[x265] [PATCH] presets: Updating presets to improve coding efficiency and speed
Pradeep Ramachandran
pradeep at multicorewareinc.com
Tue Dec 22 05:41:14 CET 2015
I still see a scroll-bar on windows 10 with Chrome :-(. My screen is
running at 1920x1080.
Pradeep Ramachandran, PhD
Solution Architect,
pradeeprama.info/
www.multicorewareinc.com/
Ph: +91 99627 82018
On Tue, Dec 22, 2015 at 9:59 AM, Tom Vaughan <
tom.vaughan at multicorewareinc.com> wrote:
> Yes, I was trying to get all the columns to be visible in our online
> documentation of presets. I’m getting frustrated with readthedocs.org,
> and .rst files. Restructured text documentation was a big improvement over
> what we were doing before, and it allows us to manage the documentation
> with our source control (allowing anyone to contribute changes as a
> patch). But there are serious limitations to reST as a markup language.
>
>
>
> Tom
>
>
>
> *From:* x265-devel [mailto:x265-devel-bounces at videolan.org] *On Behalf Of
> *Deepthi Nandakumar
> *Sent:* Monday, December 21, 2015 6:09 PM
> *To:* Development for x265
> *Subject:* Re: [x265] [PATCH] presets: Updating presets to improve coding
> efficiency and speed
>
>
>
> Yes, you're right, Tom. Also this patch will not apply at the current tip,
> since Tom committed a patch to presets.rst yesterday.
>
>
>
> On Tue, Dec 22, 2015 at 7:30 AM, Tom Vaughan <
> tom.vaughan at multicorewareinc.com> wrote:
>
> Pradeep,
> In presets.rst you put 2 rows in for limit-refs. The 2nd one should be
> limit-modes.
>
> +| limit-modes | 0 | 0 | 0 | 0 | 0
> | 0 | 1 | 1 | 1 | 0 |
>
> Also, I don't see limit-modes changed for slow, slower and veryslow.
>
> Please double-check that everything matches our planned changes before we
> commit this.
>
> Tom
>
>
> -----Original Message-----
> From: x265-devel [mailto:x265-devel-bounces at videolan.org] On Behalf Of
> Pradeep Ramachandran
> Sent: Monday, December 21, 2015 12:42 AM
> To: x265-devel at videolan.org
> Subject: [x265] [PATCH] presets: Updating presets to improve coding
> efficiency and speed
>
> # HG changeset patch
> # User Pradeep Ramachandran <pradeep at multicorewareinc.com> # Date
> 1450687279 -19800
> # Mon Dec 21 14:11:19 2015 +0530
> # Node ID df793ec188d2cb6ea9dc886296ccb4c7448e4523
> # Parent 0315c0c389c99185cbaa73a0e8acfe9ef60dd817
> presets: Updating presets to improve coding efficiency and speed.
>
> This commit results in vast improvements in encoding speed and efficiency
> for several presests. Most important features are increasing # references
> for some presets and enabling limit-refs and limit-mode to limit impact on
> speed.
>
> diff -r 0315c0c389c9 -r df793ec188d2 doc/reST/cli.rst
> --- a/doc/reST/cli.rst Thu Dec 17 23:27:58 2015 +0530
> +++ b/doc/reST/cli.rst Mon Dec 21 14:11:19 2015 +0530
> @@ -685,7 +685,7 @@
> (within your decoder level limits) if you enable one or
> both of these flags.
>
> - This feature is EXPERIMENTAL and functional at all RD levels.
> + Default 3.
>
> .. option:: --limit-modes, --no-limit-modes
>
> diff -r 0315c0c389c9 -r df793ec188d2 doc/reST/presets.rst
> --- a/doc/reST/presets.rst Thu Dec 17 23:27:58 2015 +0530
> +++ b/doc/reST/presets.rst Mon Dec 21 14:11:19 2015 +0530
> @@ -22,7 +22,7 @@
>
> +---------------------+-----------+-----------+----------+--------+------+
> --------+------+--------+----------+---------+
> | | ultrafast | superfast | veryfast | faster | fast
> | medium | slow | slower | veryslow | placebo |
>
> +=====================+===========+===========+==========+========+======+
> ========+======+========+==========+=========+
> -| ctu | 32 | 32 | 32 | 64 | 64
> | 64 | 64 | 64 | 64 | 64 |
> +| ctu | 32 | 32 | 64 | 64 | 64
> | 64 | 64 | 64 | 64 | 64 |
>
> +---------------------+-----------+-----------+----------+--------+------+
> --------+------+--------+----------+---------+
> | min-cu-size | 16 | 8 | 8 | 8 | 8
> | 8 | 8 | 8 | 8 | 8 |
>
> +---------------------+-----------+-----------+----------+--------+------+
> --------+------+--------+----------+---------+
> @@ -36,7 +36,9 @@
>
> +---------------------+-----------+-----------+----------+--------+------+
> --------+------+--------+----------+---------+
> | scenecut | 0 | 40 | 40 | 40 | 40
> | 40 | 40 | 40 | 40 | 40 |
>
> +---------------------+-----------+-----------+----------+--------+------+
> --------+------+--------+----------+---------+
> -| ref | 1 | 1 | 1 | 1 | 2
> | 3 | 3 | 3 | 5 | 5 |
> +| ref | 1 | 1 | 2 | 2 | 3
> | 3 | 4 | 4 | 5 | 5 |
> ++---------------------+-----------+-----------+----------+--------+------
> +--------+------+--------+----------+---------+
> +| limit-refs | 0 | 0 | 3 | 3 | 3
> | 3 | 3 | 2 | 1 | 0 |
>
> +---------------------+-----------+-----------+----------+--------+------+
> --------+------+--------+----------+---------+
> | me | dia | hex | hex | hex | hex
> | hex | star | star | star | star |
>
> +---------------------+-----------+-----------+----------+--------+------+
> --------+------+--------+----------+---------+
> @@ -48,6 +50,8 @@
>
> +---------------------+-----------+-----------+----------+--------+------+
> --------+------+--------+----------+---------+
> | amp | 0 | 0 | 0 | 0 | 0
> | 0 | 0 | 1 | 1 | 1 |
>
> +---------------------+-----------+-----------+----------+--------+------+
> --------+------+--------+----------+---------+
> +| limit-refs | 0 | 0 | 0 | 0 | 0
> | 0 | 1 | 1 | 1 | 0 |
> ++---------------------+-----------+-----------+----------+--------+------
> +--------+------+--------+----------+---------+
> | max-merge | 2 | 2 | 2 | 2 | 2
> | 2 | 3 | 3 | 4 | 5 |
>
> +---------------------+-----------+-----------+----------+--------+------+
> --------+------+--------+----------+---------+
> | early-skip | 1 | 1 | 1 | 1 | 0
> | 0 | 0 | 0 | 0 | 0 |
> @@ -66,7 +70,7 @@
>
> +---------------------+-----------+-----------+----------+--------+------+
> --------+------+--------+----------+---------+
> | aq-mode | 0 | 0 | 1 | 1 | 1
> | 1 | 1 | 1 | 1 | 1 |
>
> +---------------------+-----------+-----------+----------+--------+------+
> --------+------+--------+----------+---------+
> -| cuTree | 0 | 0 | 0 | 0 | 1
> | 1 | 1 | 1 | 1 | 1 |
> +| cuTree | 1 | 1 | 1 | 1 | 1
> | 1 | 1 | 1 | 1 | 1 |
>
> +---------------------+-----------+-----------+----------+--------+------+
> --------+------+--------+----------+---------+
> | rdLevel | 2 | 2 | 2 | 2 | 2
> | 3 | 4 | 6 | 6 | 6 |
>
> +---------------------+-----------+-----------+----------+--------+------+
> --------+------+--------+----------+---------+
> diff -r 0315c0c389c9 -r df793ec188d2 source/common/param.cpp
> --- a/source/common/param.cpp Thu Dec 17 23:27:58 2015 +0530
> +++ b/source/common/param.cpp Mon Dec 21 14:11:19 2015 +0530
> @@ -159,7 +159,7 @@
> param->subpelRefine = 2;
> param->searchRange = 57;
> param->maxNumMergeCand = 2;
> - param->limitReferences = 0;
> + param->limitReferences = 3;
> param->limitModes = 0;
> param->bEnableWeightedPred = 1;
> param->bEnableWeightedBiPred = 0;
> @@ -279,9 +279,9 @@
> param->bEnableWeightedPred = 0;
> param->rdLevel = 2;
> param->maxNumReferences = 1;
> + param->limitReferences = 0;
> param->rc.aqStrength = 0.0;
> param->rc.aqMode = X265_AQ_NONE;
> - param->rc.cuTree = 0;
> param->rc.qgSize = 32;
> param->bEnableFastIntra = 1;
> }
> @@ -296,9 +296,9 @@
> param->bEnableWeightedPred = 0;
> param->rdLevel = 2;
> param->maxNumReferences = 1;
> + param->limitReferences = 0;
> param->rc.aqStrength = 0.0;
> param->rc.aqMode = X265_AQ_NONE;
> - param->rc.cuTree = 0;
> param->rc.qgSize = 32;
> param->bEnableSAO = 0;
> param->bEnableFastIntra = 1; @@ -306,13 +306,11 @@
> else if (!strcmp(preset, "veryfast"))
> {
> param->lookaheadDepth = 15;
> - param->maxCUSize = 32;
> param->bFrameAdaptive = 0;
> param->subpelRefine = 1;
> param->bEnableEarlySkip = 1;
> param->rdLevel = 2;
> - param->maxNumReferences = 1;
> - param->rc.cuTree = 0;
> + param->maxNumReferences = 2;
> param->rc.qgSize = 32;
> param->bEnableFastIntra = 1;
> }
> @@ -322,8 +320,7 @@
> param->bFrameAdaptive = 0;
> param->bEnableEarlySkip = 1;
> param->rdLevel = 2;
> - param->maxNumReferences = 1;
> - param->rc.cuTree = 0;
> + param->maxNumReferences = 2;
> param->bEnableFastIntra = 1;
> }
> else if (!strcmp(preset, "fast")) @@ -331,7 +328,7 @@
> param->lookaheadDepth = 15;
> param->bFrameAdaptive = 0;
> param->rdLevel = 2;
> - param->maxNumReferences = 2;
> + param->maxNumReferences = 3;
> param->bEnableFastIntra = 1;
> }
> else if (!strcmp(preset, "medium")) @@ -348,6 +345,7 @@
> param->subpelRefine = 3;
> param->maxNumMergeCand = 3;
> param->searchMethod = X265_STAR_SEARCH;
> + param->maxNumReferences = 4;
> param->lookaheadSlices = 4; // limit parallelism as already
> enough work exists
> }
> else if (!strcmp(preset, "slower")) @@ -365,6 +363,8 @@
> param->subpelRefine = 3;
> param->maxNumMergeCand = 3;
> param->searchMethod = X265_STAR_SEARCH;
> + param->maxNumReferences = 4;
> + param->limitReferences = 2;
> param->bIntraInBFrames = 1;
> param->lookaheadSlices = 4; // limit parallelism as already
> enough work exists
> }
> @@ -384,6 +384,7 @@
> param->maxNumMergeCand = 4;
> param->searchMethod = X265_STAR_SEARCH;
> param->maxNumReferences = 5;
> + param->limitReferences = 1;
> param->bIntraInBFrames = 1;
> param->lookaheadSlices = 0; // disabled for best quality
> }
> @@ -405,6 +406,7 @@
> param->searchMethod = X265_STAR_SEARCH;
> param->bEnableTransformSkip = 1;
> param->maxNumReferences = 5;
> + param->limitReferences = 0;
> param->rc.bEnableSlowFirstPass = 1;
> param->bIntraInBFrames = 1;
> param->lookaheadSlices = 0; // disabled for best quality
> _______________________________________________
> x265-devel mailing list
> x265-devel at videolan.org
> https://mailman.videolan.org/listinfo/x265-devel
> _______________________________________________
> x265-devel mailing list
> x265-devel at videolan.org
> https://mailman.videolan.org/listinfo/x265-devel
>
>
>
>
> --
>
> Deepthi Nandakumar
>
> Engineering Manager, x265
>
> Multicoreware, Inc
>
> _______________________________________________
> x265-devel mailing list
> x265-devel at videolan.org
> https://mailman.videolan.org/listinfo/x265-devel
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20151222/a3b31314/attachment-0001.html>
More information about the x265-devel
mailing list