iOS Devices Beat Android When It Comes To Streaming Video Quality, But Challenges Remain

Today, there is a lot of growth of video usage on mobile platforms, with tablets leading the way. Based on recent numbers, tablet penetration is growing faster than any device in recent memory jumping from 45 million U.S. adults in January 2012 to 59 million in August. With smartphone penetration exceeding 50% in the U.S. late 2012, streaming to mobile devices continues to grow.

As part of the data Conviva recently released on the quality of service metrics for streaming, when it comes to reliably delivering video to mobile, nearly half of all the streams they monitored in 2012 experienced buffering issues. A for which mobile platforms performed better, iOS devices beat Android devices in every category. The average time spent buffering per 10 minutes of content for iOS was 40.2 seconds while Android was 56.7 seconds. Average stream startup time for Android was slower than iOS (2.9 seconds versus 2.4 seconds). The percentage of Android streams that failed to start was higher than iOS (17.5% versus 13.6%).

one

No clear winner has emerged in the mobile platform war when measured against the three important parameters of buffering, start-up time and failed starts. Mobile has a long way to go when it comes to being a reliable platform for high-quality video consumption and a lot of that has to do with the mobile carriers. While they imply that they want consumers to use their networks to watch video on mobile device, they really don’t. It’s one of the reasons why they have caps on data usage and keep them so low.

two

How much video can you really consumer when you only get 2GB of data transfer a month and then have to pay $10 per GB after that? When Netflix’s average mobile video stream is delivered at 400Kbps, that means you can only consume about 11 hours of video a month before you blow through your quota.  And that does not include any data transfer from surfing the web or doing other online activities, which counts toward your quota. In reality, the average person can probably consume 5-6 hours of video a month, over 3G or 4G, without going over their cap. That’s why most video consumed via tablets is done so over WiFi, not cellular connections. Mobile video still has a long, long way to go before mass adoption, at reasonable quality, for long-form content.

  • KShively

    While these types of performance stats are certainly of interest, when it comes to live video, the bigger problem of reaching Android devices continues to be the lack of standardization between the device manufacturers and how the different versions of the Android OS handle live video. While some devices are now HTML5 compliant and can receive live streams via HLS, some can only receive RSTP delivered streams and others still rely on Flash and RTMP. It remains difficult for the content creator to know how to properly identify each Android device and ensure that the proper stream and protocol is used to deliver live video to the device.

    • Very good point; that’s one of the points I address in my article from last week, surrounding Flash Player for Mobile; while one commenter took issue with the suggestion that the Android OS has flawed core media implementations, the inconsistency is apparent for anyone whose ever tried to stream live video via RTP / RTSP without using FP.

      http://www.streamingmedia.com/Articles/Editorial/Featured-Articles/Commentary-Will-Flash-Player-Survive-89240.aspx

      • Kevin Shively

        Nice article, by the way. You are exactly right in what you say about Android OS in this regard. If only the default Android media player would, at the very least, accept RTSP (which we’ve not been able to get working), you would have that option on most devices. We sometimes have to resort to offering an RTSP link, but do it by inviting the viewer to install a free third-party player in order to view it – an imperfect solution at best!

  • Ram Krishnan

    Dan,

    Usually, buffering times of less than 0.5 seconds are not noticeable by the user. So the real interesting data would be what are the individual buffering times? Are they all less than a second and add up to 40 or 50 seconds? In this case, qualitatively, the viewer may not be able to notice the difference.

  • Mary Chernuh

    Video quality often depends on the video app which is used, right? But sometimes is difficult to find a profitable app. So, is it make sense to use search engine? For example http://www.mobyss.com