Why Is Most Broadband Content Still Encoded At 300Kbps?

My earlier post got me thinking about why, in 2007, the average bitrate for a video file delivered via streaming media is still encoded in 300Kbps? Is it purely because content owners don’t want to pay more money to delivery their videos at higher bitrates? Or is the trade off between quality and price acceptable for the current content business models we have today?

Over the past 10 years, we saw the bitrates increase from 20Kbps to 37Kbps, to 80Kbps, to 100Kbps and then to 300Kbps. Yes, there is some content available in 500Kbps and 750Kbps but still very, very little. We’ve been at 300Kbps for years now and I don’t yet see a shift in that bitrate being bumped up to say 500Kbps. Why is that? Are you content owners satisfied with the quality you get now and the cost to deliver that content? Are you happy with the trade off of lower quality content delivered for a cheaper price? If that’s the case, I think that’s ok. Some trade-offs are acceptable. But it has me thinking about what is stopping the increase of what we classify today as broadband video content?

I don’t think it is a lack of capacity on the networks of the CDNs who are delivering most of this content and I don’t see the viewers having any computer or end-mile limitations stopping them from getting a 500Kbps stream today. So what is stopping the adoption of high bitrate content?

I’d love to get some thoughts and feedback to this question.

  • Viewers of online content do not have a lot of patience for choppy feeds. While the trade-off of price vs. quality is part of the issue I think the greater issue is one of the maximum number of end-users being able to view the content fluidly.
    In my experience there is a large number of users with limitations far below what they could have, or those could just be the most vocal ones. One of the issues I think is the prevalence of wireless networks used by multiple users simultaneously which does choke off available bandwidth. When this happens, users complain and they are more satisfied with a 297k feed that is fluid than a 500k feed that they cannot view.
    It also has to do with the low-cost of streaming media. Windows media encoder is free, it can be used on any PC with a video capture card. There is a great deal of user-produced content and in many cases these individuals are limited by the upload capacity of a standard DSL connection and the power of their CPU. The upload capacity available at the encoding IP address is limited even when the CDN they distribute across is not.
    The major media companies can stream video at a higher bitrate and generally do. Many use variable bitrates to allow for the differing capacities of their end-users, but in the end it comes down to simplicity and what bitrate can be viewed fluidly by the most users.
    The equilibrium highest quality that can be viewed by the most users and broadcast by producers today (many of them individuals) is at about 300K-400K, and this is a result of what end-users have available to them in addition to content-producer limitations.
    I agree that this will change very fast, windows media may become obsolete in favor higher quality technologies eventually, but that will take a lot of money to produce.
    For now the average bitrate is a function of the upload capacity of those producing streaming media as well as the availability of download capacity to end-users and this is still limited because so many choose it for it’s low-cost appeal.

  • Ben is right on the money. As a B2B news publisher, we have seen that pushing beyond 300 kbps is going to get our phones ringing and email singing with complaints. A positive end user experience is the grail that we seek for our readers and our advertisers.

  • The problem is that traditional single-source streaming has difficulty reliably delivering high bit-rate video.
    Traditional single-source streaming suffers from the “weak link” problem whereby if your server is overloaded, or there is a problem on your first hop, second hop, etc., you’ll end up with lost packets which cause scrambling, stuttering, and buffering.
    The solution is multi-source streaming, whereby streams are delivered from multiple servers and converge on the user. This allows reliable, broadcast quality video that is resilient to packet loss, and works fantastically over wifi.
    If folks want to deliver a mass-market broadcast-grade video experience over broadband, you need to go beyond antiquated single-source streaming.

  • As a higher education video producer I can say that not all broadband is created equal. Even in a well-connected college town, where most students don’t live more than 10 miles from campus, we still get tons of complaints with videos encoded over 300k.
    It seems that the last mile is sometimes the issue. ISPs and Telcos promise broadband, but sometimes the real-world results are much more spotty.
    Another issue is that students often live in apartment buildings that promise “free broadband,” but actually have dozens of apartments sharing just a few DSL or T1 lines. At 10 PM on Sunday when everyone’s cramming for exams, it seems like each user’s bandwidth slows to trickle. (And of course, some students love having the excuse “but the video wouldn’t work.”)
    For content focused on faculty, grad students and researchers who need higher resolutions, we have few problems pumping out 500 – 1000k streams. But I still have qualms about doing so for a 1200-student enrollment intro course.

  • The reason content is still being served to consumers at 300kbps is no mystery. As the head of business development at NeuLion, an iPTV service company I answer this question multiple times daily. You reach more potential audience. The NeuLion set box service is a proprietary technology that lets us deliver MPEG-4 at 700kbps to a television with a quality level that is unprecedented. Next gen is a few months will operate at 500kbps. My customers who are content aggregators/creators want to reach the largest number of customers and that can only be done at slower line speeds.

  • I totally agree. 300k is the barrier for most, but there are some interesting “pushes” of video happening here:
    The $7 TV Network: Neokast brings multicasting to the masses. -031507
    http://www.pbs.org/cringely/pulpit/2007/pulpit_20070315_001831.html
    And dont forget the guys that did SKYPE, are now doing this:
    http://www.joost.com