers, and the standout features that matter.
Automated Lip Reading
Automated Lip Reading (ALR) is a software technology developed by speech recognition expert Frank Hubner. A video image of a person talking can be analysed by the software. The shapes made by the lips can be examined and then turned into sounds. The sounds are compared to a dictionary to create matches to the words being spoken. The technology was used successfully to analyse silent home movie footage of Adolf Hitler taken by Eva Braun at their Bavarian retreat Berghof. The video, with words, was included in a documentary titled "Hitler's Private World", Revealed Studios, 2006 Source: New Technology catches Hitler off guard
Nicolò Cesa-Bianchi
Nicolò Cesa-Bianchi (Italian pronunciation: [nikoˈlɔ tˈtʃɛːza ˈbjaŋki]) is an Italian computer scientist and Professor of Computer Science at the Department of Computer Science of the University of Milan. He is a researcher in the field of machine learning, and co-author of the books "Prediction, Learning, and Games" with Gabor Lugosi and "Regret analysis of stochastic and nonstochastic multi-armed bandit problems" with Sébastien Bubeck == Education and career == Cesa-Bianchi graduated in Computer Science from the University of Milan in 1988 where he received a PhD in Computer Science in 1993 supervised by Alberto Bertoni. During his PhD, he visited UC Santa Cruz where he worked with Manfred Warmuth and David Haussler. He did his postdoctoral studies at Graz University of Technology under the supervision of Wolfgang Maass. == Research == His research contributions focus on the following areas: design and analysis of machine learning algorithms, especially in online machine learning algorithms for multi-armed bandit problems, with applications to recommender systems and online auctions graph analytics, with applications to social networks and bioinformatics == Awards and honors == Cesa-Bianchi received a Google Research Award in 2010, a Xerox University Affairs Committee Award in 2011, a Criteo Faculty Award in 2017, a Google Faculty Award in 2018, and a IBM Academic Award in 2021. Since 2023 he is corresponding member of the Accademia dei Lincei.