Cost Optimization for FAST Channels

FAST, short for Free Ad-Supported Streaming, has been gaining popularity among price-sensitive viewers. A fast emerging viewership base and increased growth in advertising revenues are favouring FAST streaming services.  An estimated advertising revenue through FAST streaming services is expected to be in the range of USD 5+ billion by the year 2025.

Ashok Magadum is the Founder and CEO of Vidarka Technologies; a company focused on the domain of High Quality, Cost Optimal, Scalable, and Reliable Video Streaming Solutions for VOD, Live and Low Latency Streaming covering OTT/FAST, EdTech, E-Shopping, Gaming, and many more application areas.

In order to achieve profit from FAST streaming services, it is very crucial to increase viewership and hence the Ad revenues and at the same time lower the Video infrastructure and streaming cost.

This article talks about how the Quality Based Tier (QBT) of the transcoder helps in achieving improved video streaming and lowering the streaming cost for FAST streaming.

Stages of Video Delivery for Video Streaming

FAST provides an experience akin to cable television without requiring any payment. FAST platforms offer access to a diverse array of content, allowing users to browse through distinct “channels” without needing a subscription.

Generic data flow for any streaming system is as below

Figure 1 below depicts the high-level block of video data flow for any video streaming/delivery system.

Content Generation

 The first step is to create the original video content. This can include recording video footage, producing animations, or generating any multimedia material you want to stream. This content should be of high quality and suitable for your target audience.

Content Editing

 After capturing or creating the content, it’s often necessary to edit and refine it. This can involve cutting, splicing, adding graphics, enhancing audio, and making other adjustments to ensure the video meets your quality and branding standards.

Transcoding

 To prepare the content for streaming, it needs to be re-encoded into different bitrates and resolutions. This allows for adaptive streaming, where the viewer’s device can select the most suitable video quality based on their available bandwidth. The encoded video is usually divided into smaller segments, typically lasting a few seconds each. These segments are more manageable for streaming and offer the flexibility to switch between different quality levels during playback referred as Adaptive Streaming

DRM

Digital Rights Management (DRM) for video streaming is a technology and set of techniques used to protect digital video content from unauthorized access, copying, redistribution, and piracy. DRM is essential for content providers, studios, and distributors to safeguard their intellectual property and revenue streams. As FAST channels run on ad-based revenue models, most of the FAST services do not include DRM in their data flow path

Packaging

Packaging refers to the process of preparing video content for delivery over the Internet using a specific streaming protocol. This process involves organizing and encoding video and audio data into a format suitable for streaming, along with creating metadata and manifest files to facilitate playback and adaptive streaming. HLS is the most popular format used in FAST channels.

Storage

The packaged data is stored in cloud memory which holds the data for streaming. In any streaming system, the storage stage grows with newer and newer content coming in. The primary requirements are durability, availability, performance, security, and virtually unlimited scalability at very low costs.

Streaming/CDN

 CDNs are distributed networks of servers strategically placed in data centres around the world. They are designed to efficiently distribute content to viewers, bringing the content closer to the end-users. CDNs cache or replicate content, such as video segments and files, on their edge servers. When a viewer requests content, the CDN delivers it from the nearest edge server rather than fetching it from the origin server. This reduces latency and speeds up content delivery. CDNs employ load-balancing techniques to distribute incoming traffic across their network of edge servers. This ensures that no single server is overwhelmed and that content is delivered quickly.

Playout

 Video playout devices, also known as video playback devices, are hardware or software tools designed to play and display video content on various screens or monitors. These devices are used in a wide range of applications, from consumer entertainment to professional broadcasting.

Here are some common video playout devices: Smartphones and Tablets, Computers and Laptops, Streaming Devices like Roku, Amazon Fire TV, Apple TV, Google Chromecast, Set-Top Boxes, Game Consoles, Smart TVs, and many more. Seamless playout, many users’ specific features, search engines, and analytics data are major requirements for players.

Ads for FAST

Ads are the monetization part of FAST where they play a key role for the FAST revenues. Video Streaming AD servers play a pivotal role in the online advertising ecosystem, helping content providers and advertisers effectively reach their target audiences while generating revenue from video content. They ensure that ads are displayed seamlessly within the video stream, enhancing the viewer’s experience while providing valuable data and analytics to optimize ad campaigns. Ad Scheduling and Management, Ad selection, insertion, Targeting, Personalization, Reporting and analysis and some of the key features of Ad servers 

Cost contributors

Content Creation

Content creation is a costly affair and is one of the major cost factors in the chain of Video delivery to end users. The cost of video content creation can vary significantly depending on several factors, including the type of video, its quality, production requirements, and the creative team involved

Here are some key considerations that can impact the cost of creating video content:

Type of Video, Quality and Complexity, Length of the Video, Crew and Talent, Equipment and Technology, Location and Set, Post-Production, Licensing and Royalties, Language and Subtitling: Legal and Compliance, Distribution and Promotion etc.

Streaming / CDN Cost

Most of the time CDN cost is one of the biggest components of operating cost for FAST channels. With an increase in the number of subscribers and their screen sizes video bandwidth consumption increases. Viewers are watching higher-resolution video content, and this increases CDN usage and CDN cost exponentially.

For SVOD and TVOD, one way to mitigate the increased cost is with higher subscription fees, but in price-sensitive markets, a jump in subscription prices usually results in a loss of subscribers.

For FAST offerings in order to make a profit, the revenue made from Ads should be as high as possible and at the same time, the operator must reduce the Video streaming cost as much as possible.

The best possible option to address this problem is to reduce the CDN costs while improving the video quality.

QBT for Cost Optimization

Quality-Based Tier of Transcoding analyses the Video complexity and understands the bitrate requirements and compression complexity. Based on this data it allocates appropriate bits for different frames in video scenes so that at minimal bitrate, high-video quality video can be achieved.

For lower complexity video scenes, lesser bitrate is consumed so that overall video bits consumption can be reduced and hence overall video bandwidth consumption can be reduced significantly. The single pass history and current video frame-based analysis helps in achieving QBT for Live as well as offline transcoding applications.

The High-Quality Lower Bandwidth videos help FAST channels to reduce the CDN cost and hence improve the cost savings. For SVOD and TVOD streaming applications, better profit can be derived from subscriptions as the CDN cost can be reduced significantly.

The below table lists the comparison data of VMAF and Bitrate for different test contents.

Test Video for H264	1080P Bitrate in Kbps	1080P VMAF	Bitrate savings w.r.t to 3Mbps	4K Bit Rate in Kbps	4K VMAF	Bitrate savings w.r.t 12 Mbps
Movie clip1	1462.02	98.49	51.27%	6435.83	99.10	46.37%
Movie clip2	2122.46	91.55	29.25%	9854.16	98.00	17.88%
Movie clip3	1463.95	98.07	51.20%	6790.74	99.62	43.41%
Movie clip4	1893.29	99.93	36.89%	6707.93	99.99	44.10%
Serial clip1	1770.20	98.63	40.99%	9302.55	99.60	22.48%
Serial clip2	1767.29	98.63	41.09%	8627.89	99.96	28.10%
Serial clip3	2244.65	99.60	25.18%	6411.67	99.86	46.57%
Serial clip4	1444.19	99.53	51.86%	6612.40	99.97	44.90%
News clip1	1839.62	98.82	38.68%	8608.66	99.72	28.26%
News clip2	1963.35	96.75	34.56%	7323.22	99.05	38.97%
News clip3	1895.84	97.02	36.81%	7749.47	99.56	35.42%
News clip4	1947.94	99.74	35.07%	7681.94	99.92	35.98%
AVERAGE	1817.90	98.06	39.40%	7675.54	99.53	36.04%

HEVC Tests	1080P Bit Rate in Kbps	1080P VMAF	Bitrate savings w.r.t to 2.2Mbps	4K Bit Rate in Kbps	4K VMAF	Bitrate savings w.r.t 6 Mbps
Tears Of steel clip	1159.00	94.83	47.32%	3308.00	98.86	44.87%
Big Buck Bunny clip	1126.00	95.68	48.82%	3269.00	97.22	45.52%

Summary and Conclusion

For FAST, Linear and VOD video streaming solutions, the major cost contributors are content production/acquisition and CDN. QBT technology can provide both bitrate savings as well as improvement in the video quality for AVC and HEVC and for 4K, 1080P and other lower resolutions.

The above clips shows that H.264 QBT achieves an average VMAF of 98 with an average bitrate of 1.8Mbps for 1080P and for 4K with an average bitrate of 7.6 Mbps, a VMAF of 99 can be achieved.

For HEVC QBT for 1080P, 95+ VMAF can be achieved at around 1.2 Mbps and for 4K 97+ VMAF can be achieved at around 3.3 Mbps.

Nearly 17% to 51% bitrate savings are seen leading to significant in savings in CDN bills.