Time filter

Source Type

Sunnyvale, CA, United States

Hua K.-L.,National Taiwan University of Science and Technology | Chiu G.-M.,National Taiwan University of Science and Technology | Chin T.-L.,National Taiwan University of Science and Technology | Pao H.-K.,National Taiwan University of Science and Technology | And 2 more authors.
2013 International Conference on Computing, Networking and Communications, ICNC 2013 | Year: 2013

ABSTRACT Peer-to-peer (P2P) media streaming systems have recently become a major type of application traffic. In these applications, an important issue has been the block scheduling problem, which determines how each peer exchanges the data blocks from others. Scalable streaming in P2P networks has recently been proposed to address the heterogeneity of the network environment. In this paper, we first define a priority function for each block according to the block's significance for video content. The block scheduling problem is then transformed to an optimization problem that maximizes the priority sum of the delivered video blocks. The simulation results show that the proposed algorithm offers excellent performance for P2P streaming service. © 2013 IEEE.

Su G.-M.,Dolby Labs | Han Z.,University of Houston
Studies in Computational Intelligence | Year: 2011

With the recent rapid growth of communication, networking, and video compression technology, the real-time video streaming applications have evolved from traditional single-stream along simple transmitter-to-receiver path to complex multiple streams through advanced full-fledged cooperative networks. In this chapter, three major emerging advanced concepts are introduced: cooperative transmission, distributed source coding (DSC), and share auction based resource allocation. Cooperative transmission has been demonstrated as an effective transmission scheme to form virtual multiple-input and multiple-output (MIMO) system and provide diversity gains. Distributed source coding brings a new coding paradigm by letting the receiver jointly exploit the statistical dependencies among multiple streams sent from different sources without coding rate penalty. Share auction brings efficient way to allocate system resources in a distributed and collaborated manner to alleviate computation complexity. Based on these advanced concepts along with the advanced video processing ability for side information generation, a wireless multi-stream video transmission framework over full-fledged cooperative networks is presented. © 2011 Springer-Verlag Berlin Heidelberg.

Daly S.,University of Utah | Xu N.,Group of Snapchat Inc. | Crenshaw J.,Dolby Labs | Zunjarrao V.J.,Microsoft
SMPTE Motion Imaging Journal | Year: 2015

There are well-known observations of movie content being displayed at different frame rates. Although the terms are not consistent across the industry, four main degradations are observed of the signal as compared to nonsampled motion (i.e., real-world motion): (1) nonsmooth motion, (2) false multiple edges, (3) flickering, and (4) motion blur. In natural imagery, all four of these effects are generally visible at typical movie frame rates. The spatiotemporal window of visibility has proved successful in describing when motion looks distorted from the realworld smooth motion. However, that model predicts only detection performance and does not address the appearance or magnitude of motion distortions. In addition, well-known image capture and display parameters are also involved with frame rate questions, such as exposure duty cycle (angle), object speed, and object contrast. There are also known interactions with brightness and contrast, which are also generally linked in the end-to-end system. For example, the Ferry-Porter law1 of psychophysics indicates the temporal frequency bandwidth of vision increases with increasing adapting luminance. We aimed to isolate the nonsmooth motion component of judder in a psychophysical study by using fundamental test signals, such as the Gabor signal. Two-interval forced choice methodology was used to generate interval scales of the magnitude of judder, or judderness. Results are presented for the viewer assessment of the magnitude of judder/judderness as a function of these key parameters tested in isolation. Copyright © 2015 by the SMPTE.

Farrell S.,Dolby Labs
SMPTE 2015 Annual Technical Conference and Exhibition, SMPTE 2015 | Year: 2015

How does one become 'the best'? You run faster; jump further; eat the most hot dogs. In these examples the winner is objective. They are the Usain Bolts, the Joey Chestnuts. But how do you define 'the best' when dealing with subjective matter? Milk chocolate or dark chocolate? Rec.2020 and 1,000 cd/m2 or P3 with 4,000? In terms of image quality, both consumers and creators want to know what makes 'the best'. Consumers want a number or a letter grade, and creators and engineers want to know how to make that image or television rate the highest. It comes down to creating a metric that factors in all image attributes and weighting them by how each are valued by consumers. We will discuss the challenges faced in creating such a metric, and our study that compared image quality dimensions such as dynamic range, color primaries, bit depth, and contrast. © 2015 Society of Motion Picture & Television Engineers® (SMPTE®).

News Article | November 24, 2015
Site: www.fastcompany.com

Working on making Buffer an inclusive place where all kinds of people feel they belong and thrive is one of my favorite things. It’s a unique time to be working on diversity goals at a tech startup. On the one hand, there is quite a bit of work to do to make the strides we’d all like to make. On the other hand, there is so much hope and so many cool innovations, programs, and even tools that can help move us toward where we want to be. We’ve shared a lot of our thoughts, strategies and mistakes so far. As I begin to embark on making diversity a bigger part of my job (Potential future title: Belonging Booster. What do you think?) I wanted to check out what some other tech companies are working on when it comes to inclusivity. My research journey took me all the way from the NFL to the symphony orchestra as I discovered tech companies are pulling inspiration from many different areas. Here are seven cool inclusivity tactics that my investigation turned up. Who’s using it: Facebook and Pinterest What does tech have in common with the National Football League? At least one element: the Rooney Rule. Introduced by the NFL in 2003, the Rooney Rule (named for Pittsburgh Steelers owner Dan Rooney, who headed the league’s diversity committee) is simple. It requires teams to interview at least one minority candidate for every head-coaching and general manager vacancy. This change quickly increased the NFL’s number of minority hires from 6% to 22%. Companies like Facebook and Pinterest have adopted their own versions of the Rooney Rule, often translating it to the idea that at least one woman and one underrepresented minority be considered for every open position (sometimes, every open senior position). Who’s using it: Twitter and Pinterest 17% of all Americans are Latino, 13% of all Americans are black, 6% of all Americans are Asian, and 62% of Americans are white. 50% of Americans identify as male, and the other 50% as female. Tech companies’ numbers generally don’t reflect these percentages very closely. Pinterest acknowledged this after releasing these demographic numbers: So Pinterest did something pretty unique and transparent to address this. From its blog post: Following suit was Twitter; here are its goals for 2016: Transparently sharing specific inclusion goals feels like a great method to create accountability and honesty around how things are going. According to Fortune, LinkedIn is among the most gender-diverse of high-profile technology companies—at the end of 2014, almost half of LinkedIn’s employees were non-white. How did it get there? One clue might be in this interview with Erica Lockheimer, LinkedIn’s director of engineering growth and women in tech. Erica explains that 20% of her time—and therefore a portion of her salary and bonus—is tied to LinkedIn’s overall diversity goals. Several others in management have a similar arrangement, and at least another 50 employees have dedicated 5% of their time—and a portion of their overall annual pay—toward working on diversity at LinkedIn through things like unconscious bias training and "acts of inclusion." 20% time is a cool strategy because it takes the mission outside the realm of just a select few at an organization and makes it something everyone can work toward, together. Here’s a cool strategy from Google: the search giant is embedding engineers at historically black colleges and universities, where they teach, mentor, and advise. Google has software engineers in residence at Howard University in Washington, D.C., Hampton University in Hampton, Virginia, Fisk University in Nashville, and Spelman and Morehouse Colleges in Atlanta. Googlers teach courses and also train students on skills like how to send a professional email and how to master a technical engineering job interview. Recently, Slack added a new way for developers to connect to the app, via the "Add to Slack" button. Quite a few users noticed that the hand in the illustration wasn’t the "usual" look for design: The seemingly small detail of skin color in the launch graphics resonated with many users. Diogenes Brito, the designer on the project, wrote about the decision beautifully: Seeing yourself recognized and included in design is a powerful feeling, and Slack recognizes this in multiple ways: For a small start in this direction, check out this great connection of stock photos of tech women of color. (You might recognize the top image in this post!) Who’s using it: Facebook, Google, and others Our brains aren’t so great at making rational decisions. There are literally hundreds of cognitive biases that trick us every day. Here are just a few: How can we retrain our flawed brains? Some say a start is to be aware of our biases. Companies like Google and Facebook are working hard on unconscious bias training for its teams, and sharing some great resources with all of us. (Seriously, both those links are gold!) Speaking of unconscious bias, here’s how orchestras in the United States uncovered and worked through one of their own biases. As late as 1970, the top five orchestras in the U.S. had fewer than 5% women. By 1997, that number was up to 30%! What changed? Orchestras began using , where performers stayed behind a screen as they played, unseen by the judges. (Kinda like TV’s The Voice). According to a 2001 study, blind orchestra auditions the probability that a woman would advance from preliminary rounds by 50%. Could blind auditions work for tech, too? The startup Gap Jumpers makes software that creates a blind audition conducted via computer. Companies like Dolby Labs and Mozilla have already signed on to try this unique method. We’re excited to explore many of these strategies at Buffer soon, and will share with you how it goes. How are you working on bringing more diverse perspectives to your team, or to make your workplace more inclusive for all? I’d love to hear what strategies you’re thinking of or working on, and any thoughts you might have for Buffer as we work on building a more diverse team. This article originally appeared on Buffer and is reprinted with permission.

Discover hidden collaborations