Commit Graph

21 Commits

Author SHA1 Message Date
James Zern
d640614d54 update copyright text
rather than symlink the webm/vpx terms, use the same header as libvpx to
reference in-tree files

based on the discussion in:
https://codereview.chromium.org/12771026/

Change-Id: Ia3067ecddefaa7ee01550136e00f7b3f086d4af4
2013-06-06 23:09:14 -07:00
Pascal Massimino
58ca6f65b7 rebalance method tools (-m) for methods [0..4]
(methods 5 and 6 are still untouched).

Methods #0 and #1 got much faster
Method #2 gets vastly improved in quality
Method #3 is noticeably faster for little lower quality
Method #4 (default) is 10-20% faster for comparable quality

+ update the internal doc about the methods' tools.

Example of speed difference:

Time to encode picture:
Method | Before | After
-m 0   | 1.272s | 0.517s
-m 1   | 1.295s | 0.623s
-m 2   | 2.217s | 0.834s
-m 3   | 2.816s | 2.243s
-m 4   | 3.235s | 3.014s
-m 5   | 3.668s | 3.654s
-m 6   | 8.296s | 8.235s

Change-Id: Ic41fda5de65066b3a6586cb8ae1ebb0206d47fe0
2013-02-27 02:19:20 -08:00
Pascal Massimino
5189957e07 describe rd-opt levels introduce VP8RDLevel enum
makes things somehow clearer compared to using magic constants

Change-Id: I9115cee71252511f722806427ee8a97f1a1cd95f
2013-02-26 02:20:59 -08:00
skal
e895059a05 add a -jpeg_like option
This option remaps internal parameters to better match
the expected compression curve of JPEG and produce output files
of similar size, but with better quality.

Change-Id: I96a1cbb480b1f6a0c6845a23c33dfd63f197b689
2013-02-06 14:19:16 +01:00
skal
0b9e682934 minor cosmetics
spotted in patch #34187

Change-Id: Ia706af6ef7674ec7a1d7250da08f718ed7c09e72
2012-10-03 15:22:33 +02:00
Pascal Massimino
fee6627538 detect and merge similar segments
similar = same quant and filter strength.
This save some bits in the segment map

Change-Id: I6f594474ad82bddf013278d47089e43a02e07e63
2012-10-01 21:18:03 +02:00
Pascal Massimino
b9600308e8 Merge "store prediction mode array as uint8_t[16], not int[16]." 2012-01-27 07:16:50 -08:00
Pascal Massimino
7b67881a23 store prediction mode array as uint8_t[16], not int[16].
This allow faster copy and gives a little bit of speed-up

Change-Id: I5f478229766098630b53b8a4982442fe29290ee6
2012-01-27 03:17:41 -08:00
Pascal Massimino
f76d3587e5 gcc warning fix: remove the 'const' qualifier.
Change-Id: I47b8f147f54072cb9b62669efd78c7b958aa5c08
2012-01-25 15:50:24 -08:00
Pascal Massimino
1336fa719d Only recompute level_cost_[] when needed
Add a dirty_ flag to keep track of updated probabilities and the need to
recompute the level costs.
This only makes a difference for "-m 2" method which was sub-optimal.
But it's overall cleaner to have this flag.

Change-Id: I21c71201e1d07a923d97a3adf2fbbd7d67d35433
2012-01-25 07:52:34 -08:00
Pascal Massimino
28a9d9b41a speed-up GetResidualCost()
+ misc cosmetics and code polishing

Change-Id: I5830cd2f268d64c072b1cbccc0a4674833875055
2012-01-23 02:36:38 -08:00
James Zern
ad1e163a0d cosmetics: normalize copyright headers
Change-Id: I5e2462b101e0447a4f15a1455c07131bc97a52dd
2012-01-06 14:49:06 -08:00
Vikas Arora
885f25bc83 Updated the includes to relative paths.
Change-Id: I0ea1a59dd0f6cb1df9f41e048e0c5a13e7fb8335
2011-12-01 15:30:41 +05:30
James Zern
964387ed19 use WEBP_INLINE for inline function declarations
removes a #define inline, objectionable in certain projects

Change-Id: Iebe0ce0b25a030756304d402679ef769e5f854d1
2011-11-11 10:53:58 -08:00
Pascal Massimino
e06ac0887f create a separate libwebpdsp under src/dsp
Gathers all DSP-related function (and SSE2 implementations).
Clean-up some unwanted symbolic dependencies so that webp_encode,
webp_decode and webp_dsp are truly independent libraries.

+ opportunistic clean-up:
  * remove unneeded VP8DspInitTables(), now integrated in VP8DspInit()
  * make consistent use of VP8GetCPUInfo() in the various DspInit() funcs
  * change OUT macro to DST
2011-09-13 12:29:44 -07:00
Pascal Massimino
900286e091 add a -partition_limit option to limit the number of bits used by intra4x4
Although it degrades quality, this option is useful to avoid the 512k
limit for partition #0.
If not enough to reach the lower bound of 4bits per macroblock header,
one should also limit the number of segments used (down to -segments 1)

See the man file for extra details.

Change-Id: Ia59ffac13176c85b809ddd6340d37b54ee9487ea
2011-08-23 15:58:22 -07:00
Pascal Massimino
cd12b4b0ac add the missing cost for I4/I16 mode selection
was missing from the RD-computation of intra-4x4 score.
Doesn't change anything significantly, it's just More Correct.

Change-Id: I25c5b53a810d97e6fb7f98c549fd23bbe55e1bf4
2011-08-18 20:48:06 -07:00
Pascal Massimino
e7ff3f9af6 merge two ITransforms together when applicable and change the TTransform
to return the sum directly.

output is bitwise the same, speed up 1-2%. This is preparatory to a
more efficient SSE2 implementation.

Change-Id: I0bcdf05808c93420fbe9dcb75e5e7e55a4ae5b89
2011-04-21 13:35:56 -07:00
Pascal Massimino
8b77c63262 move the quantization function to dsp.c
and add a function pointer for it

Change-Id: Ic7012e35c74a474bf140bb23cdaf9d6950026bc6
2011-03-23 17:41:00 -07:00
Pascal Massimino
b97a4003c0 simplify QuantizeBlock code a bit
Change-Id: I1130859756fa88e88957781f60a3701be354e2a3
2011-02-27 10:46:28 -08:00
Pascal Massimino
f61d14aabf a WebP encoder
converts PNG & JPEG to WebP

This is an experimental early version, with lot of room
of later optimizations in both speed and quality.

Compile with the usual `./configure && make`
Command line example is examples/cwebp

Usage:

   cwebp [options] -q quality input.png -o output.webp

where 'quality' is between 0 (poor) to 100 (very good).
Typical value is around 80.

More encoding options with 'cwebp -longhelp'

Change-Id: I577a94f6f622a0c44bdfa9daf1086ace89d45539
2011-02-18 23:54:59 -08:00