Beamr Technology Assessment

In December 2014, Beamr hired me to evaluate their video optimization technology and write a report on my findings. This was a paid gig, but Beamr had full control over the disclosure of the results. If positive, they could reveal them, if negative, they could withhold them. Essentially, I reviewed the product like I would any other, except that the tests were more extensive, and it was up to Beamr to make them public or not. The report was positive and Beamr has decided to make them public, and you can download the report below. 

Note that I disclosed the relationship to my editors at Streaming Media Magazine, and that I won't be reviewing products in this category to avoid any conflict of interest.

By way of background, video optimization technologies work with existing MP4 files to reduce their data rate without visible qualitative degradation. Beamr is one of a number of companies with products in this category. 

To test the technology, I used two test clips; one a movie trailer from A Walk Among the Tombstones, the other a test clip that I created. Beamr has two levels, high and best. Using the high setting, Beamr reduced the data rate in both clips by 50%; using the best setting, the data rate reduction averaged 34%. 

Qualitatively, I analyzed all test clips (there were six for both clips at different resolutions and data rates) using a "perceptually identical" standard. The high quality clips had to be perceptually identical when watched by an "average" viewer to meet the standard, while the best quality clips had to be perceptually identical when watched by an "expert" viewer to meet the standard. In the report I define the types of artifacts both types of viewers would notice and discuss how I analyzed the clips. 

Overall, the findings were impressive. Regarding the Walk trailer, I reported, "The reality was that for 99.99% of the expert and average viewers, the Beamr-processed clips would be perceptually identical to their source." 

For me, in addition to proving Beamr's claims, these results proved the reality of the product category, since the results were meaningful. On the other hand, I can't say Beamr is the best technology, as I haven't tested any others. As noted in the report, there are implementation issues related to Beamr; it's only available in the cloud today, or in a Ubuntu or RHEL command line process (I tested Ubuntu). It's applied to existing MP4 files, so integration into an adaptive workflow will require some customization.  

Overall, though, if you're spending big dollars on bandwidth, Beamr represents an opportunity to reduce it significantly, essentially delivering the benefit of HEVC/VP9 without the implementation hassles. Beyond Beamr, I can say that the product category is real and is definitely worth exploring.  

To learn more about Beamr, you can visit their website at http://beamrvideo.com/. Or, you can watch the video below from Streaming Media West, where speakers from Warner Bros, M-GO, Sony Pictures Entertainment and Yahoo! Flickr discussed their experiences with Beamr. You can read the session description below the video. 

This session discusses the best practices to improve streaming video with faster stream starts and reduced buffering through the application of media optimization. Learn how optimization of encoded files at the network ingest point helps to reduce distribution and storage costs, network peering congestion, and improve user experience for mobile devices while increasing streaming performance across Wi-Fi networks. Hear how to apply media optimization to your video content to gain smoother streaming and cheaper delivery costs, all while maintaining a high- quality user experience.


Attachments

Comments (0)

New comments are currently disabled.