Posts

Current State of HEVC/H.265

The current industry preferred codec, H.264, took over six years to mature to full end-to-end support and polished transcoding efficiency. HEVC’s timeline, however, should be accelerated compared to that of H.264. The migration to HEVC from H.264 is more of a maturation of technologies into an advanced state whereas the adoption of H.264 was a major jump from existing technologies.

The reason there’s so much fervor over HEVC is because of the many improvements it promises over its predecessors. The most prominent is its encoding efficiency. HEVC is expected to provide the same quality at half the data rate, reducing the demands on Wi-Fi and cellular networks, or twice the resolution at the same data rate, allowing HD content where it wasn’t feasible prior. Some are skeptical that this will aid in relieving network requirements arguing that HEVC will actually worsen the situation. Why? Because the lessened requirement on connection speeds is going to create and facilitate drastically greater demand. Video is already the majority of Internet traffic and is poised to become the dominant data type in the near future as demand continues to grow.

Another boon, albeit one that is much less talked about, is that image degradation is done in a manner that is much more pleasing to the human eye. As bits are lost the image takes on a smoother, almost softened, look that is less blocky than the compression artifacting of H.264.

Licensing terms for HEVC was initially a major concern for many companies until the terms were made public. H.264’s licensing incurred royalty charges whenever H.264 content was sold as pay-per-view and could have similarly, or to a greater extent, impacted HEVC. However, HEVC has no content-related royalties alleviating many costs that are currently burdened by content creators under H.264 licensing. Yet the terms are more expensive for encoder/decoder manufacturers. Where H.264 had a maximum $6.5 million in possible charges, HEVC’s cap is set at a staggering $25 million. That being said only very large companies that play in this field will be impacted; companies like Adobe, Apple, Google, Microsoft, and Mozilla. Even though this cap will only be felt by such massive companies it may greatly hinder or even prevent widespread adoption. Adobe recently announced its decision to not implement HEVC support in Flash. This may sound like an odd move, but with the global saturation of the Flash player Adobe would be facing a $25 million expenditure the quarter it implements HEVC when their quarterly net income is only around $50 million. That’s a hard pill to swallow for investors especially when Adobe and these other companies will be shouldering the cost so that content providers can distribute their content with no fees and no return for Adobe.

As with any technology in its early stages there are a few reasons to not rush in to migrating to HEVC. Currently encoding to HEVC requires drastically more power than H.264. For the individual this increase is negligible, but for massive companies such as Netflix this can be a substantial cost increase. In addition, the efficiency of current encoders is upwards of ten times that of H.264. Not only does this increase your time to deliverable file, but it also makes real-time encoders currently impossible. Early adopters are also reporting that encoding resolutions at or below 720p aren’t yielding appreciable file size savings. Even with similar file sizes they are seeing an increase in perceived quality of the video. As the technology matures the efficiency and power needs will improve addressing and alleviating these problems with real-time encoders likely just a few years away.

Several other reasons loom over HEVC, but these will also be alleviated with time. Encryption and digital rights management (DRM) for protecting the delivery of content is still emerging. Something that many content owners will require before they award approval for distribution of their content. UHDTVs, a major venue for HEVC content, are still nascent in ownership. Lastly the cost of replacing copious hardware and software systems for transcoding into HEVC looms over current companies.

With all of the benefits awaiting us with HEVC there is prudence in not rushing to quickly into adoption. As time passes and the technology matures we’ll see some truly great video in places we haven’t yet been able to.

Current State of Video in Education

Today students are intimately familiar with and expecting video in their education starting as early as elementary school.

Not only are they assuming there will be video, but their expectations are being shaped by resources like YouTube and Coursera. This trend is having a profound impact in education on instructors who embrace forward-looking pedagogies. With video, students are more engaged in creating their learning experience rather than just passively receiving information. Integration of video increases student engagement, maximizes school resources, facilitates collaboration, and can accommodate different learning styles. All of this leads to improve learning experiences and outcomes.

Flipped classroom and blended learning approaches to teaching a course are the new baseline utilizing video as a core component. However, just recording a lecture and putting it online for students to watch isn’t enough. Ideally videos need to be in short, easily digestible chunks with high production quality. This is very similar to the model that the online training company Lynda.com has adopted. An entire training module could be upwards of 24 hours of content, but it’s broken into numerous short videos typically under five minutes each, that play in sequence. As instructors have started to embrace video for their new teaching methods the biggest problem is still that do-it-yourself recordings are typically low quality plagued by simple mistakes with easy fixes, such as being back-lit creating a silhouette. Taking the time to plan, set up, and practice will make a world of difference in the end product and keep the viewer engaged.

Unfortunately, many instructors see production value as a universal solution, but condensed easily digestible content still trumps video quality. However, there’s a factor that’s less tangible while just as important as good production; the charisma of the presenter. The more engaging and entertaining they are, the more easily and likely the student is to keep watching. Monotone professors droning on will kill viewership faster than a broken overhead projector.

For entirely online courses video also serves as a reminder that the instructor is a real person not just someone hidden behind endless PowerPoint slides. Videos allow for that visual connection between the professor and their students. Not all content is conducive to having the presenter on screen all the time and can potentially create a good deal of post-production time to edit recordings to switch back and forth. At a minimum the instructor should have introductory videos for each topic chunk or week, depending on how the course is set up.

Cisco recently commissioned a review of current research on the benefits of video on learning and the quality of the educational experience and the findings are impressive. Two thirds of respondents believe that video increases student motivation, increases discussions, and helps instructors be more effective. Over 90% of university students that consumed recorded lectures felt it helped them learn course material. And almost half of elementary school children who used streaming video scored higher on their end-of-year science exam. Check out their infographic and whitepaper for more information.

Current State of Captioning

In October of 2010 President Obama signed the Twenty-First Century Communications and Video Accessibility Act (CVAA) into law. The CVAA expands upon the existing requirements set forth in the Rehabilitation Act of 1973 and the Americans with Disabilities Act of 1990 to increase access to current technologies, such as broadband and mobile, for those with disabilities. A key component of this law is the requirement that all captioned content that is delivered on television is now required to be captioned when delivered via the Internet as well.

In addition, the FCC added four non-technical quality standards in February including accuracy, completeness, placement, and synchronicity. Of those accuracy will be the most impactful to content creators. Primarily accuracy can be considered a measurement of correct words in the transcript compared to the number of incorrect words in the program. For example, a 9,050 word video with 68 errors is 99.2% accurate. However, errors could be anything from misspellings to missed words to punctuation that impedes comprehension. But the new restrictions are more comprehensive than just accuracy. “In order to be accurate, captions must match the spoken words in the dialogue, in their original language (English or Spanish), to the fullest extent possible and include full lyrics when provided on the audio track,” states the FCC. This includes the prohibition of paraphrasing and mandatory non-verbal information such as speaker identification, sound effects, and audience reactions. The FCC hasn’t given the specific accuracy requirement, but as a likely guideline the U.S. House of Representatives already requires a 98.6% accuracy for all of their floor proceedings recordings.

Schools and universities have long been held to provide accommodations under Section 504 and 508 to any student that requires them. This includes any media or online content used within a course.  The sheer volume of material that is created on a daily basis within a university can be a daunting task let alone tackling the vast amount of preexisting content. With so many organizations and businesses working just to maintain compliance many haven’t looked forward to the extended benefits of having captioned content.

Many student unions have video walls displaying several sources at once that would otherwise not be consumable.

Many student unions have video walls displaying several sources at once that would otherwise not be consumable.

With a large portion of the population experiencing hearing difficulties, but who aren’t deaf, captions can assist in digesting materials where they may struggle otherwise. This also extends to hearing individuals who may be unable to hear the content due to being in a noisy environment or where multiple pieces of content would otherwise compete with each other. From a lecture standpoint, when the material covered in a video is particularly difficult or convoluted, being able to read the transcript as the presenter is speaking can greatly assist in comprehension and notation.

Most universities host a diverse student population locally and this diversification is only compounded by new online course and program offerings. With the global popularity of massively open online courses (MOOCs) the traditional base of online students is even broader. On average OSU’s course offerings through Coursera are seeing over 60% of enrolled students coming from outside of the United States. Especially in language classes comprehending the written word is much easier before mastering speech nuances. Beyond those learning a new language or multilingual viewers, those that don’t speak the language of the content would need to rely entirely on the transcript and captions. Thick accents can pose comprehension problems even when the language between presenter and viewer is common.

Lastly, having transcripts available makes content more easily indexed by web search engines putting your content in front of more people through ease of discovery. A step beyond that, captions allow for searching within the content and being able to jump directly to the desired point of interest in certain products, such as you can with Mediasite, Ohio State’s new lecture capture solution. This increased usability gives students the flexibility to make abbreviated notes and return to the recorded lecture later for review or clarification while focusing on the lecture when in class.

With all of these benefits ready to be taken advantage of, the workflow for getting your content captioned can be discouraging to non-professionals. Fortunately, outsourcing has become a reasonably priced option. Among the many options available, OSU has its own. Transcribe OSU is a student-staffed, cost-recovery service supported by the OSU Web Accessibility Center, Office for Disability Services, and ADA Coordinator’s Office available exclusively to the Ohio State community.

You might want to wait to buy that UHDTV

What’s the problem? Your shiny new TV may be obsolete one way or another within a couple of years.
UHD supports a wider range of colors than HD does, but the parts required to support this new standard weren’t available for the production of first generation UHDTVs.

Images from Sakurambo and GrandDrake as found in an article from HD Guru

Image from HD Guru article by Sakurambo and GrandDrake

The color space used by the HDTV standard is referred to as Recommendation 709 and in addition to supported colors the recommendation also defines frame rates, pixel count, and that it’s 16:9 wide screen. The new UHD color space is Recommendation 2020. The increased color range means less banding and a higher quality color recreation. Although most displays today can’t support the new color space content creators and providers are already preparing for fully standards-capable UHDTVs.

In addition to the new UHD color space, HEVC/H.265 has multiple profiles. Every UHDTV will support the rudimentary Main profile, but content creators are encoding their material using the Main 10 profile, which uses 10-bits over Main’s 8-bits per sample, because this profile produces better quality. The problem with this lies in the fact that most UHDTVs don’t list the profiles that are supported and like early HDTVs support for the full range of specifications is hit or miss. Usually miss.

Netflix recently announced that they’re beginning the transition to HEVC Main 10 for their 4K content and many other large distributors will be quick to follow.  Hollywood as well is eager for the move to Rec. 2020 and HEVC Main 10 and are already processing their films as such. The push to move to UHD is palpable, but not everyone is ready.

So, if you choose to pick up a UHDTV soon, make sure it specifically has Main 10 profile support or if you don’t mind adding an external decoder in the future HDMI 2.0 support…unless you find a price you just can’t say no to.

Real Promise of 4K

With 4K/UltraHDTV vying for consumer adoption many are touting the new standard’s increased resolution as a way to make content even clearer. While increased clarity is certainly a boon no one seems to be talking about what could easily be a much more impactful difference, screen real estate.Television Size & Distance

Today most HD screens are overkill for where they’re used. To realize the full benefits of a shiny new 65” 4K TV you’d have to be sitting about four feet away from the screen. Even for a 1080p screen that size you’d need to be a paltry eight feet from it. (See chart at right) Once you get far enough away from any of these screens your eyes just can’t tell much of a difference.

Try scooting back a few feet from your monitor and see how much of a disparity you can discern between these two stills. (Click to view full size)
1080 vs 4K
Up close there’s an easily distinguishable difference in detail, especially in fine areas such as hair. But once you put any appreciable distance between yourself and the screen it gets lost. Since most HDTVs are used in living rooms where the viewer is ten or more feet away the increase in resolution is less and less perceivable as you move from 720p to 1080p to 4K.

Where 4K can really shine is by leveraging all of that extra real estate to display double the content at the same resolution that we are accustomed to today. By using wider angle lenses we’re able to reveal more of the scene.Take the 2014 Sochi Olympics for instance.

What you saw in HD as this …

Sochi @ 1080
… could have looked like this.
Sochi @ 4K
Swapping lenses to change how the image is captured is something that cinematographers have been doing since the early days of movies as well as in modern television content. But without increasing the resolution you’re squeezing more information onto the same number of pixels.

Movie Tight vs Wide Angle TV Tight vs Wide Angle

Video Side-by-side_thHowever, I don’t want to sell short the amazing quality that can be garnered from 4K cameras and lenses. Side-by-side comparisons of stills taken from 1080 and 4K footage are impressive, but seeing actual video is amazing. Dylan Lierz posted a great video on YouTube comparing shots taken in 1080p and 4K.  Make sure you watch it full screen at 1080p quality.

Since market saturation of 4K displays is still the vast minority one of the first benefits that the majority of consumers will see is what’s being referred to as Super 2K, 4K content scaled down to 1080p. Content creators will get all the benefits of 4K and consumers will get a crisper and higher contrast image, even though the image isn’t the native resolution of their screens. This is visible in Dylan’s comparison clip as YouTube doesn’t yet support delivery of the 4K resolution, so the 4K video is effectively Super 2K.

 

Encoding or Transcoding?

Most people use the terms encoding and transcoding interchangeably, however there is a distinct difference between the two. Encoding refers to taking an uncompressed source and converting it to a compressed file whereas transcoding is taking an already compressed file and converting it to another compression scheme. These compression schemes are commonly referred to as codecs. Codec, in this context, is the standard or format of the compression.

In order for a computer to create or decipher files using a compression format it employs a program that codes and decodes to and from that standard. This program is also referred to as a codec and is where the portmanteau is derived from; coder/decoder.  These programs are akin to their hardware counterparts endecs (encode/decode), used for encoding analog signals such as those from a VCR to files, and modems (modulator/demodulator), if you’re old enough to remember them, which were used for sending digital information over analog telephone lines.

Ah, the good old days of BBSs.  Yes, they were a thing.  Look it up.

Encoding & Transcoding

Beyond H.264

Today the most commonly consumed media format is MP4 files using the H.264 (Advanced Video Coding) compression scheme for video and AAC (Advanced Audio Coding) for audio. These provide the best, easily consumable, compression ratios and as such became the format of choice for mobile device designers and manufacturers. While H.264 does a very efficient job of compressing video, making larger sizes and higher quality possible at lower bit rates than previous formats, there’s always room for improvement. That’s where H.265, or HEVC (High Efficiency Video Coding), comes in.

HEVC is the new standard designed to supplant the currently reigning H.264. HEVC is roughly twice as efficient, allowing for the same size and quality at half the bit rate or double the size at the same bit rate. With over half of the Internet’s traffic being used for streaming video, this has a lot to offer.  Online video consumption is poised to overtake traditional cable subscriptions as tech-savvy individuals cut the cord and turn to the Internet for their TV and newer generations don’t even have a cord to cut driving Internet video consumption even higher. With 4K, or Ultra HDTV, on the horizon the better efficiency will be even more important. 4K is the new high definition standard that is twice the resolution of current HD content.Internet consumptionDigital video resolutions

All this being said, don’t expect to see HEVC in the next few years. The road to adoption is a long one. The standard was just formally published by the ITU-T Video Coding Experts Group and ISO/IEC Moving Picture Experts Group (MPEG) at the end of last year. Many companies already jumped on incorporating HEVC into new and existing products well before this, but before the average consumer will start enjoying the benefits of HEVC there are still countless encoders, servers, device manufacturers, software developers, operating system creators and others that need to commit resources to incorporating support into their products. Once HEVC has reached our hands the benefits should be many.  In addition to the commonly touted perks, such as reduced data usage for cell phone plans and prettier Netflix streams, HEVC will also make it possible for those  in remote locations on slow connections, like dial-up and ISDN, to consume media.