Cloudinary Blog

How to Implement Smooth Video Buffering for a Better Viewing Experience

How to Improve Live Streaming using Smooth Video Buffering

This article was originally posted on Scotch.io

In the early days of the web, the only thing that mattered was getting that content out to users no matter how long it took or what resources it consumed. As a matter of fact, users seemed to understand and were ready to wait till whenever their browser's loading spinner stops and the contents displayed.

Today, though, users’ patience has run out. They are no longer willing to wait because they know we – the engineers – can afford to give them a better experience.

While web content comes in different forms, today we will focus on video. Video optimization and delivery of a good viewing experience to users doesn’t come easy. Yes, flash is gone and HTML5 is here to stay; but don’t expect that you can just drop the video tag in your HTML and you’re done.

Let’s take a look at a common approach to delivering videos, popularly known as buffering.


Sign up for Cloudinary free today!


What Video Buffering Is and Why Use It

Most videos on the web do something that looks like a pre-fetch. You may notice the progress bar showing another indicator that is greater than or equal to the current play time. This is what Youtube's streaming looks like:

YouTube

The light gray portion of the progress bar indicates the buffered content.

YouTube dynamically adjusts the quality of that portion depending on the bandwidth and CPU capacity it detects. For example, if the bandwidth is poor, YouTube will pre-fetch the low-quality version of the above video. On the other hand, if the bandwidth is great, it would pre-fetch the high-quality version for rendering.

Put this together and users have a great experience whether or not their connectivity is superb or poor.

Note
The key thing to keep in mind is not the fact that the video is pre-fetched, but rather that the pre-fetching is achieved with an intelligent strategy.

We also use this strategy at Scotch School to deliver video training courses:

Scotch School

Now that we have a basic understanding of what buffering is, let’s see how we can employ an intelligent strategy that entails buffering based on the bandwidth or CPU capacity conditions.

Meet Cloudinary's Adaptive Bitrate Streaming (HLS & MPEG)

Cloudinary offers a game-changing concept known as adaptive bitrate streaming (ABS), which is a video delivery technique that adjusts the quality of a video stream in real time according to detected bandwidth and CPU capacity. This approach enables videos to start quicker, with fewer slow video buffering interruptions and at the best possible quality for the current device and network connection, to maximize user experience.

Videos provisioned using ABS are provided in versions known as representations, each with different quality and bitrates. Each video file must be accompanied by an index file that specifies predefined segments of the video. Additionally, there is a master playlist that points to the available representations with additional information about each one.

HLS

You can use this Cloudinary feature to deliver a better viewing experience for your users at no cost.

To deliver videos using adaptive streaming, Cloudinary generates multiple copies of your video prepared at different resolutions, qualities and data rates. The copies of the videos are then automatically generated and delivered from a single original video, transcoded to either or both of the following protocols:

Now that we’ve explained the concept, let’s write some code to put the pieces of these puzzles together.

Uploading Videos to Cloudinary

Cloudinary comes fully loaded with interesting features to make managing your media files an awesome experience. These features range from being a CDN for your media files (which means you can upload and store files to Cloudinary server) to manipulating your media files (as we discussed above).

Let’s take a look at how we can upload images from our Node servers to Cloudinary. Cloudinary offers a Node SDK that facilitates image uploads so you do not have to memorize the API URLs for doing this.

Note
SDKs for PHP, Python, Ruby and .Net are also available with the same API methods.

Install the SDK by running:

Copy to clipboard
$ npm install cloudinary --save

Import cloudinary to your code base:

Copy to clipboard
var cloudinary = require('cloudinary');

Configure the SDK with your credentials (you can get one here):

Copy to clipboard
cloudinary.config({ 
  cloud_name: '[CLOUD_NAME]', 
  api_key: '[API_KEY]', 
  api_secret: '[API_SECRET]' 
});

... then use the following few lines to upload to Cloudinary:

Copy to clipboard
cloudinary.uploader.upload('dog.mp4', function(result) {
  // Upload handler
  console.log('result: ', result);
}, {
  public_id: 'my_dog',
  resource_type: 'video'
});

We can request the video from Cloudinary using the public URL we specified:

Copy to clipboard
cloudinary.video('my_dog');

The output of the above will be:

Copy to clipboard
<video poster='https://res.cloudinary.com/scotch.io/video/upload/my_dog.jpg'>
    <source src='https://res.cloudinary.com/scotch.io/video/upload/my_dog.webm' type='video/webm'>
    <source src='https://res.cloudinary.com/scotch.io/video/upload/my_dog.mp4' type='video/mp4'>
    <source src='https://res.cloudinary.com/scotch.io/video/upload/my_dog.ogv' type='video/ogg'>
</video>

This is mind blowing! Cloudinary generates the HTML for the various formats suitable for any given browser so we do not have to worry about that. Awesome, right?

Transformation with Streaming Profiles and Formats

Cloudinary uses transformations to manipulate media files. Such manipulations include:

Transformations are provided as configuration options and applied either once during upload (eager) or each time a user needs the media to be delivered (per request).

Cloudinary also provides a collection of predefined streaming profiles, where each profile defines a set of representations according to suggested best practices.

For example, the 4K profile creates eight different representations in 16:9 aspect ratio, from extremely high quality to audio only. Alternatively, the SD profile creates only three representations, all in 4:3 aspect ratio. Other commonly used profiles include the HD and Full HD profiles.

We are going to apply the streaming profile as an eager transformation to our upload logic:

Copy to clipboard
cloudinary.uploader.upload('dog.mp4', 
        function(result) {console.log(result); }, 
        { resource_type: "video", 
        eager: [
            { streaming_profile: "full_hd", format: "m3u8" }],                                   
        eager_async: true,
        eager_notification_url: "http://scotch.io/upload_completed",
        public_id: "my_dog"});
  • We initiate a usual upload process with the SDK specifying the URL of the image we want to send to our Cloudinary server
  • An eager transformation is initiated. This transformation is an array that takes a streaming profile configuration. The dog.mp4 video is encoded into HLS format using a Full HD streaming profile.
  • This process could take a while so we ensure that the eager transformation is asynchronous by setting eager_async to true

Embedding Videos

You can deliver/embed your eagerly transformed videos using the .m3u8 (HLS) or .mpd (MPEG-DASH) file format (extension) and include the streaming_profile. You can as well provide other non-adaptive streaming-related transformation options.

For example:

Copy to clipboard
cloudinary.video("my_dog.m3u8", {streaming_profile: "hd"});

Final Note

Personally, I have avoided the responsibility of implementing such a responsive video feature in a website because of being held responsible for customer frustrations. Cloudinary made this so simple, and using the solutions was free for the amount of resources I needed.

Using the adaptive streaming strategy is a proven technique that enables you to deliver an optimal, smooth viewing experience. This feature is just a small part of what Cloudinary offers for video and image management. To learn more, look the documentation or get started here.


Further Reading on Video Manipulation

Recent Blog Posts

Create Lightweight Sites With Low-Code and No-Code Technology

Consumers expect modern websites to be mainly visual. But, the more compelling and complex the related media is, the more data is involved, compounding the site’s weight. In today’s content-craving world, delivering unoptimized media can cost you because it leads to sluggish page loads, resulting in visitors abandoning your site in search of a faster alternative. In fact, a page load that takes more than three seconds can cause as many as 40% of your visitors to bounce. Given this competitive, digital-first environment, you can’t afford to lose page views, for time is of the essence.

Read more
A Blueprint for AWS-Secured Webhook Listeners for Cloudinary

tl;dr: An AWS-secured and optimized Cloudinary webhook listener for extending the Cloudinary service

Code: Github

A webhook is a communication medium for sending notifications from one platform to another about events that occurred. In place are user-defined HTTP callbacks that are triggered by specific events. When a triggered event takes place on the source site, the webhook listens to the event, collects the data, and sends it to the URL you specified in the form of an HTTP request.

Read more
New Accessibility Features for Cloudinary’s Product Gallery Widget

Cloudinary’s Product Gallery widget, which launched in 2019, has enabled many brands to effectively and efficiently showcase their products in a sleek and captivating manner, saving countless hours of development time and accelerating release cycles. By adding Cloudinary’s Product Gallery widget with its customizable UI to their product page, retailers reap numerous benefits, often turning visitors into customers in short order.

Read more
Why Successful Businesses Engage With and Convert Audiences With Visual Media

Most business buyers prefer to research purchase options online, as do many shoppers. No wonder online retail sales in the U.S. rose by 32.4% in 2020—an impressive gain of $105 billion.

For B2B and B2C businesses, text-heavy websites are no longer adequate in attracting shoppers. Instead, engaging visual media—spin images, videos, 3D models, augmented reality—are becoming a must for conveying eye-catching details and differentiators about products or services.

Read more
Making User-Generated Content (UGC) Shoppable With Cloudinary

User-generated content (UGC) is a powerful marketing tool. Not only does video complement marketing efforts for e-commerce by enabling customers to explore products in greater detail, but UGC also adds an element of trust. As a bonus, user-generated video is an exceptional opportunity for e-businesses to attract website traffic without their marketing team having to create promotional videos from scratch. User-generated content drives conversions and brand loyalty as a direct result of authentic interaction.

Read more