Video: How Has Per-Title Encoding Evolved from 1.0 to 2.0?
Learn more about encoding optimization at Streaming Media's next event.
Watch the complete video of this panel, DT104A: Per-Title Optimization 2.0, in the Streaming Media Conference Video Portal.
Read the complete transcript of this clip:
Zhou Wang: What's going to be the major difference between per-title encoding 1.0 versus 2.0? The difference is really about whether you’re working like an artist or working as a scientist.
What do I mean by that? If you look at the current way most people do content Per-Title organization, they do this based on their intuition or heuristic approaches. You have some kind of measurement of the complexity of the content and use that to decide if this content is more difficult or easier, and then after that you shoot in the dark. Basically, you try multiple bitrates. I have this content. I don't know exactly what bit rate I should use to encode that content. What I'm going to do is to try a few different bitrates and see what happens.
Then after you try that, you verify the results. Either subjectively you can look at the video, if you like that or not, or you can use some objective metric to test that. But everything happens after you have already encoded that video. Also, a very common problem in encoding these days is to have cross-resolution and cross-device problems. If I have a high-resolution source video as input, I need to transcode that into multiple resolutions and different bitrates, so then I look at the reference video as input and high-speed as output.
You have to somehow solve this problem, so for example Netflix is going to say, "Okay, let me take the low-resolution video, I do some interpolation, make them the same size, so that that can compare with the reference video," and these are ad hoc approaches to do these kind of comparisons.
The difficult thing is that, for the same video content, if you show that on the big TV or show that on the smaller cell phone, your experience is going to be different. I can work with this kind of thing is because, in the end, the users of my encoded video will watch the video on different kind of devices. How am I going to decide this? Basically, you have to do some other ad hoc way to deal with these problems.
So, what do I mean by per-title optimization 2.0? In this case, we really want to do it like a scientist. For myself, I have been working as a professor, and we have to teach students to do things more scientifically. How do you do these things more scientifically? Basically, whenever you talk about optimization, the first thing is not how you are going to do optimization.
The first thing is you have to give me a cost function. If you cannot define what you mean by optimal, if you do have a quantitative way to say how optimal one thing is, you don't even talk about optimization. So the first thing you'd really need to have is to have a defined, very clear optimization goal. What are you trying to optimize for?
The second thing is, once you have these things set up, you hope to hit the right target, just one time. All these things rely on a comprehensive quality rate resolution model. You have to have a very clean, nice model in order for you to do things more scientifically, and you also need to use a perceptive quality metric that is consistent with cross-resolution analysis and cross-device analysis.
In the end, the difference you can make to go from 1.0 to 2.0 is that you start with something interactive or manually you do something interactive or you do something, or something inaccurate and expensive because you have to try again and again and again, hoping you hit the target versus something you can do something automatically, you can do something very accurate and low-cost because you only have to try it once. That's the purpose of per-title optimization 2.0.
Related Articles
Bitmovin's Steve Geiger outlines the benefits of per-title encoding, how to use it optimize your delivery, and what the workflow entails in this clip from Streaming Media West 2019.
15 Apr 2020
Streaming Learning Center Principal Jan Ozer discusses why shot-based encoding makes sense for some organizations and applications in this clip from his presentation at Streaming Media East 2019.
12 Aug 2019
Streaming Learning Center Principal Jan Ozer explains per-title encoding and rates the different per-title encoding technologies in this clip from his presentation at Streaming Media East 2019.
02 Aug 2019
Per-title encoding is now a required feature on most encoding platforms. Here's an overview of the features to look for, and how to get the most out of them
21 Mar 2019
We wrap up our survey of the encding and transcoding landscape with looks at Verizon Digital Media Services, Media Excel, Comprimato, Elecard, Capella Systems, Epic Labs, EuclidIQ, and NGCodec
14 Nov 2018
We continue our survey of the encoding & transcoding market with looks at Encoding.com, Bitmovin, Brightcove, Beamr, Cisco/Synamedia, and Ericsson/MediaKind
12 Nov 2018
Encoding and transcoding are at the heart of every OTT and online video workflow. The first article in this three-part series gives an overview of the technologies and a look at three major players in the space: Harmonic, AWS Elemental, and Telestream.
08 Nov 2018
SSIMWAVE Chief Science Officer introduces the key benefits of per-title encoding optimization in this clip from his presentation at Streaming Media East.
30 Aug 2018
Netflix Director of Video Algorithms Anne Aaron describes the encoding quality testing and optimization process used for Jessica Jones and other popular Netflix series in this clip from Streaming Media East 2018.
31 May 2018