Re: [x265] [PATCH 2 of 2 V2] framefilter: comment cleanups, use pixel data type

2013-09-15 Thread Deepthi Nandakumar
On Fri, Sep 13, 2013 at 11:11 PM, Steve Borho st...@borho.org wrote: # HG changeset patch # User Steve Borho st...@borho.org # Date 1379053732 18000 # Fri Sep 13 01:28:52 2013 -0500 # Node ID b8bb66cd21bcab6505b7fe321e95875861c84bda # Parent 2614338b90d3533c2760a94fa10ffb5dee57910c

Re: [x265] [PATCH]RDLevel: Disable RDOQTS when RDO and/or TS are disabled.

2013-09-16 Thread Deepthi Nandakumar
Agreed. Resending the patch - alongwith a few cleanups to make it easier to track all elements in the param struct. On Mon, Sep 16, 2013 at 3:27 PM, Derek Buitenhuis derek.buitenh...@gmail.com wrote: On Mon, Sep 16, 2013 at 10:47 AM, Deepthi Nandakumar deep...@multicorewareinc.com wrote

Re: [x265] [PATCH v2]: CLI: remove CLI option rdoqts; cleanup

2013-09-16 Thread Deepthi Nandakumar
Thanks for pointing that out, that was indeed unintentional. Pushed the fix. On Mon, Sep 16, 2013 at 9:11 PM, Derek Buitenhuis derek.buitenh...@gmail.com wrote: On Mon, Sep 16, 2013 at 1:30 PM, Deepthi Nandakumar deep...@multicorewareinc.com wrote: # HG changeset patch # User Deepthi

Re: [x265] [PATCH] do not use std;:list for the class AccessUnit

2013-09-19 Thread Deepthi Nandakumar
FrameEncoder and compressFrame needs to be refactored for mallocs and failures to be handled correctly. On Thu, Sep 19, 2013 at 1:31 PM, Gopu Govindaswamy g...@multicorewareinc.com wrote: # HG changeset patch # User Gopu Govindaswamy g...@multicorewareinc.com # Date 1379577682 -19800 #

Re: [x265] [PATCH] lookahead: fix crash for I frame cost estimation

2013-09-19 Thread Deepthi Nandakumar
Can you try sending this as a fresh patch to the mailing list? Does not apply cleanly to the parent node. On Thu, Sep 19, 2013 at 2:46 PM, Deepthi Devaki Akkoorath deepthidev...@multicorewareinc.com wrote: # HG changeset patch # User Deepthi Devaki deepthidev...@multicorewareinc.com # Date

Re: [x265] Commit c56e392b2c68 fails to link with MinGW

2013-09-24 Thread Deepthi Nandakumar
Thanks, pushed fix. On Tue, Sep 24, 2013 at 2:14 PM, Nikos Barkas nikbar2...@yahoo.com wrote: There is a linking problem with revision c56e392b2c68 on MinGW. The error comes from common.cpp and is our old friend: undefined reference to `__sync_val_compare_and_swap_4' The file common.cpp

[x265] [PATCH] Fixed the --me 4 option error for Full search

2013-10-15 Thread Deepthi Nandakumar
# HG changeset patch # User sai...@multicorewareinc.com # Date 1381818060 -19800 # Tue Oct 15 11:51:00 2013 +0530 # Branch hsa # Node ID 5725e4986fa142c38ed7416f303f93604135c603 # Parent 0004c182f326197019a7426f080dd04ff99f39d3 Fixed the --me 4 cli option error diff -r 0004c182f326 -r

Re: [x265] [PATCH] Fixed issue with chroma 2xN block asm code

2013-10-17 Thread Deepthi Nandakumar
Pushed. On Thu, Oct 17, 2013 at 12:52 PM, prav...@multicorewareinc.com wrote: # HG changeset patch # User Praveen Tiwari # Date 1381994531 -19800 # Node ID 68f97c7f08392d387046736ae3e86095c653fd05 # Parent 5914800260d95a9bac7ce3eec2291e5c781e8422 Fixed issue with chroma 2xN block asm

Re: [x265] [PATCH] lowres : removed duplicate code for extending right and bottom margin

2013-10-21 Thread Deepthi Nandakumar
Thanks - unintended duplicate copy. Pushed the same fix before I saw this. On Mon, Oct 21, 2013 at 1:16 PM, Gopu Govindaswamy g...@multicorewareinc.com wrote: # HG changeset patch # User Gopu Govindaswamy g...@multicorewareinc.com # Date 1382341604 -19800 # Node ID

Re: [x265] [PATCH] ratecontrol: initialize frameThreads

2013-10-21 Thread Deepthi Nandakumar
://stackoverflow.com/questions/2626898/mercurial-copy-patch-to-repository seems to say hg import should work. How did you apply this one? Thanks, Le 21/10/2013 18:47, Deepthi Nandakumar a écrit : Pushed this - but the user name has not been configured properly. On Mon, Oct 21, 2013 at 6:16 PM

Re: [x265] [PATCH] no-rdo: use bit estimates from ME to calculate RDcost

2013-10-30 Thread Deepthi Nandakumar
Steve, This is part of an ongoing change to rd 0/1 where we want to replace cost = distortion + lambda*(coeff + mv bits), as opposed to that derived from the RDO process. Here, the coeff bits have not been added, only me bits are considered. I believe we'll need an exhaustive set of tests for

Re: [x265] [PATCH] no-rdo: Use entropy encoder for bit estimation

2013-10-31 Thread Deepthi Nandakumar
Posting the efficiency results FYI. The performance drop will be taken care of once the rd 0/1 refactor is complete. Before BasketballPass_416x240_50 (11.04 fps), 328.96 kb/s, Global PSNR: 35.613 After BasketballPass_416x240_50(10.70 fps), 326.64 kb/s, Global PSNR: 35.609 Before

Re: [x265] [PATCH RFC] presets: adjust presets to increase spread and align closer with x264 presets

2013-11-07 Thread Deepthi Nandakumar
Since the default preset is medium, shouldnt x265_param_default apply medium settings only? So that the if (!strcmp(preset, medium)) block here is empty?? In that case, the preset (if specified) will change only the ones which are different from medium? Also, there are a few params here that 1.

Re: [x265] [PATCH] no-rdo: giving weightage to the cost of all CU's and neighbour CU's for early exit

2013-11-07 Thread Deepthi Nandakumar
I have a few questions. 1. Do we need so many local variables? 2. Why are we adding outTempCU-cost to totalCost and then comparing against outBestCU-cost? That doesnt make much sense to me. AFAIk, outTempCU does not contain any valid data - we should remove this. 3. Should we be adding

Re: [x265] [PATCH] TShortYUV.cpp, added code to use new pixelsub_ps asm primitives

2013-11-13 Thread Deepthi Nandakumar
This patch cannot be accepted. 1. For generic 8bpp I420 optimizations, we can avoid the extra call to partitionFromSizes in chroma by passing in part, instead of partsize. This will remove the downshift and then upscale by 2 (!!) 2. We will need to handle multiple color spaces separately as luma

Re: [x265] [PATCH] pixelsub_ps, Removed unused old code

2013-11-13 Thread Deepthi Nandakumar
On hold until earlier patch has been fixed. On Wed, Nov 13, 2013 at 12:46 PM, muru...@multicorewareinc.com wrote: # HG changeset patch # User Murugan Vairavel muru...@multicorewareinc.com # Date 1384326967 -19800 # Wed Nov 13 12:46:07 2013 +0530 # Node ID

Re: [x265] [PATCH] TComYuv.cpp, blockcpy_pp asm integration

2013-11-14 Thread Deepthi Nandakumar
This cant be applied until the csp changes have been incorporated in. On Wed, Nov 13, 2013 at 4:23 PM, prav...@multicorewareinc.com wrote: # HG changeset patch # User Praveen Tiwari # Date 1384339140 -19800 # Node ID c0da70471ba63f052bd0e0cdf81af3d0ca9150a4 # Parent

Re: [x265] [PATCH] TComYuv::addAvg, primitive function for luma and chroma loops

2013-11-18 Thread Deepthi Nandakumar
Pushed. But next time, please organize your patches more clearly. 1. Add C primitive, if it does not exist. 2. Add the function pointer declarations and new primitive declarations to EncoderPrimitives struct. 3. Add testbench code for primitives. 4. Add asm code. Once all above patches have been

Re: [x265] [PATCH] b-pyramid implementation: Allow the use of B-frames as references for non B and B frames

2013-11-18 Thread Deepthi Nandakumar
In encoder::configure, there should be some check for --b-adapt 0 --bframes 0, in which case print warning and disable b-pyramid. On Mon, Nov 18, 2013 at 3:40 PM, Gopu Govindaswamy g...@multicorewareinc.com wrote: # HG changeset patch # User Gopu Govindaswamy g...@multicorewareinc.com #

Re: [x265] [PATCH] cleanup: removed unused code in pixel-a.asm

2013-11-25 Thread Deepthi Nandakumar
Does not apply at the tip. On Mon, Nov 25, 2013 at 11:40 AM, yuva...@multicorewareinc.com wrote: # HG changeset patch # User Yuvaraj Venkatesh yuva...@multicorewareinc.com # Date 1385359751 -19800 # Mon Nov 25 11:39:11 2013 +0530 # Node ID 90a80def0f1aabdf29e1f08dd0f2263d8e6af805 #

Re: [x265] [PATCH] RD merge and cost fixes: use sa8d_inter, add early-skip param that was missed in earlier commit

2013-11-28 Thread Deepthi Nandakumar
Reverting the sa8d_inter changes. This block always uses square CUs, so sa8d primitives are sufficient. # HG changeset patch # User Deepthi Nandakumar deep...@multicorewareinc.com # Date 1385631244 -19800 # Node ID 2ba6c26c9febdc8c57d3014c0cf98d4897d3992d # Parent

Re: [x265] [PATCH] log: output intra type statistics of I frame, bug fix in intra percentage calculation

2013-11-30 Thread Deepthi Nandakumar
Build fails. On Thu, Nov 28, 2013 at 11:10 AM, kavi...@multicorewareinc.com wrote: # HG changeset patch # User Kavitha Sampath kavi...@multicorewareinc.com # Date 1385616934 -19800 # Thu Nov 28 11:05:34 2013 +0530 # Branch stable # Node ID 8519dc4a5b9e53f1ed6f2f52294d7caea2803bc3 #

Re: [x265] [PATCH] aq: bug fix for hash mismatch between recon with decoded output

2013-12-03 Thread Deepthi Nandakumar
Pushed. So what are the latest results for different sequences on enabling aq-mode? On Tue, Dec 3, 2013 at 4:31 PM, Aarthi Thirumalai aar...@multicorewareinc.com wrote: # HG changeset patch # User Aarthi Thirumalai # Date 1386068495 -19800 # Tue Dec 03 16:31:35 2013 +0530 # Node ID

Re: [x265] [PATCH] rd level: use cu coeff array while computing coefficients instead of temp buffer

2013-12-16 Thread Deepthi Nandakumar
Build fails. Restore qtlayer. On Mon, Dec 16, 2013 at 3:07 PM, deepthidev...@multicorewareinc.com wrote: # HG changeset patch # User Deepthi Devaki deepthidev...@multicorewareinc.com # Date 1387185624 -19800 # Node ID f20e5ab835cd7071d9ebeabff50f6b9fef4d3e39 # Parent

Re: [x265] [PATCH] rc: Calculate the variance data for weighted prediction if aq-mode is disable

2013-12-16 Thread Deepthi Nandakumar
Am I missing something here? You have added code in the else part - if(AQ is disabled), we are already calculating variance for weightP. On Mon, Dec 16, 2013 at 5:43 PM, Gopu Govindaswamy g...@multicorewareinc.com wrote: # HG changeset patch # User Gopu Govindaswamy g...@multicorewareinc.com

Re: [x265] [PATCH] rc: Calculate the variance data for weighted prediction if aq-mode is disable

2013-12-16 Thread Deepthi Nandakumar
Ughh - there's something wrong with my hg import. All clear now. On Mon, Dec 16, 2013 at 10:34 PM, Deepthi Nandakumar deep...@multicorewareinc.com wrote: Am I missing something here? You have added code in the else part - if(AQ is disabled), we are already calculating variance for weightP

Re: [x265] [PATCH] rc: Calculate the variance data for weighted prediction if aq-mode is disable

2013-12-16 Thread Deepthi Nandakumar
What about the variance value returned by acEnergyCu? On Mon, Dec 16, 2013 at 10:43 PM, Deepthi Nandakumar deep...@multicorewareinc.com wrote: Ughh - there's something wrong with my hg import. All clear now. On Mon, Dec 16, 2013 at 10:34 PM, Deepthi Nandakumar deep

Re: [x265] Warnings during build - on Mac OS X 10.6.8

2013-12-17 Thread Deepthi Nandakumar
Thanks, should be fixed now. On Tue, Dec 17, 2013 at 4:08 AM, Selur hyb...@selur.de wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Nothing seriously, just wanted to note that building on Mac OS throws some warnings during building.

Re: [x265] [PATCH] cutree: bug fixes. correct the timescale used in getQScale()

2013-12-27 Thread Deepthi Nandakumar
On Fri, Dec 27, 2013 at 11:07 PM, Aarthi Thirumalai aar...@multicorewareinc.com wrote: # HG changeset patch # User Aarthi Thirumalai # Date 1388165826 -19800 # Fri Dec 27 23:07:06 2013 +0530 # Node ID 964e5bc90ad2a0f80980409046a13b4cbaf11a03 # Parent

Re: [x265] [PATCH] Modifications to address review comments to support multiple color space format

2014-01-11 Thread Deepthi Nandakumar
This is a CMake + VS configuration issue. A quick Google search - please read the following: http://stackoverflow.com/questions/2849517/linking-problem-fatal-error-lnk1112-module-machine-type-x64-conflicts-with-t On Sat, Jan 11, 2014 at 12:28 AM, Purvin Pandit purv...@hotmail.com wrote: I

Re: [x265] [PATCH] rd Level: improve bitrate and psnr in rd level 2

2014-01-14 Thread Deepthi Nandakumar
On Fri, Jan 10, 2014 at 2:10 PM, sumala...@multicorewareinc.com wrote: # HG changeset patch # User Sumalatha Polureddy # Date 1389343212 -19800 # Node ID 104fb077a9813f0f3bb94c370e134d0d6d180809 # Parent 80b63c3ee144e6edbafbbe281ad3d1d8505be1f6 rd Level: improve bitrate and psnr in rd

Re: [x265] Re:x265-devel Digest, Vol 8, Issue 1

2014-01-15 Thread Deepthi Nandakumar
(default: enable) (Deepthi Nandakumar) -- Message: 1 Date: Fri, 3 Jan 2014 16:19:29 +0530 From: Deepthi Nandakumar deep...@multicorewareinc.com To: Development for x265 x265-devel@videolan.org, Tom Vaughan tom.vaug

Re: [x265] [PATCH] asm : saturation bug fix for luma_vss asm routine

2014-01-30 Thread Deepthi Nandakumar
This patch is pending, right Nabajit? I havent pushed the luma_vss /chroma_vss assembly patches or the testbench edits to luma_vss. On Wed, Jan 29, 2014 at 1:56 PM, chen chenm...@163.com wrote: @@ -5105,8 +5108,9 @@ pmaddwdm5, [r6 + 3 * 16] paddd m1, m5

Re: [x265] [PATCH] testbench: stress test support for all pixelharness functions

2014-02-04 Thread Deepthi Nandakumar
Does not apply at the tip. Please pull and resend. On Wed, Feb 5, 2014 at 12:31 PM, muru...@multicorewareinc.com wrote: # HG changeset patch # User Murugan Vairavel muru...@multicorewareinc.com # Date 1391582491 -19800 # Wed Feb 05 12:11:31 2014 +0530 # Node ID

Re: [x265] [PATCH] slicetype: bug fix for cuTree, use type int32_t for listamount and propagate_amount to calculate valid propagate_cost

2014-02-06 Thread Deepthi Nandakumar
On Fri, Feb 7, 2014 at 5:48 AM, g...@multicorewareinc.com wrote: # HG changeset patch # User Gopu Govindaswamy # Date 1391732264 28800 # Thu Feb 06 16:17:44 2014 -0800 # Node ID 0198815523c1e653fee59f8b6ee58bffbfb12131 # Parent 634bc0b1c24653dd254df77cd80f96f81e71e888 slicetype: bug

Re: [x265] I would like to contribute to x265

2014-02-10 Thread Deepthi Nandakumar
Hi, Can you take a look at what the following does? Does the decoder actually detect HRD parameters? # HG changeset patch # User Deepthi Nandakumar deep...@multicorewareinc.com # Date 1387524067 -19800 # Node ID 3e794e059f7ffe0edaaf5432df5297631a0f44f6 # Parent

Re: [x265] [Bug]reconstruction yuv picture diff with HM decoder out

2014-02-17 Thread Deepthi Nandakumar
Hi, On Mon, Feb 17, 2014 at 12:38 PM, z...@rock-chips.com z...@rock-chips.comwrote: hi, x265 members 1. [Bug report] We found that x265 (version 0.7+2-4b8901ae94ece1ac ) recon yuv data diff with HM decode out when config CQP mode with QP=34 and set rd=0 or 1 or 2,I think it's a serious

Re: [x265] [PATCH] asm: added 16bpp support for dct[4x4, 8x8], idct4x4, dst4x4 and idst4x4 primitives

2014-02-18 Thread Deepthi Nandakumar
Has this been fixed? Murugan - have you reproduced/fixed this issue? On Sat, Feb 15, 2014 at 12:13 AM, Steve Borho st...@borho.org wrote: On Fri, Feb 14, 2014 at 12:39 PM, Steve Borho st...@borho.org wrote: On Fri, Feb 14, 2014 at 4:41 AM, dnyanesh...@multicorewareinc.comwrote: # HG

Re: [x265] [Bug]reconstruction yuv picture diff with HM decoder out

2014-02-19 Thread Deepthi Nandakumar
Hi, We've checked out for hash mismatch for all our hash clips with --rd 2. Can you share the source that caused a mismatch? That will help us identify the issue. Thanks, Deepthi On Mon, Feb 17, 2014 at 12:38 PM, z...@rock-chips.com z...@rock-chips.comwrote: hi, x265 members 1. [Bug

Re: [x265] [PATCH] encoder: enable VUI; set HRD parameters in SPS

2014-02-20 Thread Deepthi Nandakumar
Deepthi Nandakumar deep...@multicorewareinc.com # Date 1392883371 -19800 # Node ID 3934859d310bcc3f54ad1855dd94bd71eb0e7457 # Parent 3389061b75a486e004409ab628c46fed39d03b72 encoder: enable VUI; set HRD parameters in SPS. You can now add a VUI on the cli. Use --vui to get a vui will all default

Re: [x265] [PATCH] Add x265 API defination to api.cpp

2014-02-25 Thread Deepthi Nandakumar
Ok, looks good except x265_ssim should not be extern, and should be a part of common.cpp and not api.cpp. On Tue, Feb 25, 2014 at 2:33 PM, sa...@multicorewareinc.com wrote: # HG changeset patch # User Sagar Kotecha sa...@multicorewareinc.com # Date 1393318766 -19800 # Tue Feb 25

Re: [x265] [PATCH] all_angs_pred_32x32, asm code improvement

2014-02-27 Thread Deepthi Nandakumar
The earlier patch was pushed, Praveen. Can you send a new patch which just removes the unused statements? ___ x265-devel mailing list x265-devel@videolan.org https://mailman.videolan.org/listinfo/x265-devel

Re: [x265] [PATCH 2/2] Fix x86_64 build

2014-03-05 Thread Deepthi Nandakumar
Can you convert this to an hg patch? Git patches don't apply cleanly On Wed, Mar 5, 2014 at 4:08 PM, Rafaël Carré fun...@videolan.org wrote: --- source/CMakeLists.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/source/CMakeLists.txt b/source/CMakeLists.txt

Re: [x265] fix chroma lambda weighting

2014-03-25 Thread Deepthi Nandakumar
Thanks, very interesting fix. Can you give us some background here? Where did we the divide by 3.0 come from? On Tue, Mar 25, 2014 at 9:01 AM, Satoshi Nakagawa nakagawa...@oki.comwrote: # HG changeset patch # User Satoshi Nakagawa nakagawa...@oki.com # Date 1395672158 -32400 # Mon Mar

Re: [x265] fix: calcrecon asm code for HIGH_BIT_DEPTH

2014-04-02 Thread Deepthi Nandakumar
Hi, The dependent patch on calcRecon has been backed out. This does not apply anymore. Deepthi On Wed, Apr 2, 2014 at 12:45 PM, Satoshi Nakagawa nakagawa...@oki.comwrote: # HG changeset patch # User Satoshi Nakagawa nakagawa...@oki.com # Date 1396422746 -32400 # Wed Apr 02 16:12:26

Re: [x265] [PATCH] cli: set aq strength to 0 in CQP

2014-04-02 Thread Deepthi Nandakumar
On Wed, Apr 2, 2014 at 5:20 PM, Aarthi Priya Thirumalai aar...@multicorewareinc.com wrote: # HG changeset patch # User Aarthi Thirumalai # Date 1396439378 -19800 # Wed Apr 02 17:19:38 2014 +0530 # Node ID 32ba982c5279dce5abd718453b3c1a6affb51ce5 # Parent

Re: [x265] [PATCH] cli: set aq strength to 0 in CQP

2014-04-02 Thread Deepthi Nandakumar
Never mind, fixed. On Wed, Apr 2, 2014 at 5:37 PM, Deepthi Nandakumar deep...@multicorewareinc.com wrote: On Wed, Apr 2, 2014 at 5:20 PM, Aarthi Priya Thirumalai aar...@multicorewareinc.com wrote: # HG changeset patch # User Aarthi Thirumalai # Date 1396439378 -19800 # Wed Apr

Re: [x265] [PATCH] vbv: log frame-average QPs for VBV; even when AQ is disabled

2014-04-11 Thread Deepthi Nandakumar
Sure. On Fri, Apr 11, 2014 at 11:41 AM, Aarthi Priya Thirumalai aar...@multicorewareinc.com wrote: hello, reminder : this bug-fix is still pending. can we have this patch pushed in to fix the bug, if there is no problems with it? On Tue, Apr 8, 2014 at 8:33 PM,

Re: [x265] fix: constrained intra

2014-04-15 Thread Deepthi Nandakumar
Thanks, verified and pushed. Do you mind including some more information on this option to the docs? This only affects intra decisions in inter-slices, I think? Deepthi On Tue, Apr 15, 2014 at 10:16 AM, Satoshi Nakagawa nakagawa...@oki.comwrote: Was it generating hash mistakes without these

Re: [x265] [PATCH] vbv:reinit cu data when row vbv re-rencode is triggered

2014-04-18 Thread Deepthi Nandakumar
I see what you're trying to do here. m_baseQP is always zero in normal vbv operation, but contains previous QP values during re-encode. A better solution would be to find out which field in TComDataCU actually contains uncleared values, and clear them in TComDataCU::initCU. On Sat, Apr 19,

Re: [x265] [PATCH] cutree: adjust the rowSatd with qpoffset only for reference frames

2014-05-15 Thread Deepthi Nandakumar
Thanks a lot, yes, it is a typo. On May 16, 2014 6:20 AM, den c dnc...@gmail.com wrote: On 5/15/14, g...@multicorewareinc.com g...@multicorewareinc.com wrote: # HG changeset patch # User Gopu Govindaswamy # Date 1400146948 -19800 # Thu May 15 15:12:28 2014 +0530 # Node ID

Re: [x265] [PATCH] fix : square chroma transform expected error message

2014-05-19 Thread Deepthi Nandakumar
Can we have more detail on this, Ashok? Both trWidthC, trHeightC and widthC and heightC should have the same values. Why exactly was this assert being triggered in the regression tests? On Mon, May 19, 2014 at 7:19 PM, as...@multicorewareinc.com wrote: # HG changeset patch # User Ashok Kumar

Re: [x265] CLI: Default rate control mode in help output - CQ or CRF?

2014-05-20 Thread Deepthi Nandakumar
On Tue, May 20, 2014 at 1:30 PM, Mario *LigH* Rohkrämer cont...@ligh.dewrote: In the help output of the x265 CLI, I see both default values for CRF (28.0) and CQ (32). Are they valid at the same time? Or is the default QP outdated since CRF is the default rate control mode? Yes, CRF is the

Re: [x265] CLI: Default rate control mode in help output - CQ or CRF?

2014-05-20 Thread Deepthi Nandakumar
Sorry, yes, typo: --crf 28.0 is the default. On Tue, May 20, 2014 at 3:14 PM, Mario *LigH* Rohkrämer cont...@ligh.dewrote: Am 20.05.2014, 11:30 Uhr, schrieb Deepthi Nandakumar deep...@multicorewareinc.com: With nothing specified --crf 32 is the default for ratecontrol. No, actually

Re: [x265] [PATCH] psyrd: use psyrdcost for intra for rdLevels 5

2014-05-22 Thread Deepthi Nandakumar
psy-rd is enabled only in full-rdo decisions (rdLevel = 5). Your patch calculates intra-psy rd cost for rd=4 and lower, which is where xEncodeIntrainInter is called. On Thu, May 22, 2014 at 3:31 PM, sumala...@multicorewareinc.com wrote: # HG changeset patch # User Sumalatha

Re: [x265] [PATCH] psyrd: use psyrdcost for intra for rdLevels 5

2014-05-22 Thread Deepthi Nandakumar
); so this patch is required for rd=4 where inter uses psyrdcost and intra was using rdcost. For other lower rd levels 4, m_psyRdScale = 0, so psycost will not have much effect on RDcost On Thu, May 22, 2014 at 3:45 PM, Deepthi Nandakumar deep...@multicorewareinc.com wrote: psy-rd is enabled

Re: [x265] [PATCH RFC] psyrd: use psyrdcost for selecting best mode across depth

2014-05-26 Thread Deepthi Nandakumar
On Mon, May 26, 2014 at 3:02 PM, sumala...@multicorewareinc.com wrote: # HG changeset patch # User Sumalatha Polureddysumala...@multicorewareinc.com # Date 1401087565 -19800 # Node ID cbe5cc0e48b4122518ca732a1b32de16ba25c467 # Parent 5134e76aa729b6fece18701fdc00390c2f2ffb32 psyrd: use

Re: [x265] [PATCH] param: log CBR if vbvMaxrate, abr bitrate and vbv bufsize are equal

2014-06-05 Thread Deepthi Nandakumar
CBR is defined when vbvMaxRate = Bitrate, the buffer size need not necessarily be 1-sec. On Thu, Jun 5, 2014 at 12:14 PM, aar...@multicorewareinc.com wrote: # HG changeset patch # User Aarthi Thirumalai # Date 1401950608 -19800 # Thu Jun 05 12:13:28 2014 +0530 # Node ID

Re: [x265] [PATCH 2 of 6] rc: define default setting and validations for 2 pass states

2014-06-15 Thread Deepthi Nandakumar
On Mon, Jun 16, 2014 at 12:20 AM, aar...@multicorewareinc.com wrote: # HG changeset patch # User Aarthi Thirumalaiaar...@multicorewareinc.com # Date 1402857077 -19800 # Mon Jun 16 00:01:17 2014 +0530 # Node ID d1f636a6dc0a11657bb25be650de60345c2952c8 # Parent

Re: [x265] [PATCH 2 of 3] lambda: change chroma lambda distortion weighting to resemble x264

2014-06-16 Thread Deepthi Nandakumar
patch # User Deepthi Nandakumar deep...@multicorewareinc.com # Date 1402916716 -19800 # Mon Jun 16 16:35:16 2014 +0530 # Node ID 4d76a9c8b5abbf143e5869d55cf80a8816d99a68 # Parent ff3a85f715d43e2c21aec295426ae9dbe7c03d75 lambda: change chroma lambda distortion weighting to resemble x264

Re: [x265] [PATCH 2 of 3] lambda: change chroma lambda distortion weighting to resemble x264

2014-06-16 Thread Deepthi Nandakumar
. not sure how the values for other chroma formats were derived. On Mon, Jun 16, 2014 at 7:42 PM, Deepthi Nandakumar deep...@multicorewareinc.com wrote: Agreed, thats what I thought too. But the g_chromatable is not in the HEVC spec at all, it's just carried over from HM. On Jun 16, 2014 5

Re: [x265] [PATCH] rc: introduce param variables for 2 pass

2014-06-19 Thread Deepthi Nandakumar
Need to bump up X265_BUILD On Tue, Jun 17, 2014 at 8:06 PM, aar...@multicorewareinc.com wrote: # HG changeset patch # User Aarthi Thirumalaiaar...@multicorewareinc.com # Date 1402997162 -19800 # Tue Jun 17 14:56:02 2014 +0530 # Node ID eb3b6d39b83102dfc7c665e3055c1ffda3322e3e #

Re: [x265] [PATCH 1 of 3] Chroma QP Offset: increase chroma QP when psy-rd is enabled

2014-06-25 Thread Deepthi Nandakumar
This is primarily a visual quality improvement/psy-rd hack. In 444, since chroma resolution is on par with luma, and our eyes arent very sensitive to chroma, we increase the chroma QP so that those bits can be used up in luma. On Wed, Jun 25, 2014 at 4:35 PM, Derek Buitenhuis

Re: [x265] [PATCH 1 of 3] Chroma QP Offset: increase chroma QP when psy-rd is enabled

2014-06-25 Thread Deepthi Nandakumar
, Derek Buitenhuis derek.buitenh...@gmail.com wrote: On 6/25/2014 12:10 PM, Deepthi Nandakumar wrote: This is primarily a visual quality improvement/psy-rd hack. In 444, since chroma resolution is on par with luma, and our eyes arent very sensitive to chroma, we increase the chroma QP so that those

Re: [x265] [PATCH 2 of 2] pass TLD into class FrameFilter

2014-06-25 Thread Deepthi Nandakumar
On Tue, Jun 24, 2014 at 5:36 AM, Min Chen chenm...@163.com wrote: # HG changeset patch # User Min Chen chenm...@163.com # Date 1403568362 25200 # Node ID efa48bc0245bded1418db3c42b042acb9969146c # Parent 12c1d8aaa8f56a8f2de74c8ff1451d99d04c817d pass TLD into class FrameFilter diff -r

Re: [x265] [PATCH 2 of 3] rc: accumulate mv bits, coeff bits per frame

2014-06-26 Thread Deepthi Nandakumar
Pls fix extra newlines and whitespace nits. On Wed, Jun 25, 2014 at 10:54 PM, aar...@multicorewareinc.com wrote: # HG changeset patch # User Aarthi Thirumalaiaar...@multicorewareinc.com # Date 1403716735 -19800 # Wed Jun 25 22:48:55 2014 +0530 # Node ID

Re: [x265] [PATCH 1 of 3] Chroma QP Offset: increase chroma QP when psy-rd is enabled

2014-06-26 Thread Deepthi Nandakumar
Ok, we dont have a separate psy option.yet On Jun 26, 2014 6:00 PM, Derek Buitenhuis derek.buitenh...@gmail.com wrote: On 6/26/2014 6:35 AM, BugMaster wrote: That is separate --psy (--no-psy) option in x264 and not --psy-rd Yeah, that was my point. :) - Derek

Re: [x265] RDOQ enabled rdLevel changed

2014-06-26 Thread Deepthi Nandakumar
Yes, the rest of the patch added support for lower rd levels. On Fri, Jun 27, 2014 at 6:17 AM, Satoshi Nakagawa nakagawa...@oki.com wrote: Is this change intended? # HG changeset patch # User Sumalatha Polureddysumala...@multicorewareinc.com # Date 1403689018 -19800 # Wed Jun 25

Re: [x265] [PATCH 2 of 3] rc: accumulate mv bits, coeff bits per frame

2014-06-27 Thread Deepthi Nandakumar
-getPPS()-getCabacInitPresentFlag()) On Thu, Jun 26, 2014 at 4:25 PM, Deepthi Nandakumar deep...@multicorewareinc.com wrote: Pls fix extra newlines and whitespace nits. On Wed, Jun 25, 2014 at 10:54 PM, aar...@multicorewareinc.com wrote: # HG changeset patch # User Aarthi Thirumalaiaar

Re: [x265] [PATCH] framefilter: remove heap corruption in tld

2014-07-02 Thread Deepthi Nandakumar
The framefilter structure needs ThreadLocalData m_tld, that has to be initialised, and then used if wpp is not enabled. Not sure what you're trying to do here? On Wed, Jul 2, 2014 at 2:20 PM, Aarthi Priya Thirumalai aar...@multicorewareinc.com wrote: sry, Thg client didnt sync the changes

Re: [x265] [PATCH] framefilter: remove heap corruption in tld

2014-07-02 Thread Deepthi Nandakumar
at 2:25 PM, Deepthi Nandakumar deep...@multicorewareinc.com wrote: The framefilter structure needs ThreadLocalData m_tld, that has to be initialised, and then used if wpp is not enabled. Not sure what you're trying to do here? On Wed, Jul 2, 2014 at 2:20 PM, Aarthi Priya Thirumalai aar

Re: [x265] A compiler warning and a question

2014-07-09 Thread Deepthi Nandakumar
Thanks, that was a bug, fixed. On Wed, Jul 9, 2014 at 1:34 PM, Mario *LigH* Rohkrämer cont...@ligh.de wrote: v1.1+260-a1e46d813642 + h:/MSYS/home/Entwicklung/x265/source/Lib/TLibEncoder/TEncCu.cpp:1192:6: warning: unused parameter 'outBestPredYuv' [-Wunused-parameter] void

Re: [x265] use std::swap() for readability

2014-07-09 Thread Deepthi Nandakumar
Hello, We spent a bunch of effort last year to remove STL dependencies, since they cause serious trouble between different compilers (even between different compiler versions). This is especially since a lot of users will use x265 as a static library. Thanks, Deepthi On Wed, Jul 9, 2014 at

Re: [x265] [PATCH 1 of 4] cu, search: remove redundant m_rdGoOnSbacCoder pointer

2014-07-10 Thread Deepthi Nandakumar
This series is awesome... On Fri, Jul 11, 2014 at 6:00 AM, Steve Borho st...@borho.org wrote: # HG changeset patch # User Steve Borho st...@borho.org # Date 1405031594 18000 # Thu Jul 10 17:33:14 2014 -0500 # Node ID d6c423c66e4d5a06dc4dccbd9eade7a21f9c8adf # Parent

Re: [x265] Many more warnings by GCC 4.8.2

2014-07-20 Thread Deepthi Nandakumar
Thanks, Mario. Except for the first one, most are harmless. But, we'll be fixing those right away. On Mon, Jul 21, 2014 at 9:59 AM, Mario Rohkrämer cont...@ligh.de wrote: No panic; I know that many reasons for warnings are less than serious. Just reporting. __

Re: [x265] Custom LowRes scale

2014-07-21 Thread Deepthi Nandakumar
Thanks, this is certainly an enhancement to x265 lookahead. We would be interested in this - especially if you can also include some efficiency (bitrate vs SSIM) metrics that describe the penalty moving from X265_LOWRES_SCALE of 4 to higher scales. On Mon, Jul 21, 2014 at 8:49 PM, Nicolas

Re: [x265] [PATCH] psyrdoq: implementation of psyrdoq

2014-07-22 Thread Deepthi Nandakumar
Does it make sense to try this for DC coefficients? On Tue, Jul 22, 2014 at 9:24 PM, Steve Borho st...@borho.org wrote: On 07/22, sumala...@multicorewareinc.com wrote: # HG changeset patch # User Sumalatha Polureddysumala...@multicorewareinc.com # Date 1406032149 -19800 # Node ID

Re: [x265] [PATCH] psyrdoq: implementation of psyrdoq

2014-07-22 Thread Deepthi Nandakumar
Ok, Thats what I thought too. On Jul 23, 2014 8:55 AM, Steve Borho st...@borho.org wrote: On 07/23, Deepthi Nandakumar wrote: Does it make sense to try this for DC coefficients? my understanding is that it is not helpful, and possibly harmful. we don't want to bias the DC coefficient

Re: [x265] [PATCH] Added fast intra search option

2014-08-13 Thread Deepthi Nandakumar
There are a couple of warnings our regression tests caught with this. Can you take a look? source\encoder\predict.cpp(78): warning C4800: 'const unsigned char' : forcing value to bool 'true' or 'false' (performance warning) (IntraFilterType can be bool, I think?).

Re: [x265] [PATCH] count_nonzero primitive, downscaling quantCoeff from int32_t* to int16_t*

2014-08-14 Thread Deepthi Nandakumar
Praveen, Can you build a mercurial queue for these quant patches - so they can be reviewed and pushed in once quant is 16-bit everywhere? Thanks, Deepthi On Thu, Aug 14, 2014 at 2:01 AM, Steve Borho st...@borho.org wrote: On 08/12, prav...@multicorewareinc.com wrote: # HG changeset patch

Re: [x265] psycho-visual feature tuning, feature freeze

2014-08-19 Thread Deepthi Nandakumar
I have removed the temporary disabling we had in place for rdoq, when psy-rd was enabled since rdoq was not psy-aware at that point. RDOQ is turned on in rdLevels 4 and above. Psy-rdoq can be turned on if rdoq is enabled. Psy-rd can be turned on with or without rdoq/psy-rdoq. On Sat, Aug 16,

Re: [x265] psycho-visual feature tuning, feature freeze

2014-08-19 Thread Deepthi Nandakumar
Our general suggestion would be to use psy-rd and psy-rdoq together for best visual quality. On Tue, Aug 19, 2014 at 2:18 PM, Deepthi Nandakumar deep...@multicorewareinc.com wrote: I have removed the temporary disabling we had in place for rdoq, when psy-rd was enabled since rdoq was not psy

Re: [x265] replace g_rasterToPelX[g_zscanToRaster[idx]] by g_zscanToPelX[idx]

2014-08-25 Thread Deepthi Nandakumar
Ashok is already working on pre-calculating these inside-picture flags along with more refactors. After his refactors are in, we can check whether padding will improve performance. In fact, very likely he already has a local version of the logic in this patch. On Mon, Aug 25, 2014 at 10:46 PM,

Re: [x265] [PATCH 1 of 3] analysis: fix inter hash mistake with --cu-lossless

2014-08-26 Thread Deepthi Nandakumar
Thanks, Min. This is a solution but will affect performance with an extra TComDataCU*. I have sent another patch where I'm just re-encoding the CU if lossless is chosen as the best mode. This will not affect normal analysis. Can you review that? Deepthi On Tue, Aug 26, 2014 at 3:47 AM, Steve

Re: [x265] fix m_initSliceContext (uninitialised m_sliceQp)

2014-08-26 Thread Deepthi Nandakumar
Thanks, queued for default (does not apply on stable due to the SAO refactor). On Tue, Aug 26, 2014 at 1:55 PM, Satoshi Nakagawa nakagawa...@oki.com wrote: # HG changeset patch # User Satoshi Nakagawa nakagawa...@oki.com # Date 1409041357 -32400 # Tue Aug 26 17:22:37 2014 +0900 # Node

Re: [x265] [PATCH] Entropy: Replaced getCtxQtCbf() with table

2014-09-02 Thread Deepthi Nandakumar
Is it also possible to change codeQtCbf so that only uint32_t cbf needs to be passed in, and not entire TComDataCU*? This will help since our coming refactors will get rid of TComDataCU. On Mon, Sep 1, 2014 at 3:17 PM, Steve Borho st...@borho.org wrote: On 09/01, as...@multicorewareinc.com

Re: [x265] [PATCH] fix: hash/binary mismatch for new CU structure holds CU-specific info

2014-09-03 Thread Deepthi Nandakumar
Thanks, Ashok. Queued. On Wed, Sep 3, 2014 at 8:20 PM, as...@multicorewareinc.com wrote: # HG changeset patch # User Ashok Kumar Mishraas...@multicorewareinc.com # Date 1409753842 -19800 # Wed Sep 03 19:47:22 2014 +0530 # Node ID 00c381bf615840180bb6ce924dc9be84c060938f # Parent

Re: [x265] [PATCH] Resolve gcc warnings

2014-09-03 Thread Deepthi Nandakumar
Thanks, pushed. On Thu, Sep 4, 2014 at 4:06 AM, dtyx...@gmail.com wrote: # HG changeset patch # User David T Yuen dtyx...@gmail.com # Date 1409783716 25200 # Node ID fed3ddf3f84f3c4e4c72bcc0818a07a99ec7312e # Parent 62c4779fb0bb35d5d8a69678e9e8aa81272f0115 Resolve gcc warnings * more

Re: [x265] [PATCH 2 of 3] asm: optimize nquant by PSIGND, improve 13k cycles - 11k cycles

2014-09-03 Thread Deepthi Nandakumar
Min, Praveen has sent a number of patches on changing the entire interface for quant such that the coefficients are now 16-bit instead of 32-bit. Your patches still assume they are 32-bit? Can you review all his patches (8-10 patches) and see if we're moving in the right direction? Thanks,

Re: [x265] [PATCH] frameencoder: remove second encodeCU() pass over CTUs when SAO is disabled

2014-09-08 Thread Deepthi Nandakumar
On Sat, Sep 6, 2014 at 10:08 PM, Steve Borho st...@borho.org wrote: # HG changeset patch # User Steve Borho st...@borho.org # Date 1409932577 -7200 # Fri Sep 05 17:56:17 2014 +0200 # Node ID 07d69bce1760a28be1b1ee1821dfeb3335602422 # Parent 795878af39730deb24e2ee0e585c625084bb031b

Re: [x265] [PATCH] search.cpp: fixed type conversion warning

2014-09-09 Thread Deepthi Nandakumar
Thanks, Praveen - an alternative patch pushed to take care of this (silly) warning. On Tue, Sep 9, 2014 at 11:29 AM, prav...@multicorewareinc.com wrote: # HG changeset patch # User Praveen Tiwari # Date 1410242347 -19800 # Node ID 5026f08bd7d64ab0ee22dcc98dd034030aa65db9 # Parent

Re: [x265] [PATCH] copy_cnt: enable avx2 version of asm code

2014-09-11 Thread Deepthi Nandakumar
Would be better to combine this asm enable with the corresponding asm patch itself. I have pushed copy_cnt8, and enabled only that for now. On Wed, Sep 10, 2014 at 3:28 PM, prav...@multicorewareinc.com wrote: # HG changeset patch # User Praveen Tiwari # Date 1410343073 -19800 # Node ID

Re: [x265] [PATCH] copy_cnt: enable avx2 version of asm code

2014-09-11 Thread Deepthi Nandakumar
at 11:29 AM, Deepthi Nandakumar deep...@multicorewareinc.com wrote: Would be better to combine this asm enable with the corresponding asm patch itself. I have pushed copy_cnt8, and enabled only that for now. On Wed, Sep 10, 2014 at 3:28 PM, prav...@multicorewareinc.com wrote: # HG changeset

Re: [x265] sao: some cleanups

2014-09-12 Thread Deepthi Nandakumar
Thanks, queued for testing. On Fri, Sep 12, 2014 at 7:34 AM, Satoshi Nakagawa nakagawa...@oki.com wrote: # HG changeset patch # User Satoshi Nakagawa nakagawa...@oki.com # Date 1410487314 -32400 # Fri Sep 12 11:01:54 2014 +0900 # Node ID 8a2312df90f99b8b479940141c6dafa4b96581cf #

Re: [x265] [PATCH] Resolved gcc compiler error of mismatched type

2014-09-12 Thread Deepthi Nandakumar
Thanks, pushed. On Fri, Sep 12, 2014 at 5:57 AM, dtyx...@gmail.com wrote: # HG changeset patch # User David T Yuen dtyx...@gmail.com # Date 1410481540 25200 # Node ID 37b199156dfc27874205d6d7fadb71a00e3257d9 # Parent 7e29b10982d2eb7fd79f581d6f04184522ba Resolved gcc compiler error of

Re: [x265] [PATCH RFC] analysis: add CU specific details to encodeCU()

2014-09-14 Thread Deepthi Nandakumar
On Fri, Sep 12, 2014 at 6:05 PM, santhosh...@multicorewareinc.com wrote: # HG changeset patch # User Santhoshini Sekar santhosh...@multicorewareinc.com # Date 1410525310 -19800 # Fri Sep 12 18:05:10 2014 +0530 # Node ID bf4ebe5df0cab013e4462597b55bd505b2a6a71a # Parent

Re: [x265] [PATCH] Search: remove redundant encode coefficients in intra for performance

2014-09-14 Thread Deepthi Nandakumar
This significantly changes outputs for P and B frames. Higher bitrates and higher SSIM. Lets do full regression testing on this - and compare the bitrate/ssim for all combinations to be reasonably sure there are no bugs. On Fri, Sep 12, 2014 at 7:47 PM, as...@multicorewareinc.com wrote: # HG

Re: [x265] [PATCH] Search: remove redundant encode coefficients in intra for performance

2014-09-15 Thread Deepthi Nandakumar
Sorry, the output mismatch was due to asm. Pushed. On Sun, Sep 14, 2014 at 4:35 PM, Deepthi Nandakumar deep...@multicorewareinc.com wrote: This significantly changes outputs for P and B frames. Higher bitrates and higher SSIM. Lets do full regression testing on this - and compare the bitrate

Re: [x265] [PATCH] analysis: add CU specific details to encodeCU()

2014-09-15 Thread Deepthi Nandakumar
Thanks, looks good. On Tue, Sep 16, 2014 at 9:45 AM, santhosh...@multicorewareinc.com wrote: # HG changeset patch # User Santhoshini Sekar santhosh...@multicorewareinc.com # Date 1410840429 -19800 # Tue Sep 16 09:37:09 2014 +0530 # Node ID 50505472d3e33b775c70f2f373e1c15d17e47e66 #

Re: [x265] [PATCH] analysis: add CU specific details to encodeCU()

2014-09-16 Thread Deepthi Nandakumar
Ok, thanks. please send a follow-on patch cleaning up both encodeCU and compressCu functions. On Tue, Sep 16, 2014 at 11:32 AM, Santhoshini Sekar santhosh...@multicorewareinc.com wrote: On Tue, Sep 16, 2014 at 10:56 AM, Deepthi Nandakumar deep...@multicorewareinc.com wrote: On Tue, Sep

  1   2   3   4   >