Video compression is always about finding a compromise between bitrate and quality.
Multimedia entertainment companies, either producers, broadcasters, technology providers, or OTT providers, are working extremely hard (recruiting top talent, investing in best tools) to find an optimal balance point to grow their business, because that’s the core value that customers paid for. However and unfortunately, conventional video compression and delivery systems are always bitrate-centric and leave no clue about what’s the quality of delivered video. We have to ask is it a FAIR way to treat the content producers and customers?
Let’s first take a look at the reasons of this bitrate-centric phenomena below (pls correct me if I’m wrong).
- Bit is the standard digital unit since computer born. From day one, it is used as a currency to tell people how much information you can get or offer to others in the computer world. Thus, it’s a historical reason that people still used it to measure the cost of transmitting a bit. However, not every bit is useful, especially when it comes to the video, which is supposed to be perceived by human visual system (HVS).
- The design of communication system is regulated by bit, i.e. the bandwidth of a channel, the structure and topology of a network, are all developed using bitrate as criteria. However, for visual communication system, using bitrate as the only criteria would be dangerous.
- Bitrate can be very straightforwardly converted to the corresponding part of cost in a business. Despite the quality variation across complex video content, once the bitrate profiles are set in a video encoding or transcoding system, it’s very easy to calculate or predict a monthly CDN cost with knowledge about the average length of videos, resolutions, total number of viewing sessions, etc. This maybe the major reason that companies use Kbps on their Service Level Agreement (SLA), because all parties know its physical meaning. A common language to talk!
- More……(for me to discover)
Now here comes the problems.
Content producers spend billions of dollars to make a movie and beatify it, because they want show a vivid story to their viewers to make attraction with the best of their efforts, like the color themes, the scene compositions, even the details of human face. They want deliver highest possible quality of their works. On the other side, customers (video watchers) want to watch highest possible quality movies in their living rooms because they just bought a brand new UHD TV and still feel excited. Unfortunately, the video delivery guys ruined everything because of compression!
There will be thousands of excuses you gonna hear when you make a complain call regarding bad video quality to your service provider. Why don’t they just do the right job? Well, they want. Truly. But more and more high-end content (4K, 8K, high frame-rate, high dynamic range, etc.) has far outweight the network capacity that can be afforded. After half decade since 4K concept been promoted, people still don’t know how to compress and deliver 4K content appropriately. Plus, the bitrate-centric video compression only make it worse! How could I know what the video looks like if I have a 5Mbps internet connection at home? The only impression I got is that the video quality sometimes perfect but sometimes extremely bad.
Yes, those’re the problems that can be addressed by a quality-driven video compression scheme. It needs to provide measurable quantitative constant video quality under the constraint of affordable bandwidth. A quality-drive compressor should be able to protect the creative intent of content producers, meanwhile offer consumers knowledgeable statistics so that they’re aware of what they paid for.
Imagine in one day, you bought a movie with quality 95 at $20 and another movie with quality 80 at $15. When you make a complain call, you’ll mention that I paid $50 per month to get videos with quality 90, but my TV tells me I got 80 only and my eyes can verify that, what’s the problem?
On the other end, when content producers (FOX, HBO, Disney, you name it) deliver a movie to their broadcasters (Comcast, CNN, Turner, etc.), the deal no the SLA would be in whatever format delivery, the video quality has to be maintained above 85.
We gonna see a different world then. Would that be more interesting and reasonable?
Is it make sense to you? Let me know in the comments or just email me @ kaizeng045 at gmail dot com.