Filters are digital image effects often used on social media. They initially simulated the effects of camera filters, and they have since developed with facial recognition technology and computer-generated augmented reality. Social media filters—especially beauty filters—are often used to alter the appearance of selfies taken on smartphones or other similar devices. While filters are commonly associated with beauty enhancement and feature alterations, there is a wide range of filters that have different functions. From adjusting photo tones to using face animations and interactive elements, users have access to a range of tools. These filters allow users to enhance photos and allow room for creative expression and fun interactions with digital content. == History == Beauty filters originate from Purikura ("print club"), a type of Japanese photographic arcade game machine conceived in 1994 by Sasaki Miho, a female employee at Atlus, and released in 1995 by Atlus and Sega primarily for female visitors at Japanese arcades. They allowed the manipulation of digital selfie photos with kawaii beauty filters similar to later Snapchat filters. Purikura filters included beautifying the image, cat whiskers, bunny ears, writing text, scribbling graffiti, selecting backdrops, borders, insertable decorations, icons, hair extensions, twinkling diamond tiaras, tenderized light effects, and predesigned decorative margins. To capitalize on the Purikura phenomenon in Japan during the late 1990s, Japanese mobile phones began including a front-facing camera, starting with the Kyocera Visual Phone VP‑210 in 1999. The Sanyo SCP-5300 released in 2002 was the first camera phone with filter effects, such as illumination, white‑balance control, sepia, black and white, and negative colors. Purikura-like beauty filters later appeared in smartphone apps such as Instagram and Snapchat in the 2010s. In 2010, Apple introduced the iPhone 4—the first iPhone model with a front-facing camera. It gave rise to a dramatic increase in selfies, which could be touched up with more flattering lighting effects with applications such as Instagram. The American photographer Cole Rise was involved in the creation of the original filters for Instagram around 2010, designing several of them himself, including Sierra, Mayfair, Sutro, Amaro, and Willow. However, the technology for virtual lens filters was invented and patented by Patrick Levy-Rosenthal in 2007. The patent received 100 citations, including Facebook, Nvidia, Microsoft, Samsung, and Snap. In September, 2011, the Instagram 2.0 update for the application introduced "live filters," which allowed the user to preview the effect of the filter while shooting with the application's camera. #NoFilter, a hashtag label to describe an image that had not been filtered, became popular around 2013. An update in 2014 allowed users to adjust the intensity of the filters as well as fine-tune other aspects of the image, features that had been available for years on applications such as VSCO and Litely. In 2014, Snapchat started releasing sponsored filters to monetize the participatory use of the application. In September 2015, Snapchat acquired Looksery and released a feature called "lenses," animated filters using facial recognition technology. Some of the early lenses available on Snapchat at the time were Heart Eyes, Terminator, Puke Rainbows, Old, Scary, Rage Face, Heart Avalanche. The Coachella filter released April 2016 was a popular early augmented reality filter. In April 2017, Facebook released the Camera Effects Platform, which is the first augmented reality platform that allows developers to create their own filters and effects on Facebook's Camera. In December 2017, Snapchat also launched their Lens Studio augmented reality developer tool that allows users and advertisers to do the same on the Snapchat application. In April 2022,TikTok joined the two, and launched their own augmented reality developer platform called Effect house. In February 2023, Effect House gave opened up the access to generative AI tools that allowed creators to change facial features in real time. In November 2023, TikTok released a feature where users no longer needed Effect House to create their own filters, as they are now able to create their own effects on the TikTok application. In August 2024, Meta announced that it would be removing third-party filter effects from its family of apps by January 14, 2025. The AR development software Meta Spark AR will also be retired at the same time; it was at one point the "world's largest mobile AR platform". Brand and creator effects represent the vast majority of filters available on Meta platforms, with over 2 million third-party filters available as of 2021. == Beauty filter == A beauty filter is a filter applied to still photographs, or to video in real time, to enhance the physical attractiveness of the subject. Typical effects of such filters include smoothing skin texture and modifying the proportions of facial features, for example enlarging the eyes or narrowing the nose. Filters may be included as a built-in feature of social media apps such as Instagram or Snapchat, or implemented through standalone applications such as Facetune. In 2020, the "Perfect Skin" filter for Snapchat and Instagram which was created by Brazilian augmented reality developer Brenno Faustino gained more than 36 million impressions in the first 24 hours of its release. In 2021, TikTok users pointed out how the default front-facing camera on the platform automatically applied the retouch and other feature-altering filters. Users noted that these filters slimmed down faces, smoothed skin, whitened teeth, and altered facial features such as nose and eye size, without the option to disable this feature through settings. In March 2023, the "Bold Glamour" filter was released on TikTok and instantly went viral with over 18 million videos created within its first week. This filter subtly enhances the user's facial features seamlessly, giving the illusion of fuller eyebrows, taller cheekbones, enhanced eye make up, a smaller nose, plumper lips, and clearer skin, giving off a natural yet distinct effect. As of May 2024, the filter has been used in over 220 million videos and has become a pivotal moment for beauty filters on digital platforms. Critics have raised concerns that the widespread use of such filters on social media may lead to negative body image, particularly among girls. Though Meta's intention of removing third-party filters will likely see all beauty filters removed, academics feel that the damage of beautifying filters is already done. === Background === The manipulation of photos to enhance attractiveness has long been possible using software such as Adobe Photoshop and, before that, analogue techniques such as airbrushing. However, such tools required considerable technical and artistic skill, and so their use was mostly limited to professional contexts, such as magazines or advertisements. By contrast, filters work in an automated fashion through the use of complex algorithms, requiring little or no input from the user. This ease of use, in combination with the increase in processing power of smartphones, and the rise of social media and selfie culture, have led to photographic manipulation occurring on a much wider scale than ever before. One of the earliest examples of a content-aware digital photographic filter is red-eye reduction. === Effects === Typical changes applied by beauty filters include: Smoothing skin texture; minimizing fine lines and blemishes Erasing under-eye bags Erasing naso-labial lines ("laugh lines") Application of virtual makeup, such as lipstick or eyeshadow Slimming the face; erasing double chins Enlarging the eyes Whitening teeth Narrowing the nose Increasing fullness of the lips Beauty filters most frequently target the face, though in some cases they may affect other body parts. For example, the app "Retouch Me" was reported to have a feature which allows users to superimpose visible abdominal muscles (a "six pack") onto photos featuring the subject's bare stomach. === Reception and psychological effects === Some commentators have expressed concern that beauty filters may create unrealistic beauty standards, particularly among girls, and contribute to rates of body dysmorphic disorder. A correlation has been established between negative body image and the use of beautifying filters, though the direction of causation is unknown. The inability to discern whether a particular image has been filtered is thought to exacerbate their negative psychological effects. Policymakers have advocated for social networks to disclose the use of filters; TikTok, Instagram, and Snapchat all label filtered photos and videos with the name of the filter applied. It has also been noted that beauty filters on social media tend to highlight Eurocentric features, like lighter eyes, a smaller nose, and flushed ch
Ulead MediaStudio Pro
Ulead MediaStudio Pro (MSP) is real-time, timeline based prosumer level video editing software by Ulead Systems. It is a suite of 5 digital video and audio applications, including: Video Capture, Video Paint, CG Infinity, Audio Editor and Video Editor. MSP is only available on the Windows platform. Since version 8.0, CG Infinity and Video Paint are separate from the MSP suite, and are being sold as a combination product called VideoGraphics Lab (VGL). On June 18, 2008, Corel formally announced that MediaStudio Pro would be discontinued. The final MediaStudio Pro version was 8.10.0039 (Pro 8 Service Pack 1) released June 2, 2006. Corel discontinued support for MediaStudio Pro in June 2009. Version 6.0 is last version to support Windows 95, although recent versions are not compatible with Windows Vista or Windows 7. == Modules == There are 5 stand-alone modules in MSP before version 8.0, they are: Video Capture – The video capturing module in MSP. Video Paint – A frame-by-frame editor which can let user to make some image or hand-drawing effects on video frames. CG Infinity – A vector-based video editing tool which allows user to create logo animation or vector graphics on video frames. Audio Editor – The audio editing tool in MSP. It can utilize DirectX audio filters and Ulead audio filters to do audio effect processing. Video Editor – The module that users do video editing with audio/video effects. It can also utilize DirectX audio filters and 3rd party video filters to do the video editing. Since version 8.0, CG Infinity and Video Paint have been separated from the MSP suite and are being sold as a combination product called VideoGraphics Lab (VGL). == Editions == Ulead MediaStudio Pro had several editions before version 7.0. They are: Full edition: this edition includes all 5 modules. Director's Cut edition: this edition has 3 modules including Video Capture, Video Editor and Audio Editor. SE edition: SE means Simple Edition or Special Edition and is an OEM bundle version. It also includes the 3 modules as Director's Cut, however, is feature limited. Sometimes it will be given freely in video magazines. After version 7.0 only Full edition is available in the MSP suite. On June 18, 2008, Corel formally announced that MediaStudio Pro would be discontinued. == Release history ==
Feature detection (web development)
Feature detection (also feature testing) is a technique used in web development for handling differences between runtime environments (typically web browsers or user agents), by programmatically testing for clues that the environment may or may not offer certain functionality. This information is then used to make the application adapt in some way to suit the environment: to make use of certain APIs, or tailor for a better user experience. Its proponents claim it is more reliable and future-proof than other techniques like user agent sniffing and browser-specific CSS hacks. == Techniques == A feature test can take many forms. It is essentially any snippet of code which gives some level of confidence that a required feature is indeed supported. However, in contrast to other techniques, feature detection usually focuses on performing actions which directly relate to the feature to be detected, rather than heuristics. === JavaScript === JavaScript feature detection can inspect the DOM and the local JavaScript environment to test whether browser features or APIs are supported. The simplest technique is to check for the existence of a relevant object or property. For example, the Geolocation API (used for accessing the device's knowledge of its geographical location, possibly obtained from a GPS navigation device) exposes a geolocation property on the navigator object in the DOM; the presence of which implies the Geolocation API is supported: if ('geolocation' in navigator) { // Geolocation API is supported } For a higher level of confidence, some feature tests will attempt to invoke the feature then look for clues that it behaved properly. For example, a test for support for cookies might attempt to set a value as a cookie and then verify it can be read back. === CSS === In CSS, the at-rule @supports introduced in 2015 allows to test if a given feature is supported. For instance the following code activates the declarations only if the user agent supports display: flex: == Undetectables == Some browser features are considered undetectable, because no clues are known to give sufficient confidence that a feature is supported. These are often because of limited information available to the JavaScript environment in the browser; generally features must be exposed via the DOM in some way in order to be detectable using JavaScript. When undetectables are encountered, it is common to turn to user agent sniffing as an alternative mechanism, or to employ defensive coding to minimise the impact if the feature turns out not to be supported. The Modernizr project maintains a record of known undetectables on their wiki.
Paperless society
A paperless society is a society in which paper communication (written documents, email, letters, etc.) is replaced by electronic communication and storage. The concept was first introduced by Frederick Wilfrid Lancaster in 1978. Furthermore, libraries would no longer be needed to handle printed documents. "Librarians will, in time, become information specialists in a deinstitutionalized setting". Lancaster also stated that both computers and libraries will not always give us the information that other people and living life will. == Literature == Brodman, E. (1979). Review of Toward Paperless Information Systems. Bulletin of the Medical Library Association, 67(4), 437–439. Buckland, M. K. (1980). Review of Toward Paperless Information Systems. Journal of Academic Librarianship, 5(6), 349. Grosch, A. (1979). Review of Toward Paperless Information Systems. College & Research Libraries, 40(1), 88–89. Kohl, D. F. (2004). From the editor . . . The paperless society . . . Not quite yet. Journal of Academic Librarianship, 30(3), 177–178. Lancaster, F. W. (1978a). Toward paperless information systems. New York: Academic Press. Lancaster, F. W. (1980b). The future of the librarian lies outside of the library. Catholic Library World, 51, 388–391. Lancaster, F. W. (1982a). Libraries and librarians in an age of electronics. Arlington, VA: Information Resources Press. Lancaster, F. W. (1982b). The evolving paperless society and its implications for libraries. International Forum on Information and Documentation, 7(4), 3–10. Lancaster, F. W. (1983). Future librarianship: Preparing for an unconventional career. Wilson Library Bulletin, 57, 747–753. Lancaster, F. W. (1985). The paperless society revisited. American Libraries, 16, 553–555. Lancaster, F. W. (1993). Libraries and the future: Essays on the library in the twenty-first century. New York: Haworth Press. Lancaster, F. W. (1999). Second thoughts on the paperless society. Library Journal, 124(15), 48– 50. Lancaster, F. W., & Smith, L. C. (1980c). On-Line systems in the communication process: Projections. Journal of the American Society for Information Science, 31(3), 193–200. Miall, D. S. (2001). The library versus the Internet: Literary studies under siege? Proceedings of the Modern Language Association, 116(5), 1405–1414. Salton, G. (1979). Review of Toward Paperless Information Systems. Journal of Documentation, 35(3), 250–252. Sellen, A. J., & Harper, R. H. R. (2003). The myth of the paperless office. Cambridge, MA: MIT Press. Stevens, N. D. (2006). The fully electronic academic library. College & Research Libraries, 67(1),5–14. Young, Arthur P. (2008).Aftermath of a Prediction: F. W. Lancaster and the Paperless Society LIBRARY TRENDS, 56(4),(“The Evaluation and Transformation of Information Systems: Essays Honoring the Legacy of F. W. Lancaster,” edited by Lorraine J. Haricombe and Keith Russell), pp. 843–858.
Coupling (electronics)
In electronics, electric power and telecommunication, coupling is the transfer of electrical energy from one circuit to another, or between parts of a circuit. Coupling can be deliberate as part of the function of the circuit, or it may be undesirable, for instance due to coupling to stray fields. For example, energy is transferred from a power source to an electrical load by means of conductive coupling, which may be either resistive or direct coupling. An AC potential may be transferred from one circuit segment to another having a DC potential by use of a capacitor. Electrical energy may be transferred from one circuit segment to another segment with different impedance by use of a transformer; this is known as impedance matching. These are examples of electrostatic and electrodynamic inductive coupling. == Types == Electrical conduction: Direct coupling, also called conductive coupling and galvanic coupling Resistive conduction Atmospheric plasma channel coupling Electromagnetic induction: Electrodynamic induction — commonly called inductive coupling, also magnetic coupling Capacitive coupling Evanescent wave coupling Electromagnetic radiation: Radio waves — Wireless telecommunications. Electromagnetic interference (EMI) — Sometimes called radio frequency interference (RFI), is unwanted coupling. Electromagnetic compatibility (EMC) requires techniques to avoid such unwanted coupling, such as electromagnetic shielding. Microwave power transmission Other kinds of energy coupling: Acoustic coupler
Convolutional neural network
A convolutional neural network (CNN) is a type of feedforward neural network that learns features via filter (or kernel) optimization. This type of deep learning network has been applied to process and make predictions from many different types of data including text, images and audio. CNNs are the de-facto standard in deep learning-based approaches to computer vision and image processing, and have only recently been replaced—in some cases—by newer architectures such as the transformer. Vanishing gradients and exploding gradients, seen during backpropagation in earlier neural networks, are prevented by the regularization that comes from using shared weights over fewer connections. For example, for each neuron in the fully-connected layer, 10,000 weights would be required for processing an image sized 100 × 100 pixels. However, applying cascaded convolution (or cross-correlation) kernels, only 25 weights for each convolutional layer are required to process 5x5-sized tiles. Higher-layer features are extracted from wider context windows, compared to lower-layer features. Some applications of CNNs include: image and video recognition, recommender systems, image classification, image segmentation, medical image analysis, natural language processing, brain–computer interfaces, and financial time series. CNNs are also known as shift invariant or space invariant artificial neural networks, based on the shared-weight architecture of the convolution kernels or filters that slide along input features and provide translation-equivariant responses known as feature maps. Counter-intuitively, most convolutional neural networks are not invariant to translation, due to the downsampling operation they apply to the input. Feedforward neural networks are usually fully connected networks, that is, each neuron in one layer is connected to all neurons in the next layer. The "full connectivity" of these networks makes them prone to overfitting data. Typical ways of regularization, or preventing overfitting, include: penalizing parameters during training (such as weight decay) or trimming connectivity (skipped connections, dropout, etc.) Robust datasets also increase the probability that CNNs will learn the generalized principles that characterize a given dataset rather than the biases of a poorly-populated set. Convolutional networks were inspired by biological processes in that the connectivity pattern between neurons resembles the organization of the animal visual cortex. Individual cortical neurons respond to stimuli only in a restricted region of the visual field known as the receptive field. The receptive fields of different neurons partially overlap such that they cover the entire visual field. CNNs use relatively little pre-processing compared to other image classification algorithms. This means that the network learns to optimize the filters (or kernels) through automated learning, whereas in traditional algorithms these filters are hand-engineered. This simplifies and automates the process, enhancing efficiency and scalability overcoming human-intervention bottlenecks. == Architecture == A convolutional neural network consists of an input layer, hidden layers and an output layer. In a convolutional neural network, the hidden layers include one or more layers that perform convolutions. Typically this includes a layer that performs a dot product of the convolution kernel with the layer's input matrix. This product is usually the Frobenius inner product, and its activation function is commonly ReLU. As the convolution kernel slides along the input matrix for the layer, the convolution operation generates a feature map, which in turn contributes to the input of the next layer. This is followed by other layers such as pooling layers, fully connected layers, and normalization layers. Here it should be noted how close a convolutional neural network is to a matched filter. === Convolutional layers === In a CNN, the input is a tensor with shape: (number of inputs) × (input height) × (input width) × (input channels) After passing through a convolutional layer, the image becomes abstracted to a feature map, also called an activation map, with shape: (number of inputs) × (feature map height) × (feature map width) × (feature map channels). Convolutional layers convolve the input and pass its result to the next layer. This is similar to the response of a neuron in the visual cortex to a specific stimulus. Each convolutional neuron processes data only for its receptive field. Although fully connected feedforward neural networks can be used to learn features and classify data, this architecture is generally impractical for larger inputs (e.g., high-resolution images), which would require massive numbers of neurons because each pixel is a relevant input feature. A fully connected layer for an image of size 100 × 100 has 10,000 weights for each neuron in the second layer. Convolution reduces the number of free parameters, allowing the network to be deeper. For example, using a 5 × 5 tiling region, each with the same shared weights, requires only 25 neurons. Using shared weights means there are many fewer parameters, which helps avoid the vanishing gradients and exploding gradients problems seen during backpropagation in earlier neural networks. To speed processing, standard convolutional layers can be replaced by depthwise separable convolutional layers, which are based on a depthwise convolution followed by a pointwise convolution. The depthwise convolution is a spatial convolution applied independently over each channel of the input tensor, while the pointwise convolution is a standard convolution restricted to the use of 1 × 1 {\displaystyle 1\times 1} kernels. === Pooling layers === Convolutional networks may include local and/or global pooling layers along with traditional convolutional layers. Pooling layers reduce the dimensions of data by combining the outputs of neuron clusters at one layer into a single neuron in the next layer. Local pooling combines small clusters, tiling sizes such as 2 × 2 are commonly used. Global pooling acts on all the neurons of the feature map. There are two common types of pooling in popular use: max and average. Max pooling uses the maximum value of each local cluster of neurons in the feature map, while average pooling takes the average value. === Fully connected layers === Fully connected layers connect every neuron in one layer to every neuron in another layer. It is the same as a traditional multilayer perceptron neural network (MLP). Each neuron in the fully connected layer receives input from all the neurons in the previous layer. These inputs are weighted and summed with the corresponding biases, and then passed through an activation function to perform a nonlinear transformation, generating the output. The flattened matrix goes through a fully connected layer to classify the images. === Receptive field === In neural networks, each neuron receives input from some number of locations in the previous layer. In a convolutional layer, each neuron receives input from only a restricted area of the previous layer called the neuron's receptive field. Typically the area is a square (e.g. 5 by 5 neurons). Whereas, in a fully connected layer, the receptive field is the entire previous layer. Thus, in each convolutional layer, each neuron takes input from a larger area in the input than previous layers. This is due to applying the convolution over and over, which takes the value of a pixel into account, as well as its surrounding pixels. When using dilated layers, the number of pixels in the receptive field remains constant, but the field is more sparsely populated as its dimensions grow when combining the effect of several layers. To manipulate the receptive field size as desired, there are some alternatives to the standard convolutional layer. For example, atrous or dilated convolution expands the receptive field size without increasing the number of parameters by interleaving visible and blind regions. Moreover, a single dilated convolutional layer can comprise filters with multiple dilation ratios, thus having a variable receptive field size. === Weights === Each neuron in a neural network computes an output value by applying a specific function to the input values received from the receptive field in the previous layer. The function that is applied to the input values is determined by a vector of weights and a bias (typically real numbers). Learning consists of iteratively adjusting these biases and weights. The vectors of weights and biases are called filters and represent particular features of the input (e.g., a particular shape). A distinguishing feature of CNNs is that many neurons can share the same filter. This reduces the memory footprint because a single bias and a single vector of weights are used across all receptive fields that share that filter, as opposed to each receptive field having its own bias and vector
MX1 Ltd
MX1 was a global media services provider founded in July 2016 from a merger between digital media services companies, RR Media and SES Platform Services, and a wholly owned subsidiary of global satellite owner and operator, SES. In September 2019, MX1 was merged into the SES Video division and the MX1 brand dropped. Broadcast and streamed content management, playout, distribution, and monetisation services from both MX1 and SES Video are now provided under the SES name. Before merger with SES, MX1 claimed to manage more than 5 million media assets and every day to distribute more than 3,600 TV channels, manage the playout of over 525 channels, distribute content to more than 120 subscription VOD platforms, and deliver over 8,400 hours of online video streaming and more than 620 hours of premium sports and live events. == Services == MX1 video and media services are provided through a single hybrid, cloud and on-premises solution, called MX1 360, which enables video and media solutions including content and metadata management, archiving, localisation solutions, channel playout, VOD, online video (OTT) and content distribution. Services provided by MX1 include: === Content aggregation === Acquisition of content via satellite, fibre or IP with satellite downlinking services (for encryption, re-encryption and re-muxing into different platforms), fibre reception from any location, and IP reception via the public Internet. Live sports, news and entertainment production (including in-studio, outside broadcasting, and SNG) with mobile live streaming and video contribution. === Content management === Digital mastering including scanning, conversion, restoration, quality control and localisation/versioning. Content archiving including secure, cloud and on-premises digital storage, and disaster recovery services. Metadata packaging and platform validation to enhance content discovery, searchability and cataloguing. Playout preparation and delivery to any format. === Channel origination and playout === Managed TV channel origination in SD, HD and UHD including 3D graphics, and video and audio effects, using cloud-based solution accessible from any location, with live content insertion and operation, and 24/7 monitoring. === Online video/VOD services === Content preparation and management for online video, VOD, live streaming services and Online video platforms using an ultra-high capacity content delivery network, including subscriber management, apps, DRM, social media, advertising tools, monetisation tools, metadata management, and analytics. === Content delivery === Delivery in all video formats over hybrid distribution network of satellite (using over 150 platforms), fibre (60 digital media hubs worldwide) and the Internet with complete downlink/uplink turnaround services and OTT content delivery. == Locations == MX1 has 16 offices worldwide, the most recent opened in March 2017 in Seoul, South Korea, as well as media centres in UK (London), US (Hawley, PA), Israel (Emeq Ha'Ela), Romania (Bucharest) and at the headquarters in Unterföhring near Munich, Germany. In the early part of 2017, significant upgrades were made to MX1's US media centre in Hawley, Pennsylvania, including expanding its capabilities for US based and global content aggregation, management and delivery to support US broadcasters and content providers. == History == RRsat was founded in Israel by David Rivel, an electronics, computers and communications engineer in 1981 as a communications provider, and in 2014 changed its name to RR Media to reflect its expanding global service offering. In 2015, RR Media acquired Eastern Space Systems (ESS), a Romanian provider of content management and content distribution services and satellite transmission services provider, SatLink Communications. Digital Playout Centre GmbH (DPC) was founded in 1996 by German media company, Kirch to provide playout, multiplexing, satellite uplinks and other broadcast services to Kirch's Premiere pay-TV platform (now Sky Deutschland) and other private and public German broadcasters. In 2005, SES Astra (a subsidiary of SES Global, now SES) bought 100% of DPC from Premiere and the company renamed ASTRA Platform Services GmbH (APS). In 2012, to reflect the company's expanding worldwide reach, the name was changed to SES Platform Services. In February 2016, it was announced that SES Platform Services had agreed, subject to regulatory approvals, to purchase RR Media. The acquisition was completed in July 2016, with the merged company renamed MX1 and headed by Avi Cohen, the former CEO of RR Media. In October 2017, Cohen was replaced as CEO by Wilfred Urner, the former CEO of SES Platform Services, CEO of SES subsidiary, HD+ and Head of Media Platforms and Product Development, SES Video.