Browsing Tag: Video

    web design

    Optimizing Video For Size And Quality — Smashing Magazine

    02/15/2021

    About The Author

    Doug leads developer relations at api.video, studying the intersection of media and web performance. A Google Developer Expert and the author of O’Reilly’s …
    More about
    Doug

    Adding video to your application can increase customer engagement and satisfaction. But the exact opposite can occur when there are issues with the video playback: video stalls are frustrating and drive customers away. In this article, we’ll walk through the steps to optimize the video on your website to ensure fast playback and reduce stalls.

    Over the last few years, more and more projects are using video as an integral part of the application. This is a great direction, as videos are more engaging than still photos (videos can double conversion rate and increase time spent on site), and as such, can really draw customers to explore details about products and services. However, it all goes sideways when there are issues related to the video playback.

    Video playback issues are directly related to the size and bitrate of the video. A video with large dimensions or a high bitrate will take longer to download and will require a higher speed network to play back smoothly. This leads to longer startup times, and if the network cannot supply the video fast enough, the video will stall during video playback.

    There is a solution though! By running basic optimizations of our videos before adding them to our websites, we can prevent these issues from occurring for good — well, most of them. All we really need to do is make the file smaller — in one way or another. So, now the trick is: how do we make the file smaller without reducing quality?

    In this article, we’ll walk through the tools and some of the steps you can take to optimize your videos for playback — all of it to avoid stalls and impress your precious customers!

    Real-World Data

    It’s not uncommon to find websites with extremely large videos — for example, used used as hero background videos. In my research, I was looking into sites found in the December 2020 mobile HTTPArchive, and it wasn’t difficult to spot a good number of sites loading huge video files by default, both on mobile and on desktop.

    It’s of course doubtful that you will be able to achieve the same savings that I’ll be showing here, but you will get some useful pointers and tips on things to keep in mind when dealing with videos. In fact, it is very easy to accidentally place extremely large videos on your website if you are not careful, resulting in them being almost unusable for most of your customers.

    The Pumpkin Patch Story

    Imagine that it’s mid-October, and you’re looking for a pumpkin patch and a corn maze to spend a weekend afternoon with your family. In the comfort of your desktop machine, you search the web for a nearby location and find the perfect one. The website looks lovely, with a beautiful drone 4K video of the fields playing at the top of the page. You pick the URL and send it to yourself and your loved ones so that you together can continue exploring this option on the go.

    But when you open the page on your phone, you notice a glitch: the video is desperately trying to play on your phone, but unfortunately fails to do so. The video keeps stalling and restarting over and over, being much more disruptive and annoying than it was on your computer. Eventually you move on, bookmark the URL, and move on with your daily routine.

    After a fun muddy day (well, I have recently lived in Seattle and the UK, so pumpkin patches are muddy), you’re back on your computer: perhaps you think yet again about that video and you wonder why it wasn’t playing well on your phone. Well, let’s diagnose what is going on.

    You might start by opening DevTools in your browser. Once the page is loaded, we can move to the Network tab, and filter by “media” to see all the video files:

    Filtering resources by ‘Media’ in DevTools. (Large preview)

    We see that an MP4 file is being downloaded. The file doesn’t come through the network as a standalone file; rather, the streaming service must be breaking up the file into a few segments, so you might see several 206 (partial content) requests for the same file.

    Looking at the response headers for this file, we can spot some details:

    accept-ranges: bytes
    access-control-allow-headers: x-test-header, Origin, X-Requested-With, Content-Type, Accept
    access-control-allow-methods: GET, POST, PUT, DELETE, OPTIONS
    Content-Length: 87690242
    Content-Range: bytes 70025216-157715457/157715458
    content-type: video/mp4
    date: Fri, 22 Jan 2021 15:27:26 GMT
    last-modified: Mon, 24 Jun 2019 05:13:04 GMT
    server: Apache
    

    Now, some of these numbers are slightly scare as they are slightly large. In fact, they are often so large that I’ve found myself getting into the habit of adding commas, so I can get an idea of how large the files actually are. In this case, the partial download is 87 MB, and the entire file is 157,715,457 bytes. Yes, that’s right, this video is 157MB, and it (tried) to load on my phone earlier today! No wonder it didn’t succeed.

    So What’s Up With This Video?

    Let’s dive a little bit deeper. Apparently, the video is way too large to play smoothly on a mobile phone with a lower memory and a slower network. But what do we need to fix it? To figure out what exactly is the problem, we can use FFMPEG, which is open source and free, and proves to be one of the most reliable tools to optimize videos. We could endlessly tweak the configuration in FFMPEG, but let’s just touch on a few important ones in this article.

    So, let’s start with the diagnosis tool called FFprobe. FFprobe gathers information from multimedia streams, and provides you with the details about how the video is encoded and how it will play. It’s a part of the FFMPEG package, and actually quite easy to use.

    Even better: if your video is online already, there is an online version of ffprobe that we can jump to right away. So, let’s just enter the URL into the form, and get the details about the video in return (e.g. video dimensions, bitrate, and quite a bit of metadata).

    When I add the MP4 URL from the pumpkin farm, we immediately see one of the issues. The show_format response from ffprobe returns a summary: apparently, there are 2 streams, and it’s 62s long (which all sounds normal enough to not raise any suspicions). But when we get to the size and bitrate, we immediately see where the video is failing.

    (Large preview)

    As mentioned above, it might be a good idea to get used to adding commas to these large numbers. As it turns out, indeed the drone footage flying over the field is 157MB, and has a bitrate of 20 MB per second! It means that for the video to play seamlessly, the network must be able to stream the video at a rate faster than 20 MBPS, which is exactly why it was stalling on the phone.

    What is the ideal playback bitrate?

    To avoid stall, we need to stream the video at an appropriate rate. That’s where bitrate becomes important. Bitrate is the playback speed of the video. For the browser to play the video smoothly, the video has to be downloading faster than it plays back — meaning that the video will only play back smoothly if the network speed is over 20 MBPS. When I think of network speeds, I tend to rely on WebPageTest’s traffic profiles:

    Traffic profiles to keep in mind on WebPageTest: ranging from Cable and DSL to 3G Slow and 2G. (Large preview)

    As we can tell from the overview abouve, the video might play well on the “Native Connection”, and on the ultra-fast optic cable FIOS connection (20 MBPS is exactly the speed required, so hopefully nothing else needs to be downloaded in the background). However, all the other connections have a downlink speed that’s significantly lower than 20 MBPS. If the video is loading at these speeds, the player will attempt to consume the video faster than it can be downloaded, and the video will permanently stall.

    The bitrate of your video sets the minimum network speed that your customers can use. In general, the bitrate of your video should be about 80% of the available throughput on the network. So a 20 MBPS video really needs 24 MBPS network throughput to play the video seamlessly. Everyone on a slower connection will have a quite poor experience and is likely to not be able to watch the video at all. More specifically, this means that for us to play smoothly and silky on 4G, the bitrate has to stay below 7.2 MBPS.

    Can We Lower This Video’s Bitrate?

    Yes! Let’s look at some of the configurations we can use to reduce this video’s bitrate. But first, let’s look at the data we get from FFprobe. One thing that is quite noticeable is the r_frame_rate value, which is the number of frames per second in the video. Its value is 60000/1001. It means that the frame rate for the video is 60 frames per second. However, typical frame rates on the web are 25–30, so the first thing we can do is to re-encode the video with a lower bitrate.

    Another thing to keep in mind is Constant Rate Factor. In FFMPEG, the principal quality/size benchmark is the Constant Rate Factor (CRF) compression, with values ranging from 0 (no compression) to 50 (high compression). The default value for CRF in FFMPEG is 23 (if you leave out the CRF parameter, your video is with that value). In my personal experience, values from 23-28 still produce high-quality videos, looking good on the web and greatly reduced in file size.

    So let’s start at 30fps and a CRF of 23. The Terminal command will look like this:

    ffmpeg -i input.mp4 -vcodec h264 -acodec aac -crf 23 -strict -2 :v fps=fps=30 output.mp4
    

    Voilà! This results in an 81.5 MB video — already a 48% improvement. But the video is still very large, with a 10 MBPS bitrate. If we set CRF to 28, the file drops to 35.4MB, with a bitrate of 4.5 MBPS which is much more likely to play well on a 4G connection.

    This is a five-time improvement over the original video. To make this video even more accessible, we can resize the video to make it smaller. That’s something we’ll discuss in the streaming section below.

    The Hungry For Pizza Story

    Imagine that you’re in Los Angeles, perhaps visiting from abroad and roaming on your phone, and of course thinking about grabbing a slice of pizza. You find a remarkable pizza place on your phone, and decide to head there. You may have noticed a few videos and hero images on the page, but really, every pizza place kind of looks the same, so you didn’t bother to watch the video. You head and grab a slice or two before heading back to your hotel.

    That night, you get a text from your carrier that you used a lot more data than you imagined (and definitely way more than you originally planned!). A couple of cabs, and the pizza website — how expensive was the pizza website again?

    You pop the pizza website into WebPageTest and check it on a mobile connection:

    A pie chart with the video taking up 80.2% of the entire data consumed. (Large preview)

    44 MB of video. Where is it coming from? Even beyond that, when we examine the source and the waterfall in a bit more detail, we can see that there are actually two videos! Fortunately (or unfortunately?), neither managed to be downloaded entirely:

    Video 1 downloaded 11.8 MB (of 121 MB total)
    Video 2 downloaded 31.1MB (of 139 MB total)

    This raises a few concerns and a few questions.

    First, why was so much video downloaded when it wasn’t autoplaying? We haven’t managed to click anything just yet, but already used almost 40 MB of data. The answer, as always, lies in the source. Well, “view source”, that is.

    <video
      id="u457537-video"
      class="video-js vjs-big-play-centered"
      controls
      preload="auto"
      width="1050"
      height="591"
      poster="assets/home_poster.jpg"
      data-setup='{"fluid": true}'>
      <source src="http://www.smashingmagazine.com/assets/home_1.mp4" type='video/mp4'>
      <source src="assets/home.webm" type='video/webm'>
      <p class="vjs-no-js">To view this video please enable JavaScript, and consider
        upgrading to a web browser that <a href="http://videojs.com/html5-video-support/"
        target="_blank">supports HTML5 video</a></p>
    </video>
    

    Off the bat, we see at least two issues:

    • preload=“auto”
      When we set preload="auto", we are overriding the browser’s default setting, enforcing video download — whether or not your customer has pressed “Play”. The default preload attribute is metadata, and would have resulted in a few 100KB downloaded. Admittedly, it is a much better outcome for site visitors who will never watch this video.
    • Video Order
      If you have multiple versions of the video (in this case: h264 .mp4 and VP8 .webm encoded videos), the browser will choose the first video it knows how to play. Now, every modern browser supports mp4, while most modern browsers also support webm (95.4% global support, according to CanIUse).

    One trick that I like to use is to insert the appropriate video source line with Javascript. That way, if you so choose to not serve video on certain screens, you just have an empty <video> tag — and no video can be downloaded.

    
        window.onload = addAutoplay();
        var videoLocation  = document.getElementById("hero-video");
    
        function addAutoplay() {
            if(window.innerWidth > 992){
                videoLocation.setAttribute("autoplay","");
          };
        }
    
    

    If we now run an ffprobe on these two videos, we’ll discover significant differences in sizes:

    The webm is 90% smaller, and yet has 0 views, since every browser supports the mp4. These two videos are both 640×360, and 140s long. Running the ffmpeg command from above on the mp4 results in a 12.4 MB video, so it’s likely that developers followed a similar process to compress and encode the .webm variant as well. Perhaps having preload="auto" for 12.5 MB would not be so bad after all.

    The second video (drone footage inside the restaurant) is filmed in Full HD (1080p), but similarly gets compressed from 140MB to 35 MB. So, 120s with FFMPEG could reduce the video weight on this page from 160 MB to 57 MB. Flipping the webm/mp4 order would save an additional few MB for 95% of the browsers that can support that format.

    What if we wanted to do even better, perhaps make the videos responsive to various sized screens? Well, let’s get even smaller videos — with responsive videos!

    The <video> tag doesn’t support media queries to serve different video files to different screens, so we need a different way to provide videos sized for the device screen. The easiest way to achieve that is by using video streaming. This will add some Javascript and other assets for the video player that will be required, but the video savings will definitely make up for this extra data.

    We can create video streams with FFMPEG (I have used bash scripts like this in the past), but this requires us to know all the sizes and settings we’d like to use (and as mentioned before, FFMPEG has a lot of settings!).

    To make it easier to stream video, there are APIs (e.g. api.video and Mux) where you upload your video, and the tools create video streams and host your video for you. For full disclouse, I do work at the former one, so to simplify my video processing pipeline, I’ll use api.video, to transcode and host my videos. With the upload API, I can upload any video, and the tool will create a streaming version at many different dimensions and bitrates (currently 240p, 360p, 480p, 720p, 1080p and 4K).

    The bitrates for the smaller videos are greatly reduced, as the dimensions of the video decrease. This means that the video will require less network capacity on smaller screens and will play on slower networks.

    For brevity, we’ll test only the Pumpkin patch video. I’ve received similar results with the drone video (the other pizza video is just 360p, so it does not greatly benefit from smaller sizes).

    Note: Please recall that this video is currently a 1080p mp4 video at 60fps, and weighs 157 MB for all visitors.

    With some optimizations (CRF 28 and reducing the framerate to 30fps), the video was reduced to 35.7 MB. Using DevTools, we can emulate devices to see how much data is used for video playback of streaming video on different sized screens.

    The table below is showing the total amount of traffic used. With HLS video, there is a JavaScript player, CSS, fonts, etc. that add about 1 MB of additional overhead. This is included in the totals below:

    Device Video Size (Pixels) Video Size (MB) Bitrate (MBPS)
    Moto G4 (Portrait) 240p 3.1 MB 0.35
    Moto G4 (Landscape) 360p 7.5 MB 0.800
    Iphone 7/7/8 (Landscape) 480p 12.1 MB 1.40
    Ipad (Landscape) 720p 21.2 MB 2.6
    Ipad Pro (Landscape) 1080p 39.4 MB 4.4

    At 1080p, there is about 4MB additional assets downloaded for the stream, but for every other size, there are significant data savings with no loss in video quality. Not only will the video be sized properly for the devices, but it is much less likely to stall, as the bitrate is reduced for the devices most likely to be on slower mobile connections.

    Video streaming takes care of framerate, video size and quality concerns — ensuring fast playback on any size screen, and any speed network.

    Another advantage to video streaming: if the network is slow (or suddenly becomes slower), the player can adjust the video being shown, and play a lower quality version of the video — ensuring playback on the device — even in poor network conditions. (You can test the different videos with StreamOrNot, a little open source project that I’ve released a while back.

    Now, isn’t it a little bit too much overhead? Couldn’t we do the same (just much faster) with YouTube or Vimeo? We surely could, but then we wouldn’t be able to completely remove the branding or advertising from the video, not to mention the overhead of scripts loaded within the video player iframe. Plus, sometimes you might want to use the video as a background video on your product page, and avoid any kind of external branding at all.

    Conclusion

    We do not deploy images from our camera directly to the web, but we compress and resize them to balance quality and web performance. The same should be done for video files as well. Smaller videos start playing faster and stall less often, improving the user experience of the website.

    In this article, we’ve walked through a few simple steps to optimize our videos, e.g. by lowering the quality and its framerate. We also looked at how video streaming can allow us to build a more responsive video experience for the web — automatically serving videos that are properly sized for the screen of the device.

    Thanks for reading, and if you’d like to learn more, you may want to read more on video best practices here, on Smashing Magazine, and on my blog:

    Smashing Editorial
    (vf, il)

    Source link

    web design

    Which Video Conferencing Tools Are Most Accessible? — Smashing Magazine

    06/15/2020

    About The Author

    Claudio is a certified accessibility professional with over 20 years’ experience in UX design and front-end development. He is an accessibility expert that …
    More about
    Claudio

    In this article, Claudio Luis Vera explains what to keep in mind when choosing between video conferencing tools to benefit everyone on your team — including those with disabilities.

    With us all suddenly working from home in the COVID era, most of us have had to make some quick choices from a huge variety of video conferencing and collaboration tools. But do all of these apps work well for everyone? If you have a co-worker or a student with a disability, would they be able to participate as well?

    If you’ve had a disability yourself — or if you work in government or higher education — you’ve probably faced this question already. If not, pay attention, because the wrong choice could leave your organization vulnerable to a discrimination lawsuit.

    With the right apps, working from home is a blessing for those with a disability. If you can’t drive, you no longer have to rely on poor transportation like paratransit or expensive ride-share services. Working from home allows you to attend to your medical needs in ways that might be disruptive in an office, like running a nebulizer or using an IV. Or you may avoid the hassle of navigating a building that’s not truly accessible.

    With the wrong tools, a person with a severe disability may actually experience new barriers. They could be left out of meetings and other types of collaboration, and that can create significant liability for the employer.

    “When people with disabilities are excluded because a product is inaccessible, it’s not only frustrating, but it [also] affects our colleagues and clients that work with us.”

    — Lynette Frison, screen reader user

    Most of us make a common mistake when we set up large meetings or online events: we assume that everyone who attends will be fully able to participate. Unless you know everyone’s ability level — like with your immediate co-workers — you should ask if someone needs accommodations or assume that someone attending may have difficulty seeing, hearing, or working a mouse.

    Zoom: The Favorite For Accessibility

    Ask anyone in the disability community about video conferencing tools, and they will almost certainly point you to Zoom. There’s no shortage of favorable opinions and anecdotal support for Zoom in the disability community, as the app has a reputation for providing a consistently high level of access. In fact, 92% listed Zoom as their tool of choice in a survey conducted by Fable, an accessibility testing firm that draws testers from a variety of different disabilities.

    An app’s accessibility comes from ensuring that the app’s functionality works for everyone, including those using assistive technologies like screen readers, magnification, and captions. Most software developers don’t test their apps thoroughly for compatibility with assistive technologies or for usability with people with disabilities.

    Even the simple task of joining a meeting can present a roadblock if the app hasn’t been developed with accessibility in mind. For example, unnecessary preference dialogs when entering a meeting can block a blind person that’s using a screen reader from attending. Zoom offers the convenience of clicking on an email link and allowing the participant to join a meeting immediately. “You follow the link and that’s it”, says Eric Barrette, director of technology for the Lighthouse of Broward, an organization for the blind and visually impaired.

    The simplicity of Zoom’s one-click sign-on belies the depth of the application behind it. Zoom’s product philosophy favors universal design over separate applications for different user types. The result is an app that is easy to learn for first-time use, but has layers of additional features for those who invest the time in becoming power users.

    Advanced Roles And Features

    Almost all video conferencing apps have three different roles that a user can play: host, presenter, and participant. The names of these roles may vary with the app, but the functions are fairly standard: A host works as the master of ceremonies and manages the permissions; a presenter is a speaker or panelist and often uses screen sharing. Naturally, playing host has the greatest functionality and requires the greatest knowledge of an app. For that reason, it’s the most difficult role to support accessibly, and Zoom supports it quite well.

    Jonathan Mosen, a blind podcaster from New Zealand, has created an in-depth audiobook titled Meet me Accessibly with tutorials that teach easy to advanced tasks to fellow blind Zoom users. These include hosting Zoom meetings, conducting screen sharing, as well as doing installations, and optimizing the settings.

    Settings And Security

    Zoom allows users to tailor the app to their needs through the app’s extensive settings menus. If anything, Zoom is the Swiss Army knife of video conferencing tools with a setting for every conceivable aspect of the application. At the Lighthouse of Broward, the IT team has taken advantage of this by creating one account with settings aimed at the blind, and another account with different settings for low-vision users.

    In fact, there are 70 settings within the desktop client, and another 95 available through Zoom’s advanced settings web user interface. With 165 settings, it’s easy for any product to stumble, especially when it comes to getting the default settings right. Many of the security issues came from overly lax defaults, like meetings not requiring passwords, or re-using personal meeting room IDs for convenience.

    A screenshot of Zoom’s desktop settings menu
    Zoom’s desktop settings menu (Large preview)

    Zoom doesn’t allow settings to be saved offline or shared — yet. In an ideal world, power users in the disability community could create settings profiles that are tailored for specific disabilities and share them with others who are less experienced with Zoom.

    Different Use Cases For Accessible Conferencing

    Video With Work Chat

    With stay-at-home orders in place, many teams are finding themselves making more use of threaded work chat throughout the day. Like an email client, this type of application is one that you’d leave on all the time, so your interactions aren’t limited to the duration of an online meeting.

    A screenshot of a chat window in Slack showing theprofile picture of the person calling along with their user name, company name, and two large bold buttons with the telephone icon on both: red button to decline call and green one to accept the call via Slack
    Slack video conferencing (Image source) (Large preview)

    Slack is a hugely popular app for threaded work chat that offers the ability to call another member in a video call. It’s great for spontaneous video conversations, and its large bold buttons and resizable text are helpful for users with moderately low vision. However, the Slack interface presents challenges to screen reader users and those with serious vision impairments, like keeping track of which Slack community that you’re in when the UI is magnified.

    Still image from a video chat including four people divided into four boxes across the screen with a little window in the bottom right corner
    Microsoft Teams website (Large preview)

    If you work in an organization with an Office 365 Enterprise license, Microsoft Teams is an alternative to Slack, with its integration into other Office apps for calendaring and file sharing. The video calls on Teams can include hundreds of participants, and they’re the app of choice inside Microsoft. Teams is also favored over Slack by many blind and low-vision users.

    One-To-One Video

    People use video conferencing tools in a variety of different ways, depending on the situation and the physical abilities of the participants. If you’re conducting a job interview though, it’s crucial to know beforehand if the job candidate needs accommodations or if they can’t work with your platform of choice. Failure to provide a viable alternative could land you in a discrimination lawsuit.

    One-to-one video calls with apps like FaceTime are well-suited for therapists and other independent healthcare providers. They’re also an excellent vehicle for those who communicate using sign language, as illustrated on the Apple accessibility website.

    A still image of a video call between two people using sign language
    FaceTime is a great way for people who use sign language to communicate easily. (Image source: Apple Accessibility) (Large preview)

    Using Captions

    But what if one of the participants doesn’t know sign language? In that case, a hard-of-hearing user would benefit from the automatic captions that are available in Google Meet. Using captions as simple as clicking on the menu in the bottom right corner and clicking on the “Turn on captions” menu item. Captions appear — but only on the devices that have them turned on — so they don’t present a distraction to attendees that don’t want them.

    A still image of a video call in Google Meet with the Turn On Captions setting zoomed in to see where it can be found in the interface
    Captions can be turned both on or off during a Google Meet call. (Image source: 9to5google) (Large preview)

    Captions are a lifeline for the deaf and the hard of hearing, and are helpful to those participants who need to keep their audio turned off or who are in a noisy situation. They’re also useful for those who aren’t native speakers of the language that’s being spoken. Auto-captions may be satisfactory for a native speaker of American or Australian English — and if you’re OK with homonyms like “meat” or “meet” being mixed up occasionally. But if the speaker has a strong accent or a speech impairment, then the accuracy of the captions drops significantly.

    There’s much debate over how much accuracy is needed for captions to be viable. According to Kristina England, an accessibility expert at the University of Massachusetts, “[in] higher ed institutions we need 99% accurate captions for accommodation requests.”

    When auto-generated captions get below 90% accuracy, they can be a serious annoyance, which has won them the nickname of “craptions” in the disability community. For that reason, it’s important to always choose a product that allows users to turn the captions off on their end.

    Avoid this easy mistake: don’t use auto-captions in PowerPoint or Google Slides if you’re presenting remotely. Limit their use to live, in-person meetings instead. By having the captions on the presenter end, they become part of the meeting’s video stream and can’t be turned off by the other attendees. If you’re recording the presentation, then the inaccurate captions become an indelible part of your recording. Also, PowerPoint and Google Slides only transcribe the presenter’s audio, leaving anything spoken by the other participants out of the captions.

    Before the COVID pandemic hit, Zoom was preparing to include Live Transcription, an automatic captioning feature, in its Enterprise plan. The feature is powered by Otter.ai, an artificial intelligence service that provides real-time automated captioning and identifies the speaker like Google Meet. With the 30-fold growth of new accounts, though, Zoom was compelled to delay the launch of Live Transcription so as not to overtax its infrastructure. In the meantime, you can still get the same functionality by running Otter.ai through Zoom’s API for third-party captioning services. but that requires an Otter for Teams account for a minimum of 3 users at $30/month each.

    One important piece of advice, though: If you’re looking to host a conference or a large public event over the internet, you should not rely on auto-captions, but consider hiring a live human captioner instead. If you use Zoom, there’s another benefit, too: your meeting participants will be able to scroll back through the captions and read any parts they may have missed.

    A Word About Webex

    Webex has gained a lot of market share in the large enterprise market, particularly with its integration with Cisco’s on-site video conferencing equipment. Webex is also tightly bound with Microsoft Outlook through the Webex Productivity Tools package. When considering these combinations, Webex offers mixed levels of accessibility for users of different abilities.

    For sighted and low-vision users who can use a mouse, Webex’s integrations are quite useful. However, when a participant enters a meeting, Webex often throws pop-up dialogs to confirm their audio and video choices, which can’t be operated without a pointing device. Outlook’s Calendar and Meeting interfaces also fail to provide keyboard access, further blocking someone who can’t work a mouse from accessing a meeting.

    If your enterprise has deaf and hard-of-hearing members, Webex doesn’t offer automatic captions but it does offer a media viewer window where a human captioner can type what is spoken by a presenter during a meeting. Webex also provides transcription services for recorded meetings for customers who have Network Based Recordings enabled.

    Webex’s bold visual design works well for people who have low vision but are not blind. “The text in Webex is easy to see visually and works well when utilizing screen magnification software. The settings in Webex are intuitive and easy to figure out,” says Elizabeth Novak, a tester with Fable.

    For those users who are blind and rely on screen readers, though, Webex presents numerous barriers. Many of its interface elements have missing or confusing labels. Others can’t be reached with a keyboard.

    Webex’s spotty support for screen readers has led to a lawsuit against one of its customers: in October 2019, a blind plaintiff in Massachusetts sued her employer for not providing accessible alternatives like Zoom for Webex meetings. Beacon Health Options, the employer, was using Webex for virtual conferences and mandatory employee training in which Amy Ruell, the plaintiff, was unable to participate. The case is still ongoing at the time of this writing.

    One of the quirks of digital accessibility lawsuits is that they are rarely filed against software makers or implementers. Instead, they are almost always filed against the organization that purchases the software and makes it available to its customers, employees, and other constituents. The courts’ reasoning is that the organization should be conducting due diligence in vetting the products that it uses.

    Judging Apps By Their Accessibility Pages

    If you’re the one deciding which video conferencing platform to use in your organization, the lack of formal studies and the constant change will pose challenges. One good way to gauge a company’s commitment to accessibility is to look for an accessibility page on their website. Usually, it’s posted as a link in the footer of every page. A company that is at least aware of accessibility will have a posted accessibility policy.

    There are a handful of companies that need to meet accessibility standards quite strictly and take their documentation much further. These companies will post Voluntary Product Accessibility Template (VPAT) statements to meet Section 508 requirements for selling to the government or higher education. VPAT statements can be a goldmine of accessibility compliance information, and should be reviewed by anyone in charge of making a purchasing decision.

    Zoom and Webex both provide extensive VPAT statements on their websites with detailed information on accessibility. These disclosure statements classify the level of support for users with disabilities and assistive technologies into four categories:

    Supports Supports With Exceptions Does Not Support Not Applicable

    From these options, anything other than “Supports” should be looked at carefully or treated as a potential problem area.

    A real red flag is when a company does not have a posted policy, and where a search for “accessibility” yields results that are unrelated to disabilities or assistive technologies. This is the case with Crowdcast, where the results for “access” are related to connectivity and security. With Jitsi, developers contribute their efforts to building and refining the application as an open-source project. However, accessibility at Jitsi is sparsely resourced:

    Community post that says “Maybe after few months or more when things slow down, we will be able to take a look, but for now our hands are full and any help is welcome.”
    (Large preview)

    While this doesn’t help a person with a disability who may want to use Jitsi, it does present opportunities to developers looking to contribute where they can make a large social impact.

    If an app has advanced features or functionality that’s not directly connected to video conferencing, then support can be spotty as well. This often the case with polling, Q&A and other ancillary features, even with Zoom. For those products that do post VPATs on their websites, you can see which features are problematic. A good rule of thumb would be to avoid these features when accessibility is a concern.

    Building Accessible Products

    How do the makers of video conferencing platforms create a product that’s truly accessible for all? At Zoom accessibility is treated as a core value, beginning with CEO Eric Yuan. According to Alex Mooc, Zoom’s accessibility lead, “that executive buy-in trickles down to every team leader, every platform. Every platform owner is accountable for these accessibility issues.” The product teams are organized so that accessibility experts are embedded into every development team, unlike most other companies where accessibility comes in as oversight from a separate entity.

    Zoom has fast release cycles, where the product is tested for accessibility as it’s built. Zoom’s engineers aren’t necessarily well-versed in accessibility when they’re hired, but they’re expected to learn quickly from the company’s extensive internal knowledge.

    Most companies design and build their product for their abled users, and might consider disabled users as an afterthought, if at all. The opposite is true at Zoom: accessibility is treated as part of the product’s usability for all users, abled and otherwise. “Someone who is disabled is like a canary in a coal mine,” says Mooc, because resolving issues that they encounter helps with usability for all other users. In other words, accessibility is a means to deliver the frictionless, easy-to-use experience that Zoom strives for.

    Zoom also seeks out feedback from people in the disability community, leaning heavily on its relationships with educational institutions like the University of Washington. Some of Mooc’s research involves going to disabled users’ homes and observing how they work. With stay-at-home orders and social distancing, though, this type of research has been put on hold.

    Of all the ways to gain insight into a product’s accessibility, the best is to have live user testing by people with disabilities. Fable Tech Labs uses Zoom almost exclusively to conduct functional testing remotely with participants who are blind, have low vision, or have mobility impairments and rely on assistive technologies like switch systems and voice navigation. But doing this type of testing remotely presents its own technical challenges and requires a lot of skill and knowledge of hardware to conduct successfully. This can lead to elaborate workarounds with multiple devices and video feeds. Ideally, there would be a mixing-board-style product to handle multiple feeds in a remote connection, but that’s not in Zoom’s roadmap.

    Growth And Change

    Over the past few months, both Zoom and Google Meet have grown from 10 million users to over 300 million daily participants each. The sudden explosion in popularity has attracted malicious users who found ways to hijack insecure meetings in Zoom in a practice called “zoom-bombing”. The few incidents generated news stories that became ubiquitous in the news media, raising concerns about Zoom’s security. Some organizations like Google and the New York City school system have gone as far as banning Zoom from their computers.

    For many in the disability community, this turn of events has had an unfortunate impact on access. If organizations ban Zoom from their computers, then meetings need to be conducted and classes need to be taught online using less accessible products, which can mean trading a usable-but-insecure platform for one that simply can’t be used.

    “Security is very important, but there’s no point trying to use an inaccessible tool no matter how secure it is.”

    — Erik Burggraaf, screen reader user

    Fortunately, Zoom has been quick to respond with a series of security fixes within a couple of weeks. The overly lax defaults in the app’s 165 settings were tightened up, and a variety of changes were made to the user experience. Now it’s mostly up to users to update their Zoom apps, settings, and desktop clients.

    Since the beginning of the COVID pandemic, the rate of development in video conferencing has accelerated in a business that already had frequent releases. Major product announcements are happening weekly, and Zoom is releasing patches nearly daily, along with a major upgrade to Zoom 5.

    The net: if you use any video conferencing software, you should be diligent about updating your software as frequently as you can — or enabling auto-updates on your device. You may even find that a serious accessibility issue may be resolved with a simple software update.

    The future is looking promising though, according to accessibility tester Shane: “Accessibility is getting better each day within online communication tools. However, some still need work to be completely inclusive to all people.”

    Further Reading

    Smashing Editorial
    (ra, il)

    Source link