[HEVC] x265.EXE: mingw builds

[LigH.de]

x265 1.9+200-6098ba3e0cf16b11 (oh, revision hashes are longer now): some thread pool changes, git version ID and other fixes

[-Habanero-]

Ligh, are you able to download and compile with Linux? For some reason I'm getting this error:

Edit: My bad, was using the wrong network controller for that VM.

[El Heggunte]

Nothing important (as usual ), just testing my MSYS2 environment...

commit dcd65c5ff838

no XP-compatible build this time — gomen nasai.

[El Heggunte]

commit 106a5a7dc4b3

Changelog since commit dcd65c5ff838:

go-see https://bitbucket.org/multicoreware/x265/commits/all

[El Heggunte]

commit 78ffb67a844e

Changelog since commit 106a5a7dc4b3:

go read https://bitbucket.org/multicoreware/x265/commits/all

[El Heggunte]

commit 626fcbac7ffb

Unzip, unrar, profit

[LigH.de]

A bit later, but still the same version: x265 1.9+217-626fcbac7ffb

[El Heggunte]

commit 17c0c875f27d

yes$S, it seems 2.0 is coming

EDIT: commit 836a870ba76b has been added.

[LigH.de]

x265 1.9+227-836a870ba76b: merge with stable

[El Heggunte]

commit a932b4366235

https://bitbucket.org/multicoreware/x265/commits/all

[El Heggunte]

commit 4fe3a8090c96

rcStats: add more fields to rcStats

[El Heggunte]

Today we present x265 2.0 for our intrepid guinea pigs

commit 70581d6cd065

[El Heggunte]

Just quoting:

INFO

x265 version 2.0 has been released. This release supports many new features as well as support for arm assembly optimizations for most basic pixel and me operations, as well as sao cleanups and a fully tested reconfigure functionality.

Full documentation is available at http://x265.readthedocs.org/en/stable/

===== new features =====

• uhd-bd: Enforce ultra-hd blu-ray disc parameters (overrides any other settings)
• rskip: Enables skipping recursion to analyze lower cu sizes using heuristics at different rd-levels. Provides good visual quality gains at the highest quality presets.
• rc-grain: Enables a new rate control mode specifically for grainy content. Strictly prevents qp oscillations within and between frames to avoid grain fluctuations.
• tune grain: A fully refactored and improved option to encode film grain content including qp control as well as analysis options.
• asm: Arm assembly is now enabled by default, native or cross compiled builds supported on armv6 and later systems.

===== api and key behaviour changes =====

• x265_rc_stats added to x265_picture, containing all rc decision points for that frame
• ptl: High tier is now allowed by default, chosen only if necessary
• multi-pass: First pass now uses slow-firstpass by default, enabling better rc decisions in future passes
• pools: Fix behaviour on multi-socketed windows systems, provide more flexibility in determining thread and pool counts
• abr: Improve bits allocation in the first few frames, abr reset, vbv and cutree improved

===== misc =====

• an ssim calculation bug was corrected

[LigH.de]

^ That would have fit better in the general x265 thread, not in this MinGW builds specific one.
_

x265 2.0+2-70581d6cd065 (MSYS/MinGW, GCC 5.3.0, 32+64bit 8/10/12bit EXEs+DLLs + multilib EXEs)

[El Heggunte]

{{ d e l e t e d }}

[El Heggunte]

Compiled by Barough@forum.doom9.org:

x265-v2.0+4-43ca554799c2 (MSYS/MinGW, GCC 5.4.0, 32 & 64bit 8/10/12bit multilib EXEs)

http://www42.zippyshare.com/v/y8N5UeTc/file.html

[El Heggunte]

Compiled by Barough@forum.doom9.org:

x265-v2.0+5-98a948623dc (MSYS/MinGW, GCC 5.4.0, 32 & 64bit 8/10/12bit multilib EXEs)

rc: add qpmin and qpmax options

http://www91.zippyshare.com/v/MZaPppbS/file.html

[El Heggunte]

commit 669dc9bfe7eb

µchangelog:

669dc9b     doc: clarify aq-mode/aq-strength effects

114eab1     arm: Implement psyCost_pp_8x8_neon

dfd5744     constants: g_zscanToRaster, g_rasterToZscan as const table

[El Heggunte]

To whom this may interest:

I've decided to share my new (MSYS2-based) MinGW_w64 toolchain.
The directory includes CMake, Git and Mercurial. Download, unzip, unrar and profit.
Post #1 has been updated accordingly.

https://www.mediafire.com/?3ib29yauv5kdw

[LigH.de]

Nice.

I still have my good old reliable MSYS/MinGW/GCC530 environment by XhmikosR to build x265 2.0+9-2737c6ff5f80.

[El Heggunte]

That's right, if it works, then it doesn't need to be fixed

As I explained here , mine stopped working properly after I switched to Windows 7 x64 ,
so the "upgrade" was more or less "inevitable" to me

[LigH.de]

XhmikosR recently released an MSYS / MinGW / GCC 6.1.0 package (2016-07-08); so I built x265 binaries for you to cross-compare compilers and linking types of mostly the same generic options, just including the pentium4/generic flag pair Ma suggested for Win32 builds to enforce a coarser stack alignment (at least I hope I did it right, please check).

GCC 5.3 works OK. The problem is with stack alignment. In Windows and ancient CPU the stack alignment was 4 which is too small for SSE2 functions. GCC needs to analyze code and realign the stack in functions that calls SSE2 functions -- GCC 6.1 works wrong in this area.

x265 2.0+10-5a0e139e2938 (GCC 5.3.0)
x265 2.0+10-5a0e139e2938 (GCC 6.1.0)

[El Heggunte]

commit 4be91b287e3c

asm: fix typo mistake

Note 1 — now the 32-bit build is multilib as well.

Note 2 — when encoding to high bit-depth (10 bits and 12 bits), the 32-bit build does NOT use ASM optimizations

[Ma0]

Speed comparison of 32-bit builds LigH GCC 6.1 vs. El Heggunte GCC 5.3 on i5 3450S, input file 1920x800, no options:

 bits |      GCC 6.1       |      GCC 5.3
 8bit |  83.20s (9.01 fps) |  88.42s (8.48 fps)
10bit | 328.99s (2.28 fps) | 316.26s (2.37 fps)
12bit | 328.90s (2.28 fps) | 315.68s (2.38 fps)

32-bit GCC 6.1 wins at 8-bit, GCC 5.3 wins at 10 & 12-bit.

[El Heggunte]

While the x265 developers do not release any (important) new commit...

1) encoder.pw's repository is outdated, and more importantly, its security certificate has expired
— so I removed the link to it from post #1.

2) I've just updated my MSYS2 toolchain, so that the users don't need to mess around with fstab anymore;
and if you feel that GCC 6.x already is "trustable enough", you can simply delete the contents of the mingw* folder(s) and replace them with the most recent release(s) from Nevcairiel's site (https://files.1f0.de/mingw/).

[El Heggunte]

commit 215eedc9ecc0

Changelog:

asm: replace mova by movu to avoid crashing x265_denoise_dct_sse4() 32-bit version

Expose log2_max_pic_order_count as a param

sei: Enable param option to not dump SEI messages

bug: Fix log2_max_pic_order_count when too few bits are specified

[LigH.de]

x265 2.0+16-215eedc9ecc0 (GCC 5.3.0)
x265 2.0+16-215eedc9ecc0 (GCC 6.1.0)

New CLI options:

   --qpmin <integer>             sets a hard lower limit on QP allowed to ratecontrol. Default 0
   --qpmax <integer>             sets a hard upper limit on QP allowed to ratecontrol. Default 69
...
   --log2-max-poc-lsb <integer>  Maximum of the picture order count
   --[no-]sei-dump               Emit SEI packets in bitstream. Default enabled

[El Heggunte]

Just testing GCC 6.2.0...

commit b0ec268aef60

rc: add support for qg-size 8

cli: Change name of sei discard parameter

[El Heggunte]

commit df559450949b

Latest changes:

Enable PropagateCost assembly primitive

lookahead: overloading addPicture() function

{ https://bitbucket.org/multicoreware/x265/commits/all }

P.S.: compiled with GCC 6.2 — use at your own risk :^P

[El Heggunte]

commit 7f9aeed70c0d

Latest changes:

doc: Add documentation for experimental feature '--slices'

[slice] fix bug in generate weight frame

[slice] fix bug on sliceId assign

[slice] fix warning shadowed declaration on sliceId
	
[slice] fix output change when slice=1

[slice] typecast to avoid warning, cleanup debug code

[slice] fix bug in preset medium and below (rd 0_4)

[slice] fix hash mistake bugs
	
[slice] reduce operator on calculate picHeight in SAO

[slice] Enable Slices with threading

[slice] change reconRowCount to row based reconRowFlag[]

[slices] fix compiler limit in number of options

[slice] fix chroma hash mistake when enable both slices and sao

[slice] fix slice header mistake

[slice] fix multiple-slices bug in SAO::calcSaoStatsCu_BeforeDblk

[slice] support multiple slice with sao

[slice] fix HM 16.12 crash with DECODER_CHECK_SUBSTREAM_AND_SLICE_TRAILING_BYTES,
the substream generate dependency on syntax element entry_point_offset_minus1

[slice] support multiple slice with deblock

[slice] support multiple slice

[slices] new option --slices

test: Adding tests for qg-size 8