[x264-devel] how to improve compression efficiency of x264 encoder

[Vasiliy Volkov]

Looks interesting, but their promotion materials has no technical data
about how video was encoded with libx264. At least what preset was used?
What profile?

They say that they analyze image and find which parts of image should stay
blurred (movements or something like this) and which parts should use more
bits, for better human perception. Also x264 has some options which defines
some same prerequisites for bits allocation like: "psy-rd" or simple "tune"
options and adaptive quantization

Maybe they made some improvements in these algorithms?

2016-05-12 5:10 GMT+03:00 qw <applemax82 at 163.com>:

> //Are you see below comment in there: "In a recent subjective test that
> included video clips distributed across the motion-complexity spectrum,
> //IQ264x achieved a median gain of 26.5% when compared to the reference
> x264 encoder."
>
> //It means in the specially case, their algorithm may reduce bitrate (or
> call improve compression efficiency).
> //I guess the algorithm based on Dynamic Programming, it made MV and Mode prediction
> become more effect on these environment (this technology have more
> //powerful on HEVC since these longer candidate list)
>
> I think, the article tries to show that, the new algorithm can improve
> video quality, especially for video sequence with motion-complexity
> spectrum.
>
>
>
>
>
>
> At 2016-05-11 22:44:29, "chen" <chenm003 at 163.com> wrote:
>
> Hi Andrew,
>
> Are you see below comment in there: "In a recent subjective test that
> included video clips distributed across the motion-complexity spectrum,
> IQ264x achieved a median gain of 26.5% when compared to the reference x264
> encoder."
>
> It means in the specially case, their algorithm may reduce bitrate (or
> call improve compression efficiency).
> I guess the algorithm based on Dynamic Programming, it made MV and Mode prediction
> become more effect on these environment (this technology have more powerful
> on HEVC since these longer candidate list)
>
> Regards,
> Min
>
> At 2016-05-11 19:33:10,qw <applemax82 at 163.com> wrote:
>
> Hi simon,
>
> I found one article that says the current version of x264 is not good
> enough, and there is some space to improve its compression efficiency.
>
> http://www.prweb.com/releases/2016/03/prweb13296419.htm
>
>
>
>
>
> At 2016-05-11 18:03:56, "Simon Horlick" <simonhorlick at gmail.com> wrote:
>
> Hi Andrew,
>
> What algorithm is this? Is there a research paper you're referring to? I'd
> be very surprised if there were a way to make x264 20% better.
>
> Simon
>
> On 11 May 2016, at 17:20, qw <applemax82 at 163.com> wrote:
>
> Hi,
>
> I found one impressed video algorithm that can yield an average of more
> than 20% bandwidth savings, doing so without impacting video quality,
> compared with x264 encoder. The impressed video algorithm has perceptual
> quality optimization method to select optimal qp value, where visually
> important or less important regions of a video frame are determined in
> terms of spatial and temporal information.
>
> Can x264 developers do some research, and design and implement some
> similar algorithm to improve compression efficiency?
>
> Thanks!
>
> B.R.
>
> Andrew
>
>
>
>
> _______________________________________________
> x264-devel mailing list
> x264-devel at videolan.org
> https://mailman.videolan.org/listinfo/x264-devel
>
>
>
>
>
>
>
>
>
> _______________________________________________
> x264-devel mailing list
> x264-devel at videolan.org
> https://mailman.videolan.org/listinfo/x264-devel
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x264-devel/attachments/20160513/5a09bb90/attachment.html>