There is some uncertainty in the market place about the deployment of HEVC.

The situation now it is no very different than how it was at the beginning of H.264/AVC and NGCodec is expecting that things will evolve on a similar way with similar conclusions.

In our opinion the main competitors of HEVC (and of NGCodec ) it is H.264/AVC and Dolby because of HDR.

•       H.264/AVC is a pretty good codec, good performance, good encoders available, decoded everywhere and the royalties are quite reasonable. 

•       Dolby has identified the HDR is needed for UHDTV earlier than many others and they are great at communicating their technology.  As they did with Dolby Digital they may be deployed in some markets.

In NGCodec opinion moving from one codec to another is based on two main reasons:

•       Is the old codec missing something that is offered by the new one. H.264/AVC is is not great in a number of places: 1080p and 4K support, HDR and 10 bits.

•       Economics. How much the cost is reduced if the new codec is used.

The main economic issue behind HEVC today seems to be the performance of encoders is not as good as people are expecting, that means the cost advantage of using HEVC is not that high yet, but those encoders will improve over time and they will get to be closer to the 50% gains that the specification has demonstrated. Then the cost advantage of the deploying the new codec will start being more apparent.

In addition there is still limited number of HEVC decoders available and that also greatly affects the situation as the cost reduction is limited until a substantial number of decoders are deployed. This will be addressed this year when nearly every new PC computer, TV and phone will support HEVC and specially the Main10 profile.

Economics lead by the constrained compression efficiency and not patents we think was the main problem with MPEG-4 and that was the reason why MPEG-4 was not as popular as H.264/AVC and many applications continued using MPEG-2 .

The current situation with VP9 is similar to how it was in the past with VC-1. The first encoders for H.264/AVC where not as good as the VC-1 encoders that were developed faster and by a single company, that actually made a very large investment on developing that video technology. Nevertheless, over time the H.264 encoders got better and they showed a much larger advantage over VC-1. Again cost playing a role and VC-1 progressively was used in less applications.

VP9 and AOM as a opportunity

VP9 is a good codec for some applications and specially for on-line streaming where the encoding is done in Software VP9 has shown to be better than H.264/AVC and that is why it makes economic sense for Youtube to use it, but HEVC needs a much lower bit rate than H.264/AVC or VP9 and once the encoders improve it will deliver a cost advantage compared with H.264/AVC or VP9.

VP9 has a number of mayor issues:

Real time V8 or VP9 encoder deliver much lower performance than the software encoders used off line. And they are not very competitive with the performance of HEVC or H.264/AVC real time encoders and that means that for many applications VP9 has actually no performance advantage. This is an economic disadvantage for VP9 against H.264/AVC. There is also a very limited number of decoders for VP9 that can support large resolutions that is another economic disadvantage.

I addition it is unlikely that companies will take the risk of not supporting HEVC decoding, that means that actually VP9 is an additional cost on top of the original cost of supporting HEVC and H.264/AVC.

This means that VP9 does not offer much of a advantage but as VC-1 did in the past can be used to try to reduce the cost of HEVC patents and offer competition.

The IP situation may actually be worse for VP9 than HEVC as there are many patents that Google does not have control and the companies owning those patents do not have to follow RAND rules as they did not commit to do so as they do when they develop standard based codecs. This means that thinking that the royalties for VP9 are cheaper than HEVC still needs to be proved and there are a number of companies that owns IP on VP9 that may asking for royalties for VP9 if it actually becomes more popular.

In relation to patents the situation was similar with VC-1, Microsoft was offering it for free and that was the perception for some time, but at the end a patent pool was created and companies also needed to pay for VC-1, that means that advantage also disappeared (they are 313 companies that are Licensees of that VC-1 patent pool).

VC-1 was very effective on reducing the royalties paid to the AVC/H.264 patent pool an applied competition to H.264/AVC and that led to a royalty for H.264/AVC that is very reasonable and much lower than MPEG-2 for example. After the royalties for AVC/H.264 where announced by MPEG-LA, the investment on the VC-1 project was greatly reduced by the company supporting it. The main goal of VC-1 was achieved.

Standard based royalties and patent pools

In relation to royalties, once one pool with a substantial number of essential patents for an standard set up a price that sets the precedence for the other and that has been proved in case on the Google vs Microsoft case that you see some more information bellow.

As a follow up in relation to patents for H.264/HEVC and the legal case that has created a precedent you can read the following.

http://www.fosspatents.com/2012/03/these-two-charts-show-that-motorolas.html

http://www.fosspatents.com/2012/06/googles-newest-h264-royalty-demand-is.html  

Our conclusion on this, one company having patents on a standard where it has claimed that will operate under FRAND cannot ask for numbers that are much larger than what the other parties are asking for, it is very difficult to ask for a percentage of the price of the final product and it is not possible to ask for an injunction.

These means that all of this is a question of cost of deploying technology and risk management. That mean that they are going to be some market where the cost of the royalties for HEVC is not a problem. Moreover in relation to risk management it is currently extremely risky to ignore HEVC and in our opinion most companies will not take such a high risk.

In the future it could be possible that a new technology will replace HEVC, and it may be developed somewhere else, but in NGCodec opinion this can only be achieved with the large investment needed to develop this very complex codecs, this technology (as wireless standards) need many people working in a collaborative way. The members of AOM could have the capital required to do the investment required on video compression to developed a much better codec than HEVC, but without that high level of investment it will be very difficult to compete with the international standard approach used on MPEG/VCEG/JCT-VC/JVET. 4 of the 7 AOM founders continue being very active at the international standard level that means that they continue to invest also on the standard way of developing multimedia technology, as the 150-200 companies that contributed to the development of HEVC.

Conclusion

HEVC encoders will get better over time and NGCodec is contributing to raise the bar on this area.

NGCodec believes that HEVC will be widely deployed in a large number of applications, it will deliver the cost advantages required to be able to do so. The royalty situation for HEVC will take some time to get clarified but it will end up being reasonable for a large number of applications and products and will allow HEVC to be widely used.

 

Comment