“Time is money”, a phrase used by Benjamin Franklin in Advice to a Young Tradesman, Written by an Old One, and applies to you, me, him, her and everyone. Life is short and no one wants to waste time on useless things. It’s a way of describing treasure, or anything precious.
When it comes to the video world, we have to admit QUALITY IS MONEY.
Content producers make big investment into professional filming equipment, use expensive reference monitors to adjust every details of their works, and are hoping the audience become big fan of them. On the other side, highly entertaining content with premium quality is all what consumers paid for. The motivations of maintaining high video quality from both ends have never lacked. But undoubtedly, it puts heavy burdens to the video delivery system which always favor light traffic. Thus, a combat without smoke had started long long ago and may last forever.
The principle of “the customer is king” seems not applied to this industry. The demands of premium quality videos from the consumers are never higher before and they keep put money into high-end TVs and better service, but in the end, there is no way to tell what they got. Even now, the video service providers are still playing this game by promote the concept of HD, 4K, HDR, etc. and trying to “steal” more money out of consumers’ pocket without justifying it, objectively.
Let’s be more specific, the main problem of video delivery is that it’s extremely hard to maintain high quality and smooth delivery at the same time. The major source of quality degradation is compression or encoding, the lossy version. The size of video files is reduced (sometime significantly) after encoding, and so as quality. But the major problem is that it’s a bitrate-centric system, in which every module is optimized for bitrate. (Practically, it does consider quality by using PSNR as indicator. But this is only worse because PSNR had been widely accepted as a very poor quality indicator, and the encoder can be optimized into totally wrong direction in this way) Back to the old time, the main reason people used bitrate based encoding was because they wanted to fit movies on 716 MB CDs. That’s why the files were usually ~716 MB for movies, ~358 MB for 45 minute TV episodes, and ~179 MB for 23 minute TV episodes. But the modern compression system deserves some changes, and brings real perceptual quality in.
Actually, the famous (one of the best) H.264 encoder, x264, has a magic mode, called constant rate factor (CRF), which just needs one pass coding and is claimed to generate constant quality streams. (Quality fluctuation has been proved to hurt the overall video QoE) However, two pass Bitrate mode is used more widely than CRF, because it provides the generated file size before encoding. But Bitrate usually inferior than CRF in terms of rate-distortion performance and almost double the encoding time!
Having said above, CRF (value range 0 to 51) is not true constant quality mode. Same CRF value for different content would leads to compressed videos with drastically different perceptual quality. If the content type in a video is changing a lot, constant quality won’t be achieved for sure.
Interestingly, the major problem for video delivery now is that people know huge amount of bits are wasted during compression but can’t do anything about it. During daily operation, a large portion of 1080p videos in a fully automatic transcoding system doesn’t actually need 5Mbps to be with excellent quality, but this bitrate has to be used because they cannot afford letting the other small portion streams to be with bad quality. Let’s see whether this problem can be addressed by a quality-driven compression scheme or not.
All in all, quality is money and can be the life-saver of a business in a such crucial competitive landscape.
Is it make sense to you? Let me know in the comments or just email me @ kaizeng045 at gmail dot com.