Nearly every encoding technology company talks about bandwidth savings as a differentiator of their technology, but consumers want faster video starts. In this blog post we show you how bandwidth savings equates directly to faster video starts and how our innovative product, the Rithm AI Cloud Transcoder, can help you grow revenue and retain your customers.
EuclidIQ’s breakthrough transcoder results in industry leading stream sizes; and those stream size reductions range from 15 to 75 percent depending on your video content. Obviously having smaller streams means that storage costs and bandwidth costs are proportionally lower, which is excellent for reducing your expenses. While those figures are staggering and will make a meaningful impact to your CDN costs, those same savings have a more meaningful impact to the revenue side of your business. The study below looks at how the Rithm Transcoder improves video start times with smaller stream sizes and how that impacts revenue.
Our Video Startup Time Study
Our study consisted of comparing Rithm generated H.264 streams against an industry standard transcoder and the impact upon playback of the streams. We selected Amazon AWS Elastic Transcoder to compare our Rithm Transcoder against as Elastic Transcoder is a popular choice for many businesses and uses FFmpeg, the same core technology found within Rithm.
We started with a typical Apple HLS profile that was used for both the Amazon AWS Elastic Transcoder and the Rithm Transcoder. Next, we set the Rithm MOS quality value to 3.2 to make the output video quality equal to the quality level of Amazon AWS Elastic Transcoder. Prior tests concluded that a quality level of 3.2 used within the Rithm Transcoder produced videos with quality equivalent to videos produced by Elastic Transcoder.
Apple profile input transcoder settings used in our A:B test:
Resolution Max Bitrate
1920x1080px 7800 Kbps
1920x1080px 6000 Kbps
1280x720px 4500 Kbps
1280x720px 3000 Kbps
960x540px 2000 Kbps
768x432px 1100 Kbps
640x360px 730 Kbps
We choose 11 videos to be used within the test. The videos ranged from simple talking heads to higher complexity and higher motion water scenes. The videos tested represent a broad slate of content that is commonly found on television, social media and the majority of movies.
Below are the results of 1 of 11 input videos used in the test. Note the % bitrate reduction with the Rithm HLS files shown in the right-most column. The bitrate reduction was calculated by comparing the resulting size of the video produced from Elastic Transcoder with the file produced from the Rithm Transcoder.
This dataset yields a reduction of 41.38% to 79.65% and shows the ability of removing a rung for 1080p. Another benefit of Rithm is rung removal as the second rung (1630 Kbps) can be removed automatically because it is close to the 1636 Kbps rung above. Rung reduction can reduce unnecessary and sometimes disruptive stream switching.
Next, we tested our 11 resulting streams on WiFi and 3G simulated networks. We used the Google Chrome developer network console profiles to constrain the bandwidth. We also instrumented a JW Player to measure video start times. The duration times from when you press the play button to the first rendered video was captured and logged. We also scripted the video to play sequentially at least 50 times to ensure we had enough play events sampled.
Start times on WIFI network
Rithm 1.526 seconds AWS 4.630 seconds
Start times on 3G Network
Rithm 6.776 seconds, AWS 8.964 seconds
What does 3 faster video start times mean for your revenue? Well, if you are running a video advertising-based business, it means you will be very interested in using Rithm to recapture lost potential advertising revenue. Conviva recently published a study that concluded that 11% of their 16.8 billion video views on the Conviva instrumented devices in Q1 2018 resulted in a video being exited before it started because the video startup time was simply too long. In this same report Conviva determines an average video start time of 4.53 seconds, not far off from our 4.63 seconds on the WIFI simulator, corroborating our methods and finding.
These are tremendous numbers and figures that deserves serious attention. While we do not yet know if the Rithm Transcoder benefits to video startup time is linear, or not, we can conservatively assume a 40% improvement or a 4.4% reduction (40% x 11%) in video abandonment. For instance, for every $100M of ad revenue lost you could acquire back $4.4M.
Our Rithm Transcoder is in the cloud and can be easily integrated into any existing workflow. You can quickly see how Rithm can help regain lost revenue by trying our Rithm Transcoder that is now available for use. Our simple browser UI allows you to upload videos and quickly see the savings. Why wait when you can save?